Dear Colleagues,

I'd like to draw your attention to RuleML-2008
(http://2008.ruleml.org) and ask you to consider submitting a paper or
demo (or both) to it.

We are building an interesting program. Highlights include:

- Keynotes:
    * David Luckham (Stanford University, USA) on complex event
      processing.
    * Michael Kifer (State University of New York at Stony Brook, USA),
      on W3C's Rule Interchange Format (RIF).
    * Paul Haley (Haley Ltd) on business rules.
    * 4th speaker (pending)
- Joint Lunch Panel with the Business Rules Forum about “Rules on the
   Web”
- RuleML-2008 Challenge with prestigious prices
- Lightning talks / Highlight talks

Papers will be published in Springer LNCS proceedings and a special
journal issue (IEEE TKDE pending) is forthcoming.

Please find below the link to the CFP. I would also like to ask you to
forward the CFP to your colleagues.

http://2008.ruleml.org
Deadline 2 June

CFP: http://www.cs.manchester.ac.uk/ruleML/cfp.pdf
The submission Site is open:
    http://www.easychair.org/conferences/?conf=ruleml2008
PRESS RELEASE can be found at: http://www.targetwire.com/vpo/rm/
Sponsoring opportunities: http://2008.ruleml.org/sponsoring/


Thanks a lot.


Best regards,
Adrian




[ our apologies should you receive this message more than one time ]

             2008 International RuleML Symposium
        on Rule Interchange and Applications (RuleML-2008)

           October 30-31, 2008, Orlando, Florida
                http://2008.ruleml.org

--------------------------------------------------------------------
Deadlines are approaching: June 2

Proceedings published by Springer LNCS

Submission Site is open:
        http://www.easychair.org/conferences/?conf=ruleml2008

PRESS RELEASE: http://www.targetwire.com/vpo/rm/

Sponsoring opportunities: http://2008.ruleml.org/sponsoring/
--------------------------------------------------------------------

=====================================================================
Co-located with:

          The 11th International Business Rules Forum
               http://www.businessrulesforum.com
=====================================================================
Sponsored by:

Silver level: Model Systems
Bronze level: STI Innsbruck, ruleCore, JBoss
=====================================================================
In Co-operation with:

AAAI, W3C, BPM-Forum, Business Rules Forum , ECCAI, OASIS, OMG,
European Business Rules Conference, Belgium Business Rules Forum,
MIT Sloan CIO Symposium,
ACM, ACM SIGART, ACM SIGMIS, ACM SIGWEB, Open Research Society,
IEEE Systems Man and Cybernetics Society
IEEE SMCS TC on Intelligent Internet Systems
IEEE SMCS TC on Distributed Intelligent Systems
IEEE Computer Society TC on Autonomous and Autonomic Systems
=====================================================================
Media Partners:

                   Springer LNCS, MoDo Marketing
=====================================================================


Call for Papers

Collocated with the 11th International Business Rules Forum, the 2008
International Symposium on Rule Interchange and Applications
(RuleML-2008) is
the second symposium (after last year's highly successful RuleML-2007 -
http://2007.ruleml.org/) devoted to work on practical distributed rule
technologies and rule-based applications which need language standards for
rules operating in the context of, e.g., the Semantic Web, Intelligent
Multi-
Agent Systems, Event-Driven Architectures and Service-Oriented Computing
Applications. The RuleML symposium is a new kind of event where the Web
Rules
and Logic community joins the established, practically oriented Forum
of the
Business Rules community (http://www.businessrulesforum.com) to help
cross-
fertilizing between Web and Business Logic technology.

The goal of RuleML-2008 is to bring together rule system providers,
representatives of, and participants in, rule standardization efforts
(e.g.,
SBVR, RuleML, RIF, PRR, CL) and open source rules communities (e.g., jBoss
Rules, CLIPS/Jess, Prova, OO jDrew, Mandarax, XSB, XQuery),
practitioners and
technical experts, developers, users, and researchers. They will be
offered
an exciting venue to exchange new ideas, practical developments and
experiences on issues pertinent to the interchange and application of
rules
in open distributed environments such as the Web.

The Symposium gives emphasis on practical issues such as technical
contributions and show case demonstrations of effective, practical,
deployable rule-based technologies, rule interchange formats and
applications
as well as discussions of lessons learned that have to be taken into
account
when employing rule-based technologies in distributed, (partially) open,
heterogeneous environments. We also welcome groundwork that helps to
build an
effective, practical, and deployable rule standard, improve rule
technology,
provide better understanding of the integration and interchange of
rules, and
make the current generation of rule engines and rule technology more
usable
for advanced Web and Service Oriented Architectures.

RuleML-2008 highlights include:
- Plenary keynotes, highlight/lightning session, and a joint Boxed Lunch
Panel
   about "Rules on the Web" together with the Business Rules Forum
featuring
   prominent and visionary speakers.
- Keynote speakers:
   * Michael Kifer (State University of New York at Stony Brook, USA),
     on W3C's Rule Interchange Format (RIF).
   * David Luckham (Stanford University, USA) on complex event processing.
   * Paul Haley (Haley Ltd) on business rules.
- A RuleML-2008 Challenge with prizes to demonstrate tools, use cases, and
   applications
- Industry, demo and scientific research & development papers and
   presentations advancing and assessing the state of the art in event and
   rule- based systems selected in a peer-reviewed fashion by an
international
   program committee
- Invited talks given by leaders from industry and world-class experts
   featuring practical topics on event and rule-based computing and
industry
   success stories
- Social events to promote networking among the symposium delegates in an
   informal setting.


Topics of Interest
------------------

We invite industry practitioners, rule system providers, technical experts
and developers, rule users, and researchers who are using rule-based
systems,
developing systems and applications, or exploring problems and best
practices
(especially in the areas of system interoperability, rule interchange, or
business agility), to share their ideas, results, and experiences. We
invite
submissions related (but not limited to) to one or more of the following
topics:

- Representation and meta-annotation of rules and rule sets (modules) for
   publication and interchange
- Collaborative authoring, modeling and engineering of rule specifications
   and rule repositories
- Information integration of external data and domain knowledge into rules
- Homogeneous and heterogeneous integration of  rules and ontologies
- Rules in Web 2.0 and Web 3.0
- Rules in Semantic Web Technologies
- Rules in Web Intelligence Research
- Hybrid rule systems
- Management and maintenance of distributed rule bases and rule
repositories
   during their lifecycle
- Interchange and refactoring of rule bases in heterogeneous execution
   environments
- Verification and validation of interchanged rule bases in heterogeneous
   execution environments
- Contributions on effective, practical, and deployable Web standards on
   rules as well as special purpose, vertical domain rule languages
- Rule-based agility and its role in middleware
- Communication between rule based systems using interchange formats and
   processing / communication middleware
- Applications, products, research, and development in rule-based,
   distributed complex event processing, event communication and reaction
   rules
- Event-driven/action rule languages and models
- Rule-based Event Processing Languages and rule-based CEP
- Rule patterns and CEP patterns
- Practical solutions tackling the real-world Software Engineering
   requirements of rule-based systems in open, distributed environments
- Modeling of executable rule specifications and tool support
- Execution models, rule engines, and environments
- Compilation vs. interpretation approaches of rules
- Applications and integration of rules in web standards
- Rule-based software agents and (web) services
- Applications of rules in the Semantic Web and Pragmatic Web
- Comparing and advancing the state of current business rules engines and
   management system tools
- Rule interchange standards and related industry interchange formats
- Interoperation between different rule formats and ontological domain
   conceptualization
- Applications based on (Semantic) Web rule standardization or standards-
   proposing efforts
- Translation of interchangeable and domain-independent rule formats and
rule
   models into executable technical rule specifications
- Extraction and reengineering of platform-independent, interchangeable
rules
   and rule models from existing platform-specific resources
- Natural-language processing of rules
- Graphical processing, modelling and rendering of rules
- Incorporation of rule technology into distributed enterprise application
   architectures
- Rule-based policies and electronic contracts: their specification,
   execution and management
- Languages for exchanging and processing information through the web
- E-contracting and automated negotiations with rule-based declarative
   strategies
- Applications of rules in e.g. legal reasoning, compliance rules,
security,
   IT government, security, risk management, trust and proof reasoning,
etc.
- Rule-based (multi-valued) reasoning with and representing uncertain and
   fuzzy information
- Rule-based reasoning with non-monotonic negation, modalities, deontic,
   temporal, priority, scoped or other rule qualifications
- Rule-based default reasoning with default logic, defeasible logic, and
   answer set programming


RuleML-2008 Challenge
---------------------

The RuleML-2008 Challenge is one of the highlights of RuleML-2008. It
addresses the system demonstration for practical use of rule
technologies in
distributed and/or Web-based environments. The focus of the challenge
is on
rule technologies (including rule languages and engines),
interoperation and
interchange. The challenge offers participants the chance to demonstrate
their commercial and open source tools, use cases, and applications.
Prizes
will be awarded to the two best applications. All accepted demos will be
presented in a special Challenge Session.

A submission to RuleML challenge has to meet the requirement that
declarative
rules explicitly play a central role in the application. Basically this
means
that:
- Rules are explicitly represented in a declarative format and they are
   decoupled from the application (rather than being compiled or hard-coded
   into the application logic).
- Rules are used in interesting and practically relevant ways to, e.g.,
   derive useful information, transform knowledge, provide decision
support,
   provide automated rule-based monitoring, enforcement, validation or
   management of the behavioural logic of the application.

The demo should preferably (but not necessarily) be embedded into a
web-based
or distributed environment so that there will be a need for features
related
to the RuleML conference topics, as listed in the call for papers.

For more details please consult the RuleML-2008 Challenge website
(http://ruleml-challenge.cs.nccu.edu.tw).


Important Dates
---------------

- Paper Submissions due        June 2, 2008
- Notification of acceptance    July 18, 2008
- Final submissions due        August 9, 2008
- Symposium date        October 30-31, 2008
- RuleML Challenge        October 30, 2008


Symposium Proceedings and Submission Details
--------------------------------------------

Authors are invited to submit original contributions of practical
relevance
and technical rigor in the field, experience reports and show case / use
case
demonstrations of effective, practical, deployable rule-based
technologies or
applications in distributed environments. Papers must be in English
and may
be submitted at http://www.easychair.org/conferences/?conf=ruleml2008 as:

- Full Papers (15 pages in the proceedings)
- Short Papers (8 pages in the proceedings)
- RuleML-2008 Challenge Demo Paper + Show Cases (up to 8 pages in the
   proceedings)

Please upload all submissions as PDF files in LNCS format
(http://www.springer.de/comp/lncs/authors.html). To ensure high quality,
submitted papers will be carefully peer-reviewed by 3 PC members based on
originality, significance, technical soundness, and clarity of exposition.
Authors are requested to upload their complete papers by June 2, 2008. The
selected papers will be published in book form in the Springer Lecture
Notes
in Computer Science (LNCS) series along with a CD with demo software and
documents. The best paper from all submissions will be determined by
the PC
and a Best Paper Award will be handed over at the Symposium by a
Sponsor. All
submissions must be done electronically via
http://www.easychair.org/conferences/?conf=ruleml2008. We will pursue the
publication of a selection of revised papers to a special issue of a high-
quality journal.

Submissions to the RuleML Challenge 2008 consist of a demo paper of up
to 8
pages, describing the demo show case, and a link to more information about
the demo/show case, e.g. a project site, an online demonstration, a
presentation, or a download site. Demo papers should contain a substantial
presentation of the system to enable a proper evaluation of the techniques
used. The content of papers should be sufficiently substantial for
publication in the conference proceedings. The demo paper should be
submitted
at http://www.easychair.org/conferences/?conf=ruleml2008, while the demo
link
should be submitted at http://ruleml-challenge.cs.nccu.edu.tw/, and it
will be
immediately publicly available. If the link is password-protected, then
please submit a password for anonymous login from any Web browser,
giving us
the permission to pass the password on to 3 PC members. The submissions
should satisfy the minimal requirements defined in the topics of interest.
The demos will be evaluated during RuleML-2008 and prizes will be
awarded to
the first two best applications.

Please do not hesitate to email to the appropriate Symposium Chair(s),
if you
have any questions.


Organizing Committee
--------------------

General Chair

Adrian Paschke, Technical University Dresden, Germany
        adrian.paschke AT biotec.tu-dresden.de

Program Co-Chairs

Nick Bassiliades, Aristotle University of Thessaloniki, Greece
        nbassili AT csd.auth.gr
Guido Governatori, University of Queensland, Australia
        guido AT itee.uq.edu.au

Challenge Co-Chairs

Costin Badica, University of Craiova, Romania
        badica_costin AT software.ucv.ro
Yuh-Jong Hu, National Chengchi University, Taiwan
        hu AT cs.nccu.edu.tw

Panel Co-Chairs

John Hall, Model Systems, UK
        john.hall AT modelsys.com
Axel Polleres, DERI Galway, Ireland
        axel AT polleres.net

Liaison Co-Chairs

Mark Proctor, JBoss Rules, UK
        mproctor AT redhat.com
Rainer von Ammon, CITT GmbH, Germany
        vonammon AT t-online.de
Jan Vanthienen, Katholieke Universiteit Leuven, Belgium
        Jan.Vanthienen AT econ.kuleuven.be


Publicity Co-Chairs

Matthias Nickles, University of Bath, UK
        M.L.Nickles AT cs.bath.ac.uk
Tracy Bost, Valocity, USA
        tbost AT valocity.com
(Sponsoring levels: http://2008.ruleml.org/sponsoring/)

Web Chair

Suzanne Embury, University of Manchester, UK
        Suzanne.Embury AT manchester.ac.uk


Program Committee
--------------------

- Asaf Adi, IBM, Israel
- Grigoris Antoniou, University of Crete-FORTH, Greece
- Sidney Bailin, Knowledge Evolution, USA
- Matteo Baldoni, University of Torino, Italy
- Cristina Baroglio, University of Torino, Italy
- Claudio Bartolini, HP Labs
- Tim Bass, SilkRoad Inc.
- Bernhard Bauer, University of Augsburg, Germany
- Mikael Berndtsson, University of Skovde, Sweden
- Leopoldo Bertossi, Carleton University, Canada
- Pedro Bizarro, University of Coimbra, Portugal
- Peter Bollen, University of Maastricht, Netherlands
- Christian Brelage, SAP Research, Germany
- Donald Chapin, Business Semantics Ltd, UK
- Shyi-Ming Chen, National Taiwan University of Science and Technology,
   Taiwan
- Jorge Cuellar, Siemens, Germany
- Mike Dean, BBN Technologies, USA
- Stan Devitt, Agfa Healthcare, Ontario, Canada
- Jens Dietrich, Massey University, New Zeeland
- Jurgen Dix, Technische Universitaet Clausthal, Germany
- Schahram Dustdar, Vienna University of Technology, Austria
- Andreas Eberhart, fluid Operations, Germany
- Opher Etzion, IBM Research Laboratory Haifa, Israel
- Dieter Fensel, DERI Innsbruck, Austria
- Dragan Gasevic, Athabasca University, Canada
- Adrian Giurca, Brandenburg University of Technology at Cottbus, Germany
- Stijn Goedertier, Katholieke Universiteit Leuven, Belgium
- Robert Golan, DBmind Technologies, USA
- Christine Golbreich, University of Versailles, France
- Tom Gordon, Fraunhofer Institute for Open Communication Systems, Germany
- Marek Hatala, Simon Fraser University, Canada
- Ioannis Hatzilygeroudis, University of Patras, Greece
- Martin Hepp, University of Innsbruck, Austria
- Elisa Kendall, Sandpiper Software, USA
- Yiannis Kompatsiaris, Informatics and Telematics Institute, Greece
- Manolis Koubarakis, National and Kapodistrian University of Athens,
Greece
- Alex Kozlenkov, Betfair Ltd., UK
- Holger Lausen, DERI Innsbruck, Austria
- John Lee, Polytechnic University, Hong Kong
- Mark Linehan, IBM Research, USA
- Heiko Ludwig, IBM Watson Research, USA
- Mirko Malekovic, University of Zagreb, Croatia
- Christopher J. Matheus, Vistology, USA
- Craig McKenzie, Science Applications International, USA
- Jing Mei, IBM Research Lab China, China
- Zoran Milosevic, Deontik Inc, Australia
- Jang Minsu, E&T Research Institute, Korea
- Leora Morgenstern, Stanford, USA
- Gero Muehl, Technische Universitat Berlin, Germany
- Jorg Muller, TU Clausthal, Germany
- Chieko Nakabasami, Toyo University, Japan
- Ilkka Niemela, Helsinki University of Technology, Finland
- Bart Orriens, Tilburg University, Netherlands
- Jeff Pan, University of Aberdeen, UK
- Paula-Lavinia Patranjan, Skytec AG, Germany
- Jon Pellant, Pega Systems Inc., USA
- Jeff Pollock, Oracle, USA
- Alun Preece, Cardiff University, UK
- Maher Rahmouni, HP Labs
- Girish Ranganathan, University of New Brunswick, Canada
- Dave Reynolds, HP, England
- Graham Rong, MIT Sloan School of Management, USA
- Antonio Rotolo, CIRSFID, University of Bologna, Italy
- Norman Sadeh, Carnegie Mellon University, USA
- Christian de Sainte Marie, ILOG, France
- Marco Seirio, ruleCore, Sweden
- Timos Sellis, Institute for the Management of Information Systems and
   National Technical University of Athens, Greece
- Michael Sintek, DFKI, Germany
- Silvie Spreeuwenberg, LibRT, The Netherlands
- Giorgos Stamou, National Technical University of Athens, Greece
- Giorgos Stoilos, National Technical University of Athens, Greece
- Terrance Swift, XSB Inc., USA
- Kuldar Taveter, University of Melbourne, Australia
- James Taylor, Fair Isaac Corp., USA
- Vagan Terziyan, University of Jyvaskyla, Finland
- Paul Vincent, TIBCO Software, USA
- George Vouros, University of the Aegean, Greece
- Kewen Wang, Griffith University, Australia
- Mehmet Emre Yegen, Ygntec Inc., USA

--- In CEP-Interest@yahoogroups.com, "dcluckham" <pussyfatcat@...> wrote:
>
> Tim et al:
> CEP Forum appears to be up and running again:
>
> Webmeister: "I have got the forum up and running again, although I
> have removed a couple of custom features during the course of testing
> - Advanced RSS Syndication and Last Post Preview on the board index.
>
> I'll reapply the features tomorrow when I have adequate testing time"
>
>
> Any suggestions towards not repeating this situation during the course
> of upgrades to new versions of Wordpress are welcome!
> I am told some custom features, like the ones currently omitted
> (above) were incompatible with security.
>
> I think our regression testing needs to be tightened.
> - David
>
>
>
> --- In CEP-Interest@yahoogroups.com, "Tim Bass" <tim.silkroad@> wrote:
> >
> > Hi David,
> >
> > Yes, is certainly a permissions problem.  For example, if a regular
> > user logs in and clicks on the link to "show your own post"
> >
> > http://forum.complexevents.com/search.php?search_id=egosearch
> >
> > You get the message:
> >
> > Sorry but you are not permitted to use the search system.
> >
> > To reproduce the error, have your webguy register a normal new
user id
> > and the login under that user id.
> >
> > Yours faithfully, Tim
> >
> > ... and Stay tuned for the never ending "blogosphere POSETS event
cloud
> > debate" to resume on David's board when it is working again, LOL
> >
>

Tim et al:
CEP Forum appears to be up and running again:

Webmeister: "I have got the forum up and running again, although I
have removed a couple of custom features during the course of testing
- Advanced RSS Syndication and Last Post Preview on the board index.

I'll reapply the features tomorrow when I have adequate testing time"


Any suggestions towards not repeating this situation during the course
of upgrades to new versions of Wordpress are welcome!
I am told some custom features, like the ones currently omitted
(above) were incompatible with security.

I think our regression testing needs to be tightened.
- David



--- In CEP-Interest@yahoogroups.com, "Tim Bass" <tim.silkroad@...> wrote:
>
> Hi David,
>
> Yes, is certainly a permissions problem.  For example, if a regular
> user logs in and clicks on the link to "show your own post"
>
> http://forum.complexevents.com/search.php?search_id=egosearch
>
> You get the message:
>
> Sorry but you are not permitted to use the search system.
>
> To reproduce the error, have your webguy register a normal new user id
> and the login under that user id.
>
> Yours faithfully, Tim
>
> ... and Stay tuned for the never ending "blogosphere POSETS event cloud
> debate" to resume on David's board when it is working again, LOL
>

Hi David,

Yes, is certainly a permissions problem.  For example, if a regular
user logs in and clicks on the link to "show your own post"

http://forum.complexevents.com/search.php?search_id=egosearch

You get the message:

Sorry but you are not permitted to use the search system.

To reproduce the error, have your webguy register a normal new user id
and the login under that user id.

Yours faithfully, Tim

... and Stay tuned for the never ending "blogosphere POSETS event cloud
debate" to resume on David's board when it is working again, LOL

It looks like a permissions problem. If the administrator and guest
accounts are granted access to the boards, then he's not be able to
reproduce the problem because he has the required permission. It might
just be a matter of granting the "normal user" group access to the boards.

--- In CEP-Interest@yahoogroups.com, "dcluckham" <pussyfatcat@...> wrote:
>
> The CEP forum has been down for 48 hours.
> There have been numerous complaints.
> If you login you can't see the Boards.
>
> The webmeister says he cannot reproduce this error???????
> Probably due to an update from Wordpress 3.0 to wordpress 3.1!!!
> Bear with us!!!
>
> thanks
> - David
>
> --- In CEP-Interest@yahoogroups.com, "hansgilde" <hgilde@> wrote:
> >
> > Same issue here.
> >
> > --- In CEP-Interest@yahoogroups.com, "Szabolcs Rozsnyai"
> > <s.rozsnyai@> wrote:
> > >
> > > Can anyone confirm that it is currently not possible to post
> > anything on the
> > > CEP Board:
> > >
> > > forum.complexevents.com
> > >
> > > If I log in I receive the message:
> > >
> > > *"This board has no forums."
> > >
> > > Just to be sure that this is not an issue of my account I create a
> > new one
> > > and the same thing happens.
> > >
> > > Regards,
> > >
> > > Szabolcs
> > > *
> > >
> >
>

The CEP forum has been down for 48 hours.
There have been numerous complaints.
If you login you can't see the Boards.

The webmeister says he cannot reproduce this error???????
Probably due to an update from Wordpress 3.0 to wordpress 3.1!!!
Bear with us!!!

thanks
- David

--- In CEP-Interest@yahoogroups.com, "hansgilde" <hgilde@...> wrote:
>
> Same issue here.
>
> --- In CEP-Interest@yahoogroups.com, "Szabolcs Rozsnyai"
> <s.rozsnyai@> wrote:
> >
> > Can anyone confirm that it is currently not possible to post
> anything on the
> > CEP Board:
> >
> > forum.complexevents.com
> >
> > If I log in I receive the message:
> >
> > *"This board has no forums."
> >
> > Just to be sure that this is not an issue of my account I create a
> new one
> > and the same thing happens.
> >
> > Regards,
> >
> > Szabolcs
> > *
> >
>

Same issue here.

--- In CEP-Interest@yahoogroups.com, "Szabolcs Rozsnyai"
<s.rozsnyai@...> wrote:
>
> Can anyone confirm that it is currently not possible to post
anything on the
> CEP Board:
>
> forum.complexevents.com
>
> If I log in I receive the message:
>
> *"This board has no forums."
>
> Just to be sure that this is not an issue of my account I create a
new one
> and the same thing happens.
>
> Regards,
>
> Szabolcs
> *
>

Rainer,
Opher just said it all!

- David

ps
see also the Forum on what's missing from engines.

--- In CEP-Interest@yahoogroups.com, "rainer93138" <vonammon@...> wrote:
>
> I suspect that there is still a misunderstanding of the term CEP in
> the last blog entry of John, perhaps hanging around a bit jet lagged
> in Hongkong:-)
>
> http://apama.typepad.com/my_weblog/
>
> @David Luckham: Do you want to answer once more?
> I'm sure to remember that John Bates has attended the 2nd CEP
> symposion at Redwood Shores/Oracle hosted where once again we decided
> to keep the term CEP. Not Event Processing is Complex, but we talk
> about Complex Events, aggregated/correlated from basic low level
> events to complex business events on which we can act anything. Great
> CEP platforms like Apama, Coral8, BEA, Esper, Tibco, IBM, Oracle,
> StreamBase etc. can do this job very easily on the base of their easy-
> to-use EPL and with high scalability/low latency by processing
> hundred thousands of low level events per second.
>
> This is our main job at present to show and explain it potential
> customers. What is not so easy is to explain what "Predictive
> Business" (TM from Tibco) means and how to find the right event
> patterns and strategies. This is not a complex technical problem, but
> rather a business logical
>
> Just my 2 cents...
>
> --Rainer
>

Rainer - you are right, I have put a comment about it in the original Blog
- before seeing your comment (see:
http://apama.typepad.com/my_weblog/2008/04/on-complex-even.html?cid=112010062#co\
mment-112010062)
  -  CEP is not about complex processing but about complex events.

BTW - the person who signed the Apama Blog that we are talking about is
Richard Bentley and not John Bates









Cheers,
Opher
--------------------------------------------------------------------------------\
--------------------------------------------------------
Dr. Opher Etzion
Senior Technical Staff Member,   Master Inventor
Event Processing Scientific Leader
IBM Research Lab in Haifa
Phone: +972-4-829-6230; Cell: +972-54-790-2086; Fax: +972-4-829-6116
Homepage:
http://domino.research.ibm.com/comm/research_people.nsf/pages/opher.index.html
Blog: http://epthinking.blogspot.com/
--------------------------------------------------------------------------------\
-------------------------------------------------------



              "rainer93138"
              <vonammon@t-onlin
              e.de>                                                      To
              Sent by:                  CEP-Interest@yahoogroups.com
              CEP-Interest@yaho                                          cc
              ogroups.com
                                                                    Subject
                                        [CEP-Interest] Misunderstanding
              24/04/2008 15:21          about term definition of Complex
                                        Event Processing at FeedBlitz

              Please respond to
              CEP-Interest@yaho
                 ogroups.com






I suspect that there is still a misunderstanding of the term CEP in
the last blog entry of John, perhaps hanging around a bit jet lagged
in Hongkong:-)

http://apama.typepad.com/my_weblog/

@David Luckham: Do you want to answer once more?
I'm sure to remember that John Bates has attended the 2nd CEP
symposion at Redwood Shores/Oracle hosted where once again we decided
to keep the term CEP. Not Event Processing is Complex, but we talk
about Complex Events, aggregated/correlated from basic low level
events to complex business events on which we can act anything. Great
CEP platforms like Apama, Coral8, BEA, Esper, Tibco, IBM, Oracle,
StreamBase etc. can do this job very easily on the base of their easy-
to-use EPL and with high scalability/low latency by processing
hundred thousands of low level events per second.

This is our main job at present to show and explain it potential
customers. What is not so easy is to explain what "Predictive
Business" (TM from Tibco) means and how to find the right event
patterns and strategies. This is not a complex technical problem, but
rather a business logical

Just my 2 cents...

--Rainer


------------------------------------

Yahoo! Groups Links

I suspect that there is still a misunderstanding of the term CEP in
the last blog entry of John, perhaps hanging around a bit jet lagged
in Hongkong:-)

http://apama.typepad.com/my_weblog/

@David Luckham: Do you want to answer once more?
I'm sure to remember that John Bates has attended the 2nd CEP
symposion at Redwood Shores/Oracle hosted where once again we decided
to keep the term CEP. Not Event Processing is Complex, but we talk
about Complex Events, aggregated/correlated from basic low level
events to complex business events on which we can act anything. Great
CEP platforms like Apama, Coral8, BEA, Esper, Tibco, IBM, Oracle,
StreamBase etc. can do this job very easily on the base of their easy-
to-use EPL and with high scalability/low latency by processing
hundred thousands of low level events per second.

This is our main job at present to show and explain it potential
customers. What is not so easy is to explain what "Predictive
Business" (TM from Tibco) means and how to find the right event
patterns and strategies. This is not a complex technical problem, but
rather a business logical

Just my 2 cents...

--Rainer

--- In CEP-Interest@yahoogroups.com, don-cep-list@... wrote:
>  > 1) Pattern Matching Basics
>  > negative events (A followed by NOT B within X seconds)
> I think you probably mean "A not followed by B".
> Do you consider "not B" and event?
> I don't, because if I did then I would view it as happening
infinitely
> often.
> Also, I notice that all of your examples include "within X seconds".
> Perhaps you don't mind just supplying a large number for X, but my
> view is that the "within" clause should be optional.
I guess you're right about the "not B" point.  I translated my code
directly into english...sorry :)  The CEP engines that I've studied
all use some (A 'followed by' !B) syntax instead of (A !'followed by'
B).  The same goes for "within".  I think unbounded times for pattern
matches would be great.

>  > 2) Event style (non-Cartesian product) Patterns
> This seems a good example of the problems/questions above.
> Surely you agree that it should be POSSIBLE to specify cartesian
> products?  It should also be possible to specify
>  A followed by B without any other B-like event (C) occuring between
> i.e., the first B following A.  For that matter it should be
possible
> to specify the SECOND B following A.
> I think you are mistaken in identifying "A then B" with A causing B.
> If you want cause then you'll have to describe how it can be
determined.
> Timing is surely not adequate to define causality (if you believe in
> causality at all).
If they want to allow you to get a Cartesian product, that's fine,
but it shouldn't be the only behavior of the pattern matcher.  My
point with this is that for the use cases I've discovered events have
a one-to-one relationship (an "trade" event is associated with
one "new order" event).  The developer should be able to reflect this
reality in the event processor.

>
>  > 3) Query Order
> I didn't realize this was a problem.
In some CEP engines you basically declare a set of actions that are
taken when an event "happens" and all of the scheduling work is
handed over the the engine itself.  In cases like this, the developer
should be given some tool to explicetly specify the order of the
actions.  Unfortunately these capabilities aren't always provided
natively.

>  > 5) Startup/shutdown events
> In general it is not possible to react to shutdown events, e.g.,
> the power goes off so the machine goes down so the program stops
running.
Sure, if you pull the plug on a server, I don't expect the CEP engine
to gracefully shut down.  Basically I want something like
Java's "finally" block that allows me to easily do some cleanup when
the event processing application is halted.

>  > 8) Seamless integration with external databases
> It's not quite clear what this means, but even in the most
restricted
> meaning (from your example of reading a date) this seems an
unreasonably
> tall order, since there are lots of different DBs out there that
> differ in small but important details.  Why is it the job of the CEP
> vendor to know all about every DB rather than the other way around?
My point was that CEP engines and databases are very ofthen used
closely together.  I think it's reasonable to say that every company
that's looking into CEP already has a database.  It's very
frustrating to bring in a new CEP engine that has all this great
potential only to have to struggle to get it to talk to a DB.  CEP
vendors will only benefit from making this integration more seamless.

We've been working a lot with CEP and I've been keeping a list of
things that were too hard to accomplish - just standard use cases that
are part of any company's workflow.  We mostly discovered limitations
by implementing our business use cases in an event processor.
Eventually the "this is too hard" list was distiled into a list of
features that would address the majority of the issues that we
encountered.  After sitting on the feature list for a few months, it
didn't really make sense to keep it private anymore.  I really believe
that these features will make event processing systems much more useful
to an enterprise.

--- In CEP-Interest@yahoogroups.com, John Stoneham <lyric@...> wrote:
>
> > Below is a list of 9 features that I consider to be requirements for
> > a CEP engine to be maximally useful to an enterprise. This list
> > assumes that the CEP engine supports all of the "basics" like
sliding
> > and jumping windows, filtering, aggregation, etc.
>
> Tom,
>
> Fantastic list. I'll be keeping this for reference.
>
> May I ask what led you to create and post this list?
>
> - John
>

> Below is a list of 9 features that I consider to be requirements for
> a CEP engine to be maximally useful to an enterprise. This list
> assumes that the CEP engine supports all of the "basics" like sliding
> and jumping windows, filtering, aggregation, etc.

Tom,

Fantastic list. I'll be keeping this for reference.

May I ask what led you to create and post this list?

- John

Mostly my reactions below are small details about what you should be
asking for.  Other than those I'd be somewhat surprised to hear that
many of the features you list are missing from many products.  But
since I don't use many products, perhaps you could tell me which ones
are missing (or present) in which products.

tom.puzak writes:

  > In my mind, the greatest benefit of CEP is that it reverses the "80-
  > 20" rule
What's this 80-20 rule?
If it's something like 80% of the effort goes into 20% of the value then
- it's hard to see how to measure value
- my experience suggests that neither CEP nor any other software will
   change that

  > 1) Pattern Matching Basics
  > negative events (A followed by NOT B within X seconds)
I think you probably mean "A not followed by B".
Do you consider "not B" and event?
I don't, because if I did then I would view it as happening infinitely
often.
Also, I notice that all of your examples include "within X seconds".
Perhaps you don't mind just supplying a large number for X, but my
view is that the "within" clause should be optional.
In general I agree that it should be possible to specify a very large
set of conditions.  However, it is not possible to specify a very
large number of conditions and express them all in a very simple way.
Which conditions do you think should NOT be possible to specify?
Which conditions do you think must be specified "simply" and which not?

  > 2) Event style (non-Cartesian product) Patterns
This seems a good example of the problems/questions above.
Surely you agree that it should be POSSIBLE to specify cartesian
products?  It should also be possible to specify
  A followed by B without any other B-like event (C) occuring between
i.e., the first B following A.  For that matter it should be possible
to specify the SECOND B following A.
I think you are mistaken in identifying "A then B" with A causing B.
If you want cause then you'll have to describe how it can be determined.
Timing is surely not adequate to define causality (if you believe in
causality at all).

  > 3) Query Order
I didn't realize this was a problem.

  > 4) Absolute time
Isn't absolute time available everywhere?
Perhaps it's just that "clock events" (the clock now reads 4:00) are
not supported everywhere?

  > 5) Startup/shutdown events
In general it is not possible to react to shutdown events, e.g.,
the power goes off so the machine goes down so the program stops running.
I assume all software supports startup as a special case.  That seems
appropriate to me.

  > 6) Full state management capabilities

  > 7) Backtesting Capabilities

  > 8) Seamless integration with external databases
It's not quite clear what this means, but even in the most restricted
meaning (from your example of reading a date) this seems an unreasonably
tall order, since there are lots of different DBs out there that
differ in small but important details.  Why is it the job of the CEP
vendor to know all about every DB rather than the other way around?

  > 9) A nice development environment
  > As far as IDEs go, I'd say that 99% of IDE using developers (sorry
  > emacs lovers) are covered by three platforms...
(I'd much prefer a normal lisp environment to any of those listed.
But then, I'm an emacs lover.)

====
I see I'm not the first to reply.
David writes
  I suggest you post this on the CEP forum http://forum.complexevents.com/
If you do so, feel free to put my reply there as well.

Tom,
I suggest you post this on the CEP forum http://forum.complexevents.com/
under the discussion section.
You'll need to join in order to post, but that's a formality.
- David

--- In CEP-Interest@yahoogroups.com, "tom.puzak" <tom.puzak@...> wrote:
>
> Below is a list of 9 features that I consider to be requirements for
> a CEP engine to be maximally useful to an enterprise.  This list
> assumes that the CEP engine supports all of the "basics" like sliding
> and jumping windows, filtering, aggregation, etc.
>
> It's obvious that event processing vendors have spent a lot of time
> and effort making their systems fast, scalable, etc, but sometimes I
> can help but feel like several very natural use cases have been
> overlooked to the overall detriment of the available products.
>
> In my mind, the greatest benefit of CEP is that it reverses the "80-
> 20" rule � with an event processor you get all of the plumbing that
> allows you to process events thereby allowing the developer to focus
> most of his effort implementing business logic.  However every CEP
> system that I have encountered makes the developer jump through hoops
> to accomplish conceptually simple tasks that go hand in hand
> with "standard" event processing use cases.  I hope that this list of
> features will start a discussion about "real world" applications of
> CEP and how existing event processors can incorporate this
> functionality into their frameworks.  I truly believe that these
> features are required for CEP applications to fully live up to their
> potential.
>
> 1) Pattern Matching Basics
> A CEP engine must allow the user to simply and concisely represent a
> pattern of events that occur within a specified time period.  The
> pattern matcher should support Boolean matches (A and/or B within X
> seconds), sequence relations (A followed by B within X seconds),
> negative events (A followed by NOT B within X seconds), and sub-
> patterns (A followed by (B followed by C within X seconds) within Y
> seconds).  The pattern matching operator must also allow the
> developer to specify a predicate (guard) on the events in the
> pattern.  Of course, the developer must be able to apply the
> predicate to a negative event in the pattern (A followed by NOT B
> when B.p < 50 within X seconds).  Additionally, the developer must be
> able to write a predicate that restricts a successful evaluation of
> the pattern match based on the fields contained within the events
> themselves (A followed by B when A.p==B.q within X seconds).
>
> 2) Event style (non-Cartesian product) Patterns
> Some subset of the capabilities listed in #1 is provided by the
> majority of the CEP systems that I have encountered.  However, I
> think it is important that we actually nail down the exact meaning of
> a pattern match.  Most vendors' pattern matching implementations seem
> to be based on incremental joins of the fields in the events.
> Consequently it is quite common for these systems to produce a
> Cartesian product when matching a pattern.  For example, given the
> pattern (A followed by B within 60 seconds), assume that we receive
> the following events (taking the numbers as subscripts) within a 60
> second window:
>
> A1 A2 B1 A3 B2 B3
>
> A standard pattern matching implementation would produce matches on
> the following events:
>
> (A1, B1) (A1,B2) (A1,B3) (A2, B1) (A2, B2) (A2, B3) (A3, B2) (A3, B3)
>
> While this implementation may be academically correct, I can't see
> how this is useful from an event processing perspective.  Consider
> the output of the pattern above.  How does a downstream application
> (the one consuming the output of the pattern matcher) know the cause
> of the events?  In the real world event B2 may have been caused by
> event A1, but you'd never be able to know that by looking at the
> output from a pattern matcher that is implemented to produce a
> Cartesian product, because the CEP engine is saying that B2 was also
> caused by A1 and by A3.
>
> In my opinion, a fully functional pattern matcher must provide the
> capability to produce pattern matches that are based on the order
> that the events were created.  In this case, with the sample input
> above, a pattern matcher of this type would produce matches on ONLY
> these events:
>
> (A1, B1) (A2, B2) (A3, B3)
>
> With this behavior there is no ambiguity about which `A' event caused
> the related `B' event.  For all of the use cases I've encountered
> this is far more useful than the Cartesian product style of pattern
> matching because domain specific events usually have a one-to-one
> relationship.
>
> Telling the developer to look for the pattern (A followed by NOT B
> followed by B within X seconds) is not good enough.  Code like that
> is obtuse, and there is absolutely no chance that someone who's not
> intimately familiar with the EPL could read a section of code like
> that and understand what it's supposed to do.
>
> Consider a concrete example.  If American Airlines stock goes up and
> Delta's stock goes up within 10 seconds then say we want to buy 100
> shares of Jet Blue.  If we see AA go up and then Delta goes up 3
> times within 10 seconds, our algorithm should buy 300 shares.  If the
> pattern matcher is doing a Cartesian product we'll buy 600 shares.
> That's not good.
>
> For academic completeness it's probably okay to implement a pattern
> matcher to produce a Cartesian product, but support for the "match
> the first" style of pattern matching is an absolute requirement.
>
> 3) Query Order
> Oftentimes it's necessary to perform multiple actions on the
> reception of a single event.  For example, when an event is received
> we may want to query an in memory table and then update a record in
> the same table.  Because it's probably important for these actions to
> occur in a particular order, the event processing engine must provide
> some mechanism for the developer to specify the order that the
> actions occur when an event is received.  Without this capability,
> the developer needs to make complicated event routing logic that adds
> a lot of complexity the code to accomplish a conceptually simple task.
>
> 4) Absolute time
> Like it or not, batch processing is still a valuable tool that is
> applicable to many tasks within an organization.  A fully featured
> event processor should support the capability to kick of "batch"
> style jobs at a particular time of day.  For example, imagine that a
> CEP engine used for algorithmic trading is storing all of the day's
> stock tics in memory.  At 4:00PM the market closes, so at 4:30 we
> might want to dump all of the data stored in the event processor's
> cache to an external database and clear the cache so we're ready to
> roll the next morning.  Any way you slice it, this is a batch job.
> For maximum flexibility the event processor must provide the
> capability to perform tasks at a specific time of day.
>
> If the event processor doesn't natively support this functionality,
> the developer will probably be stuck creating a cron job that
> executes a shell script to send an event to the CEP engine at 4:30
> and kick off the archiving job.  This is less than desirable because
> the functionality is split over two completely separate entities (the
> CEP engine, and a shell script running on cron).  Maintainability
> will undoubtedly suffer.
>
> 5) Startup/shutdown events
> As another example of batch processing that should be integrated with
> event processing engines, there are certain tasks that always need to
> be performed whenever an event processing application is
> started/stopped.  For example, on startup a particular application
> may need to initialize itself with some standing data from an
> external database.  Likewise, it might be necessary to write all of
> the state in the event processor to a database or file on shutdown.
> For maximum flexibility the CEP engine should provide some
> capabilities to perform tasks at startup and shutdown natively.  This
> could be as simple as sending "startup" and "shutdown" events
> whenever an application is started/stopped by default.  The developer
> could choose to process those events if necessary.
>
> 6) Full state management capabilities
> Real world event processing applications need to correlate events
> with more static data being cached by the CEP system.  A CEP engine
> must provide the capabilities to effectively maintain this cache and
> to relate received events to the data stored in the cache.  Cache
> management needs to be simple and straightforward.  If the CEP
> engine's event processing language (EPL) is implemented as a superset
> of SQL (as most EPLs are) then the EPL needs to support all of the
> standard operations supported by standard DB SQL.  This means INSERT,
> UPDATE, and DELETE.
>
> 7) Backtesting Capabilities
> Because CEP is about dealing with transient events, it can be
> difficult to determine how a particular application would have
> performed in the past after changes are made.  Therefore, the CEP
> engine must provide some capability to "record" the events that it
> receives for future playback.  Playback must preserve the relative
> orderings of the original events.   For effective testing, the
> developer needs to be able to play back the stored events at an
> increased rate.  The ability to start or stop recording events should
> be a switch that can be thrown at runtime.  Of course the performance
> impact of recording all the incoming events needs to be minimized.
>
> Lots of CEP engines out there provide some way for the developer to
> perform this task manually (attach an output adapter to a file, and
> when you want to do testing swap your middleware's input adapter with
> the "read a file" input adapter) but that's a pain and it's
> inflexible.  Backtesting can be a major sticking point with CEP's
> adoption, so a system that natively implements this capability will
> have already cleared a major hurdle.
>
> 8) Seamless integration with external databases
> It's common for a CEP engine to interact with an external DB.  This
> interaction must be natural and very easy.  The data types used
> internally by the CEP engine should map exactly into the external
> database's types without casting (the developer shouldn't ever have
> to write code like
> 'to_date(<a date from the CEP engine cast as a string>, "MM/DD/YYYY
> HH:MI:SS.FF")'
> just to record a date in an Oracle DB).  The same is true of reading
> data from a database.  The data types of the external table's columns
> must be preserved and seamlessly mapped into the CEP engine's data
> types without casting or translation.
>
> 9) A nice development environment
> As far as IDEs go, I'd say that 99% of IDE using developers (sorry
> emacs lovers) are covered by three platforms.  For everything non-
> Java the clear choice is Microsoft's Visual Studio.  The Java
> developers are split between Eclipse and IntelliJ IDEA.  Most
> developers who find themselves writing code for a CEP platform are
> already very familiar with the aforementioned IDEs.  Integration,
> adoption, and development will be easier if the developer can work on
> CEP applications in an environment that he is already comfortable
> using.
>
>
> Thanks,
> Tom
>
> Tom Puzak
> Strategic Research
> Liquidnet Holdings, Inc.
>

Below is a list of 9 features that I consider to be requirements for
a CEP engine to be maximally useful to an enterprise.  This list
assumes that the CEP engine supports all of the "basics" like sliding
and jumping windows, filtering, aggregation, etc.

It's obvious that event processing vendors have spent a lot of time
and effort making their systems fast, scalable, etc, but sometimes I
can help but feel like several very natural use cases have been
overlooked to the overall detriment of the available products.

In my mind, the greatest benefit of CEP is that it reverses the "80-
20" rule � with an event processor you get all of the plumbing that
allows you to process events thereby allowing the developer to focus
most of his effort implementing business logic.  However every CEP
system that I have encountered makes the developer jump through hoops
to accomplish conceptually simple tasks that go hand in hand
with "standard" event processing use cases.  I hope that this list of
features will start a discussion about "real world" applications of
CEP and how existing event processors can incorporate this
functionality into their frameworks.  I truly believe that these
features are required for CEP applications to fully live up to their
potential.

1) Pattern Matching Basics
A CEP engine must allow the user to simply and concisely represent a
pattern of events that occur within a specified time period.  The
pattern matcher should support Boolean matches (A and/or B within X
seconds), sequence relations (A followed by B within X seconds),
negative events (A followed by NOT B within X seconds), and sub-
patterns (A followed by (B followed by C within X seconds) within Y
seconds).  The pattern matching operator must also allow the
developer to specify a predicate (guard) on the events in the
pattern.  Of course, the developer must be able to apply the
predicate to a negative event in the pattern (A followed by NOT B
when B.p < 50 within X seconds).  Additionally, the developer must be
able to write a predicate that restricts a successful evaluation of
the pattern match based on the fields contained within the events
themselves (A followed by B when A.p==B.q within X seconds).

2) Event style (non-Cartesian product) Patterns
Some subset of the capabilities listed in #1 is provided by the
majority of the CEP systems that I have encountered.  However, I
think it is important that we actually nail down the exact meaning of
a pattern match.  Most vendors' pattern matching implementations seem
to be based on incremental joins of the fields in the events.
Consequently it is quite common for these systems to produce a
Cartesian product when matching a pattern.  For example, given the
pattern (A followed by B within 60 seconds), assume that we receive
the following events (taking the numbers as subscripts) within a 60
second window:

A1 A2 B1 A3 B2 B3

A standard pattern matching implementation would produce matches on
the following events:

(A1, B1) (A1,B2) (A1,B3) (A2, B1) (A2, B2) (A2, B3) (A3, B2) (A3, B3)

While this implementation may be academically correct, I can't see
how this is useful from an event processing perspective.  Consider
the output of the pattern above.  How does a downstream application
(the one consuming the output of the pattern matcher) know the cause
of the events?  In the real world event B2 may have been caused by
event A1, but you'd never be able to know that by looking at the
output from a pattern matcher that is implemented to produce a
Cartesian product, because the CEP engine is saying that B2 was also
caused by A1 and by A3.

In my opinion, a fully functional pattern matcher must provide the
capability to produce pattern matches that are based on the order
that the events were created.  In this case, with the sample input
above, a pattern matcher of this type would produce matches on ONLY
these events:

(A1, B1) (A2, B2) (A3, B3)

With this behavior there is no ambiguity about which `A' event caused
the related `B' event.  For all of the use cases I've encountered
this is far more useful than the Cartesian product style of pattern
matching because domain specific events usually have a one-to-one
relationship.

Telling the developer to look for the pattern (A followed by NOT B
followed by B within X seconds) is not good enough.  Code like that
is obtuse, and there is absolutely no chance that someone who's not
intimately familiar with the EPL could read a section of code like
that and understand what it's supposed to do.

Consider a concrete example.  If American Airlines stock goes up and
Delta's stock goes up within 10 seconds then say we want to buy 100
shares of Jet Blue.  If we see AA go up and then Delta goes up 3
times within 10 seconds, our algorithm should buy 300 shares.  If the
pattern matcher is doing a Cartesian product we'll buy 600 shares.
That's not good.

For academic completeness it's probably okay to implement a pattern
matcher to produce a Cartesian product, but support for the "match
the first" style of pattern matching is an absolute requirement.

3) Query Order
Oftentimes it's necessary to perform multiple actions on the
reception of a single event.  For example, when an event is received
we may want to query an in memory table and then update a record in
the same table.  Because it's probably important for these actions to
occur in a particular order, the event processing engine must provide
some mechanism for the developer to specify the order that the
actions occur when an event is received.  Without this capability,
the developer needs to make complicated event routing logic that adds
a lot of complexity the code to accomplish a conceptually simple task.

4) Absolute time
Like it or not, batch processing is still a valuable tool that is
applicable to many tasks within an organization.  A fully featured
event processor should support the capability to kick of "batch"
style jobs at a particular time of day.  For example, imagine that a
CEP engine used for algorithmic trading is storing all of the day's
stock tics in memory.  At 4:00PM the market closes, so at 4:30 we
might want to dump all of the data stored in the event processor's
cache to an external database and clear the cache so we're ready to
roll the next morning.  Any way you slice it, this is a batch job.
For maximum flexibility the event processor must provide the
capability to perform tasks at a specific time of day.

If the event processor doesn't natively support this functionality,
the developer will probably be stuck creating a cron job that
executes a shell script to send an event to the CEP engine at 4:30
and kick off the archiving job.  This is less than desirable because
the functionality is split over two completely separate entities (the
CEP engine, and a shell script running on cron).  Maintainability
will undoubtedly suffer.

5) Startup/shutdown events
As another example of batch processing that should be integrated with
event processing engines, there are certain tasks that always need to
be performed whenever an event processing application is
started/stopped.  For example, on startup a particular application
may need to initialize itself with some standing data from an
external database.  Likewise, it might be necessary to write all of
the state in the event processor to a database or file on shutdown.
For maximum flexibility the CEP engine should provide some
capabilities to perform tasks at startup and shutdown natively.  This
could be as simple as sending "startup" and "shutdown" events
whenever an application is started/stopped by default.  The developer
could choose to process those events if necessary.

6) Full state management capabilities
Real world event processing applications need to correlate events
with more static data being cached by the CEP system.  A CEP engine
must provide the capabilities to effectively maintain this cache and
to relate received events to the data stored in the cache.  Cache
management needs to be simple and straightforward.  If the CEP
engine's event processing language (EPL) is implemented as a superset
of SQL (as most EPLs are) then the EPL needs to support all of the
standard operations supported by standard DB SQL.  This means INSERT,
UPDATE, and DELETE.

7) Backtesting Capabilities
Because CEP is about dealing with transient events, it can be
difficult to determine how a particular application would have
performed in the past after changes are made.  Therefore, the CEP
engine must provide some capability to "record" the events that it
receives for future playback.  Playback must preserve the relative
orderings of the original events.   For effective testing, the
developer needs to be able to play back the stored events at an
increased rate.  The ability to start or stop recording events should
be a switch that can be thrown at runtime.  Of course the performance
impact of recording all the incoming events needs to be minimized.

Lots of CEP engines out there provide some way for the developer to
perform this task manually (attach an output adapter to a file, and
when you want to do testing swap your middleware's input adapter with
the "read a file" input adapter) but that's a pain and it's
inflexible.  Backtesting can be a major sticking point with CEP's
adoption, so a system that natively implements this capability will
have already cleared a major hurdle.

8) Seamless integration with external databases
It's common for a CEP engine to interact with an external DB.  This
interaction must be natural and very easy.  The data types used
internally by the CEP engine should map exactly into the external
database's types without casting (the developer shouldn't ever have
to write code like
'to_date(<a date from the CEP engine cast as a string>, "MM/DD/YYYY
HH:MI:SS.FF")'
just to record a date in an Oracle DB).  The same is true of reading
data from a database.  The data types of the external table's columns
must be preserved and seamlessly mapped into the CEP engine's data
types without casting or translation.

9) A nice development environment
As far as IDEs go, I'd say that 99% of IDE using developers (sorry
emacs lovers) are covered by three platforms.  For everything non-
Java the clear choice is Microsoft's Visual Studio.  The Java
developers are split between Eclipse and IntelliJ IDEA.  Most
developers who find themselves writing code for a CEP platform are
already very familiar with the aforementioned IDEs.  Integration,
adoption, and development will be easier if the developer can work on
CEP applications in an environment that he is already comfortable
using.


Thanks,
Tom

Tom Puzak
Strategic Research
Liquidnet Holdings, Inc.

RuleCore is today extending its OEM/ISV license offering with end-user access to
ruleCore.

We do this by providing a subscription based service where we host a ruleCore
CEP Server
for the user.

Priced at 0.0001 euro/event, complex event processing is no longer expensive or
complicated to get started with.

As the ruleCore CEP Server is designed to be a smart building block of an Event
Driven
Architecture, everything is accomplished using events. The ruleCore service
provides
access to the user's private ruleCore CEP Server using JMS.

Inbound events are used to manage rules, monitor the state of the server and
deliver
your business events into ruleCore. Outbound events are delivered with reaction
events
with notifications about detected situations and description of the live state
of the
rule evaluation.

And as all events are just plain XML, every tool and system which can generate
and parse
XML can feed ruleCore with events and listen on reaction events. A typical use
would be
to connect ruleCore to a message broker or ESB.

/Marco

Call for EPTS Founding Members


EPTS - the Event Processing Technical Society - is going to be
officially launched in late May or early June (exact date � TBD)

The EPTS goals are to:

Increase awareness (promote mindshare) about event processing,
including understanding what its value is, producing a common
glossary, and highlighting the current state of best practices,
Collaborate with various standards organizations making sure all
standard efforts are in sync (EPTS will not become a standards
organization)
Help establish Event Processing as an academic discipline

EPTS has existed as an informal group for the last two years,
holding three Event Processing Symposiums to date; for details see:
http://complexevents.com/?page_id=87,
http://www.complexevents.com/?page_id=129, and
http://complexevents.com/?p=277

We invite participation of all vendors, individual participants
(academic people, independent consultants, and employees of non-
member organizations), and customers who would like to contribute to
the evolution of the event processing area.  There is no
participation fee, but all members (organizational and individual
members) must sign the EPTS Members Agreement.

EPTS activities will be performed through workgroups, and periodic
general meetings.  Examples of current workgroup tasks include
glossary creation, use cases analysis, event processing meta-
modeling, the study of event formats, and event processing academic
education.

Benefits to members include the ability to identify new requirements
or topics for discussion, lead or participate in workgroups and
other community activities, being listed and potentially quoted in
official announcements and press releases, and the ability to be
elected to the EPTS Steering Committee,

Please respond immediately by email to Opher Etzion:
opher@...  if you are interested in becoming a founding
member.
If you are able to sign the membership agreement by May 9th, 2008,
you'll have the opportunity to appear in the list of founding
members in the launch press release.

In order to join, please take the following actions:

1. Print the EPTS Member Agreement document from URL:
http://www.ep-
ts.com/component/option,com_docman/task,doc_download/gid,54/Itemid,84
/ - please print it as is, without any change,
2.   Sign (for individual member) or have an authorized
representative of your company sign (for a company member) two
hardcopies of the agreement.   Keep one for your files.   You will
need to produce this copy if the EPTS Steering Committee ever makes
a request for it at some point in the future.
3 Fax one copy to:

        Attention:          Mike Kaiser
        Fax #:                1-845-489-9958

4. Courier the second hard copy signature to the following
address for EPTS archival purposes:

         IBM Corporation
         Attn: Mike Kaiser (JGWA/062/M313)
         3039 Cornwallis Rd.
         Research Triangle Park, NC 27709
         Phone: 919-254-7605
Note: Along with this hardcopy, please include the name, e-mail and
phone contact information of the PR person that EPTS should
coordinate with for the launch press release.

[ our apologies should you receive this message more than one time ]


             2008 International RuleML Symposium
        on Rule Interchange and Applications (RuleML-2008)

           October 30-31, 2008, Orlando, Florida
                http://2008.ruleml.org

=====================================================================
Co-located with:

          The 11th International Business Rules Forum
               http://www.businessrulesforum.com
=====================================================================
Sponsored by:

Silver level: Model Systems
Bronze level: STI Innsbruck, ruleCore, JBoss
=====================================================================
In Co-operation with:

AAAI, W3C, BPM-Forum, Business Rules Forum , ECCAI, OASIS, OMG,
European Business Rules Conference, Belgium Business Rules Forum,
ACM, ACM SIGART, ACM SIGMIS, ACM SIGWEB, Open Research Society,
IEEE Systems Man and Cybernetics Society
IEEE Computer Society TC on Autonomous and Autonomic Systems (pending)
IEEE Computer Society TC on Intelligent Informatics (pending)
IEEE SMCS TC on Intelligent Internet Systems
IEEE SMCS TC on Distributed Intelligent Systems

=====================================================================
Media Partners:

          Springer LNCS, MoDo Marketing
=====================================================================


Call for Papers

Collocated with the 11th International Business Rules Forum, the 2008
International Symposium on Rule Interchange and Applications
(RuleML-2008) is
the second symposium (after last year's highly successful RuleML-2007 -
http://2007.ruleml.org/) devoted to work on practical distributed rule
technologies and rule-based applications which need language standards for
rules operating in the context of, e.g., the Semantic Web, Intelligent
Multi-
Agent Systems, Event-Driven Architectures and Service-Oriented Computing
Applications. The RuleML symposium is a new kind of event where the
Web Rules
and Logic community joins the established, practically oriented Forum
of the
Business Rules community (http://www.businessrulesforum.com) to help
cross-
fertilizing between Web and Business Logic technology.

The goal of RuleML-2008 is to bring together rule system providers,
representatives of, and participants in, rule standardization efforts
(e.g.,
SBVR, RuleML, RIF, PRR, CL) and open source rules communities (e.g., jBoss
Rules, CLIPS/Jess, Prova, OO jDrew, Mandarax, XSB, XQuery),
practitioners and
technical experts, developers, users, and researchers. They will be
offered
an exciting venue to exchange new ideas, practical developments and
experiences on issues pertinent to the interchange and application of
rules
in open distributed environments such as the Web.

The Symposium gives emphasis on practical issues such as technical
contributions and show case demonstrations of effective, practical,
deployable rule-based technologies, rule interchange formats and
applications
as well as discussions of lessons learned that have to be taken into
account
when employing rule-based technologies in distributed, (partially) open,
heterogeneous environments. We also welcome groundwork that helps to
build an
effective, practical, and deployable rule standard, improve rule
technology,
provide better understanding of the integration and interchange of
rules, and
make the current generation of rule engines and rule technology more
usable
for advanced Web and Service Oriented Architectures.

RuleML-2008 highlights include:
- Plenary keynotes and a joint Boxed Lunch Panel about "Rules on the Web"
   together with the Business Rules Forum featuring prominent and visionary
   speakers.
- There will be a joint keynote between RuleML-2008 and RR2008
   (http://www.rr-conference.org/RR2008/), which will be broadcast from
RR2008
   to RuleML-2008.
- A RuleML-2008 Challenge with prizes to demonstrate tools, use cases, and
   applications
- Industry, demo and scientific research & development papers and
   presentations advancing and assessing the state of the art in event and
   rule- based systems selected in a peer-reviewed fashion by an
international
   program committee
- Invited talks given by leaders from industry and world-class experts
   featuring practical topics on event and rule-based computing and
industry
   success stories
- Social events to promote networking among the symposium delegates in an
   informal setting.


Topics of Interest
------------------

We invite industry practitioners, rule system providers, technical experts
and developers, rule users, and researchers who are using rule-based
systems,
developing systems and applications, or exploring problems and best
practices
(especially in the areas of system interoperability, rule interchange, or
business agility), to share their ideas, results, and experiences. We
invite
submissions related (but not limited to) to one or more of the following
topics:

- Representation and meta-annotation of rules and rule sets (modules) for
   publication and interchange
- Collaborative authoring, modeling and engineering of rule specifications
   and rule repositories
- Information integration of external data and domain knowledge into rules
- Homogeneous and heterogeneous integration of  rules and ontologies
- Rules in Web 2.0 and Web 3.0
- Rules in Semantic Web Technologies
- Rules in Web Intelligence Research
- Hybrid rule systems
- Management and maintenance of distributed rule bases and rule
repositories
   during their lifecycle
- Interchange and refactoring of rule bases in heterogeneous execution
   environments
- Verification and validation of interchanged rule bases in heterogeneous
   execution environments
- Contributions on effective, practical, and deployable Web standards on
   rules as well as special purpose, vertical domain rule languages
- Rule-based agility and its role in middleware
- Communication between rule based systems using interchange formats and
   processing / communication middleware
- Applications, products, research, and development in rule-based,
   distributed complex event processing, event communication and reaction
   rules
- Event-driven/action rule languages and models
- Rule-based Event Processing Languages and rule-based CEP
- Rule patterns and CEP patterns
- Practical solutions tackling the real-world Software Engineering
   requirements of rule-based systems in open, distributed environments
- Modeling of executable rule specifications and tool support
- Execution models, rule engines, and environments
- Compilation vs. interpretation approaches of rules
- Applications and integration of rules in web standards
- Rule-based software agents and (web) services
- Applications of rules in the Semantic Web and Pragmatic Web
- Comparing and advancing the state of current business rules engines and
   management system tools
- Rule interchange standards and related industry interchange formats
- Interoperation between different rule formats and ontological domain
   conceptualization
- Applications based on (Semantic) Web rule standardization or standards-
   proposing efforts
- Translation of interchangeable and domain-independent rule formats
and rule
   models into executable technical rule specifications
- Extraction and reengineering of platform-independent,
interchangeable rules
   and rule models from existing platform-specific resources
- Natural-language processing of rules
- Graphical processing, modelling and rendering of rules
- Incorporation of rule technology into distributed enterprise application
   architectures
- Rule-based policies and electronic contracts: their specification,
   execution and management
- Languages for exchanging and processing information through the web
- E-contracting and automated negotiations with rule-based declarative
   strategies
- Applications of rules in e.g. legal reasoning, compliance rules,
security,
   IT government, security, risk management, trust and proof reasoning,
etc.
- Rule-based (multi-valued) reasoning with and representing uncertain and
   fuzzy information
- Rule-based reasoning with non-monotonic negation, modalities, deontic,
   temporal, priority, scoped or other rule qualifications
- Rule-based default reasoning with default logic, defeasible logic, and
   answer set programming


RuleML-2008 Challenge
---------------------

RuleML-2008 Challenge addresses the practical use of rule technologies in
distributed and/or Web-based environments. The focus of the challenge
is on
rule technologies (including rule languages and engines),
interoperation and
interchange. The challenge offers participants the chance to demonstrate
their commercial and open source tools, use cases, and applications.
Prizes
will be awarded to the two best applications. All accepted demos will be
presented in a special Challenge Session.

Intentionally, this year's challenge (as last year) does not define a
specific task, data set, application domain, or technology to be used
because
the potential applicability of rules is very broad. Instead, a number of
minimal criteria were defined which allow people to submit a broad
range of
applications.

A submission to RuleML challenge has to meet the minimal requirement that
declarative rules explicitly play a central role in the application.
Basically this means that:

- Rules are explicitly represented in a declarative format and they are
   decoupled from the application (rather than being compiled or hard-coded
   into the application logic).
- Rules are used in interesting and practically relevant ways to, e.g.,
   derive useful information, transform knowledge, provide decision
support,
   provide automated rule-based monitoring, enforcement, validation or
   management of the behavioural logic of the application.

The demo should preferably (but not necessarily) be embedded into a
web-based
or distributed environment so that there will be a need for features
related
to the ruleml conference topics, as listed in the call for papers. More
details will be available on the RuleML-2008 web site.


Important Dates
---------------

- Paper Submissions due  June 2, 2008
- Notification of acceptance July 18, 2008
- Final submissions due  August 9, 2008
- Symposium date  October 30-31, 2008
- RuleML Challenge  October 30, 2008


Symposium Proceedings and Submission Details
--------------------------------------------

Authors are invited to submit original contributions of practical
relevance
and technical rigor in the field, experience reports and show case /
use case
demonstrations of effective, practical, deployable rule-based
technologies or
applications in distributed environments. Papers must be in English
and may
be submitted at http://www.easychair.org/conferences/?conf=ruleml2008 as:

- Full Papers (15 pages in the proceedings)
- Short Papers (8 pages in the proceedings)
- RuleML-2008 Challenge Demo Paper + Show Cases (3-5 pages in the
proceedings)

Please upload all submissions as PDF files in LNCS format
(http://www.springer.de/comp/lncs/authors.html). To ensure high quality,
submitted papers will be carefully peer-reviewed by 3 PC members based on
originality, significance, technical soundness, and clarity of exposition.
Authors are requested to upload their complete papers by June 2, 2008. The
selected papers will be published in book form in the Springer Lecture
Notes
in Computer Science (LNCS) series (pending) along with a CD with demo
software and documents. The best paper from all submissions will be
determined by the PC and a Best Paper Award will be handed over at the
Symposium by a Sponsor. All submissions must be done electronically via
http://www.easychair.org/conferences/?conf=ruleml2008. We will pursue the
publication of a selection of revised papers to a special issue of a high-
quality journal.

Submissions to the RuleML Challenge 2008 consist of a demo paper of 3-5
pages, describing the demo show case, and a link to more information about
the demo/show case, e.g. a project site, an online demonstration, a
presentation, or a download site. If the link is password-protected: then
please submit a password for anonymous login from any Web browser,
giving us
the permission to pass the password on to 3 PC members. The submissions
should satisfy the minimal requirements defined in the topics of interest.
The demos will be evaluated by an independent jury of experts during
RuleML-2008 and prizes will be awarded to the first two best applications.

Please do not hesitate to email to the appropriate Symposium Chair(s),
if you
have any questions.


Organizing Committee
--------------------

General Chair

Adrian Paschke, Technical University Dresden, Germany
        adrian.paschke AT biotec.tu-dresden.de

Program Co-Chairs

Nick Bassiliades, Aristotle University of Thessaloniki, Greece
        nbassili AT csd.auth.gr
Guido Governatori, University of Queensland, Australia
        guido AT itee.uq.edu.au

Challenge Co-Chairs

Costin Badica, University of Craiova, Romania
        badica_costin AT software.ucv.ro
Yuh-Jong Hu, National Chengchi University, Taiwan
        hu AT cs.nccu.edu.tw

Panel Co-Chairs

John Hall, Model Systems, UK
        john.hall AT modelsys.com
Axel Polleres, DERI Galway, Ireland
        axel AT polleres.net

Liaison Co-Chairs

Mark Proctor, JBoss Rules, UK
        mproctor AT redhat.com
Rainer von Ammon, CITT GmbH, Germany
        vonammon AT t-online.de
Jan Vanthienen, Katholieke Universiteit Leuven, Belgium
        Jan.Vanthienen AT econ.kuleuven.be


Publicity Co-Chairs

Matthias Nickles, University of Bath, UK
        M.L.Nickles AT cs.bath.ac.uk
Tracy Bost, Valocity, USA
        tbost AT valocity.com
(Sponsoring levels: http://2008.ruleml.org/sponsoring/)

Web Chair

Suzanne Embury, University of Manchester, UK
        Suzanne.Embury AT manchester.ac.uk


Program Committee
--------------------

- Asaf Adi, IBM, Israel
- Grigoris Antoniou, University of Crete-FORTH, Greece
- Sidney Bailin, Knowledge Evolution, USA
- Matteo Baldoni, University of Torino, Italy
- Cristina Baroglio, University of Torino, Italy
- Claudio Bartolini, HP Labs
- Tim Bass, SilkRoad Inc.
- Bernhard Bauer, University of Augsburg, Germany
- Mikael Berndtsson, University of Skovde, Sweden
- Leopoldo Bertossi, Carleton University, Canada
- Pedro Bizarro, University of Coimbra, Portugal
- Peter Bollen, University of Maastricht, Netherlands
- Christian Brelage, SAP Research, Germany
- Donald Chapin, Business Semantics Ltd, UK
- Shyi-Ming Chen, National University of Science and Technology, Taiwan
- Jorge Cuellar, Siemens, Germany
- Mike Dean, BBN Technologies, USA
- Stan Devitt, Agfa Healthcare, Ontario, Canada
- Jens Dietrich, Massey University, New Zeeland
- Jurgen Dix, Technische Universitaet Clausthal, Germany
- Schahram Dustdar, Vienna University of Technology, Austria
- Andreas Eberhart, Portland State University, USA
- Opher Etzion, IBM Research Laboratory Haifa, Israel
- Dieter Fensel, DERI Innsbruck, Austria
- Dragan Gasevic, Athabasca University, Canada
- Adrian Giurca, Brandenburg University of Technology at Cottbus, Germany
- Stijn Goedertier, Katholieke Universiteit Leuven, Belgium
- Robert Golan, DBmind Technologies, USA
- Christine Golbreich, University Rennes, France
- Tom Gordon, Fraunhofer Institute for Open Communication Systems, Germany
- Marek Hatala, Simon Fraser University, Canada
- Ioannis Hatzilygeroudis, University of Patras, Greece
- Yiannis Kompatsiaris, Informatics and Telematics Institute, Greece
- Manolis Koubarakis, National and Kapodistrian University of Athens,
Greece
- Alex Kozlenkov, Betfair Ltd., UK
- Holger Lausen, DERI Innsbruck, Austria
- John Lee, Polytechnic University, Hong Kong
- Mark Linehan, IBM Research, USA
- Heiko Ludwig, IBM Watson Research, USA
- Mirko Malekovic, University of Zagreb, Croatia
- Christopher J. Matheus, Vistology, USA
- Craig McKenzie, Science Applications International, USA
- Jing Mei, IBM Research Lab China, China
- Zoran Milosevic, Deontik Inc, Australia
- Jang Minsu, E&T Research Institute, Korea
- Leora Morgenstern, Stanford, USA
- Gero Muehl, Technische Universitat Berlin, Germany
- Jorg Muller, TU Clausthal, Germany
- Chieko Nakabasami, Toyo University, Japan
- Ilkka Niemela, Helsinki University of Technology, Finland
- Bart Orriens, Tilburg University, Netherlands
- Jeff Pan, University of Aberdeen, UK
- Paula-Lavinia Patranjan, Rewerse, Germany
- Jon Pellant, Pega Systems Inc., USA
- Jeff Pollock, Oracle, USA
- Alun Preece, Cardiff University, UK
- Maher Rahmouni, HP Labs
- Girish Ranganathan, University of New Brunswick, Canada
- Dave Reynolds, HP, England
- Graham Rong, MIT Sloan School of Management, USA
- Antonio Rotolo, CIRSFID, University of Bologna, Italy
- Norman Sadeh, Carnegie Mellon University, USA
- Christian de Sainte Marie, ILOG, France
- Marco Seirio, ruleCore, Sweden
- Timos Sellis, Institute for the Management of Information Systems and
   National Technical University of Athens, Greece
- Michael Sintek, DFKI, Germany
- Silvie Spreeuwenberg, LibRT, The Netherlands
- Giorgos Stamou, National Technical University of Athens, Greece
- Giorgos Stoilos, National Technical University of Athens, Greece
- Terrance Swift, XSB Inc., USA
- Kuldar Taveter, University of Melbourne, Australia
- James Taylor, Fair Isaac Corp., USA
- Vagan Terziyan, University of Jyvaskyla, Finland
- Paul Vincent, TIBCO Software, USA
- George Vouros, University of the Aegean, Greece
- Kewen Wang, Griffith University, Australia
- Mehmet Emre Yegen, Ygntec Inc., USA

Hi CEP friends

I think that the CEP page on wikipedia could *easily* be written without talking about vendors. The vendors don't have to worry, we are all easy to find with google anyway.

There are a number of strong concepts within the event processing community. Rules, SQL based approaches, event oriented programming languages and so on. All these can be describes in an interesting way without any reference to any vendor. Vendors come and go, vendor information is not terribly encyclopedic (or whatever the term is).

Obvisouly all vendors would like to be mentioned on the Wikipedia, but if we think for a moment of the readers of the Wikipedia. They are better served with information about event processing basics and not about vendor implementations which will change every year anyway.

For vendors, there's www.eventprocessing.eu wiki which happily accepts vendor information.


/Marco

On fre 08/03/21 10:58 , "Alexandre Vasseur" avasseur@... sent:

I think the point on wikipedia is only about agreeing that EPL is an
acronym and stands for Event Processing Language, ie a domain specific
language commonly found in ESP and CEP products, and that it is not a
single vendor defined term, nor a standard.
I cannot comment on each vendor own strategy between multiple
overlapping products / risk of soon end of lifed products, as well as
approach to standardization in a defacto or dejure form but I agree
with your observation on the time it would take to reach a true
standardization. Mitigating vendor lockin is not only a matter of EPL
standardization.
Alex

On Fri, Mar 21, 2008 at 8:51 AM, Rainer von Ammon <vonammon@...> wrote:
>
>
>
>
>
>
>
>
>
> I think Hans is right, could be frustrating. I remember the discussion about
> Wikipedia as well and of course the most EPLs are from commercial
> enterprises and are proprietary, because there is no standard so far. The
> CEP community has estimated that it'll take more than 5 years to establish a
> standard for EPL. And some of us mean that a standard EPL wouldn't make
> sense and there MUST be different EPL approaches for different purposes like
> algotrading, fraud detection, EDBPM/BAM and so on.
>
> Also, there are some movements ongoing, e.g. BEA brought its Event Server to
> the market some months ago and OEM-licensed Esper, now BEA is going to be
> acquired by Oracle, but Oracle just announced its CEP-approach and its EPL
> is named CQL (Continuous Query Language) and has then 2 CEP-approaches, how
> long? The same goes for IBM, just acquired AptSoft, additionally to its
> AMiT, but stopped ACT because acquired another company… But nevertheless, we
> could write just about that in Wikipedia in a neutral way and it seems to be
> interesting for some potential CEP-adopters. In January 2008 at the OOP2008,
> Clemens Utschig, a product manager from Oracle, was asked by the auditorium
> whether EPL like CQL would already be standardized, otherwise they wouldn't
> use it because of vendor lockin, perhaps typical for Germans… How long is
> the lifetime for software applications? How long could we wait for a
> standard regarding our projects while the competitors earn the money by
> algorithmic trading….
>
> I just wait on the standardized Easter Bunny, also he will bring
> not-standardized Easter eggs this year again, but my children won't care
> about:-)
>
>
>
>
>
> -----Ursprüngliche Nachricht-----
> Von: CEP-Interest@yahoogroups.com [CEP-Interest@yahoogroups.com] Im
> Auftrag von hansgilde
> Gesendet: Donnerstag, 20. März 2008 18:39
> An: CEP-Interest@yahoogroups.com
> Betreff: [CEP-Interest] Re: EPL on wikipedia
>
>
>
>
>
>
>
>
> Just a general comment about Wikipedia. If you look at the talk
> (Discussion) page for Complex Event Processing on Wikipedia, you will
> see that far more writing has been devoted to arguing about this page
> than to adding content.
>
> It is very hard to talk about CEP without mentioning commercial
> products or vendors, but certain Wikipedia editors (who are not
> involved with the CEP community) insist on deleting all content
> referring to commercial entities. They also delete content referring
> to non-commercial projects that in turn link to a supporting vendor
> (for example, they delete references to the open source Esper project
> because there is a commercial entity supporting it).
>
> I suspect that you will find the same thing if you begin to write a
> page about EPLs. It may be very frustrating or impossible to complete
> an article to your satisfaction (this has been the case with the
> Complex Event Processing article).
>
> So I suggest that you consider another venue for publishing this
> information.
>
> --- In CEP-Interest@yahoogroups.com, "Alexandre Vasseur"
> wrote:
> >
> > I have found this EPL article on wikipedia
> > http://en.wikipedia.org/wiki/Event_Programming_Language
> >
> > I am calling the group to enhance this article that I personnaly find
> > misleading. Before starting the process on wikipedia, I think we
> > should iterate some in this expert group.
> > Alex
> >
>
>

------------------------------------

Yahoo! Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/CEP-Interest/

<*> Your email settings:
Individual Email | Traditional

<*> To change settings online go to:
http://groups.yahoo.com/group/CEP-Interest/join
(Yahoo! ID required)

<*> To change settings via email:
CEP-Interest-digest@yahoogroups.com
CEP-Interest-fullfeatured@yahoogroups.com

<*> To unsubscribe from this group, send an email to:
CEP-Interest-unsubscribe@yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/

I think the point on wikipedia is only about agreeing that EPL is an
acronym and stands for Event Processing Language, ie a domain specific
language commonly found in ESP and CEP products, and that it is not a
single vendor defined term, nor a standard.
I cannot comment on each vendor own strategy between multiple
overlapping products / risk of soon end of lifed products, as well as
approach to standardization in a defacto or dejure form but I agree
with your observation on the time it would take to reach a true
standardization. Mitigating vendor lockin is not only a matter of EPL
standardization.
Alex

On Fri, Mar 21, 2008 at 8:51 AM, Rainer von Ammon <vonammon@...> wrote:
>
>
>
>
>
>
>
>
>
> I think Hans is right, could be frustrating. I remember the discussion about
> Wikipedia as well and of course the most EPLs are from commercial
> enterprises and are proprietary, because there is no standard so far. The
> CEP community has estimated that it'll take more than 5 years to establish a
> standard for EPL. And some of us mean that a standard EPL wouldn't make
> sense and there MUST be different EPL approaches for different purposes like
> algotrading, fraud detection, EDBPM/BAM and so on.
>
> Also, there are some movements ongoing, e.g. BEA brought its Event Server to
> the market some months ago and OEM-licensed Esper, now BEA is going to be
> acquired by Oracle, but Oracle just announced its CEP-approach and its EPL
> is named CQL (Continuous Query Language) and has then 2 CEP-approaches, how
> long? The same goes for IBM, just acquired AptSoft, additionally to its
> AMiT, but stopped ACT because acquired another company� But nevertheless, we
> could write just about that in Wikipedia in a neutral way and it seems to be
> interesting for some potential CEP-adopters. In January 2008 at the OOP2008,
> Clemens Utschig, a product manager from Oracle, was asked by the auditorium
> whether EPL like CQL would already be standardized, otherwise they wouldn't
> use it because of vendor lockin, perhaps typical for Germans� How long is
> the lifetime for software applications? How long could we wait for a
> standard regarding our projects while the competitors earn the money by
> algorithmic trading�.
>
> I just wait on the standardized Easter Bunny, also he will bring
> not-standardized Easter eggs this year again, but my children won't care
> about:-)
>
>
>
>
>
> -----Urspr�ngliche Nachricht-----
>  Von: CEP-Interest@yahoogroups.com [mailto:CEP-Interest@yahoogroups.com] Im
> Auftrag von hansgilde
>  Gesendet: Donnerstag, 20. M�rz 2008 18:39
>  An: CEP-Interest@yahoogroups.com
>  Betreff: [CEP-Interest] Re: EPL on wikipedia
>
>
>
>
>
>
>
>
> Just a general comment about Wikipedia. If you look at the talk
>  (Discussion) page for Complex Event Processing on Wikipedia, you will
>  see that far more writing has been devoted to arguing about this page
>  than to adding content.
>
>  It is very hard to talk about CEP without mentioning commercial
>  products or vendors, but certain Wikipedia editors (who are not
>  involved with the CEP community) insist on deleting all content
>  referring to commercial entities. They also delete content referring
>  to non-commercial projects that in turn link to a supporting vendor
>  (for example, they delete references to the open source Esper project
>  because there is a commercial entity supporting it).
>
>  I suspect that you will find the same thing if you begin to write a
>  page about EPLs. It may be very frustrating or impossible to complete
>  an article to your satisfaction (this has been the case with the
>  Complex Event Processing article).
>
>  So I suggest that you consider another venue for publishing this
>  information.
>
>  --- In CEP-Interest@yahoogroups.com, "Alexandre Vasseur"
>  <avasseur@...> wrote:
>  >
>  > I have found this EPL article on wikipedia
>  > http://en.wikipedia.org/wiki/Event_Programming_Language
>  >
>  > I am calling the group to enhance this article that I personnaly find
>  > misleading. Before starting the process on wikipedia, I think we
>  > should iterate some in this expert group.
>  > Alex
>  >
>
>

Just a general comment about Wikipedia. If you look at the talk
(Discussion) page for Complex Event Processing on Wikipedia, you will
see that far more writing has been devoted to arguing about this page
than to adding content.

It is very hard to talk about CEP without mentioning commercial
products or vendors, but certain Wikipedia editors (who are not
involved with the CEP community) insist on deleting all content
referring to commercial entities. They also delete content referring
to non-commercial projects that in turn link to a supporting vendor
(for example, they delete references to the open source Esper project
because there is a commercial entity supporting it).

I suspect that you will find the same thing if you begin to write a
page about EPLs. It may be very frustrating or impossible to complete
an article to your satisfaction (this has been the case with the
Complex Event Processing article).

So I suggest that you consider another venue for publishing this
information.

--- In CEP-Interest@yahoogroups.com, "Alexandre Vasseur"
<avasseur@...> wrote:
>
> I have found this EPL article on wikipedia
> http://en.wikipedia.org/wiki/Event_Programming_Language
>
> I am calling the group to enhance this article that I personnaly find
> misleading. Before starting the process on wikipedia, I think we
> should iterate some in this expert group.
> Alex
>

Note that the article in question is not for "EPL" or "Event
Processing Language", but for "Event Programming Language", which is
what iSpheres called their language. Perhaps a new article should be
created and the old one deleted?

Hans

--- In CEP-Interest@yahoogroups.com, "Rainer von Ammon" <vonammon@...>
wrote:
>
> Hi Alex, you're right, this entry should be improved.
>
> We tried to give a summary of the EPL-discussion in the CEP community in
> the thesis of David Guschakowski and Hans-Martin Brandl on the basis of
> the 5th expert meeting
> http://www.citt-online.com/downloads/5-bpm-bam-cep-soa-agenda-v3.doc ,
> slot 2pm first day, and our CEP-course at the University of Appl Sc. of
> Regensburg http://www.citt-online.com/index.php?id=lehre
> <http://www.citt-online.com/index.php?id=lehre&id3=course_eventprocessin
> g&id4=more> &id3=course_eventprocessing&id4=more . Actually the
> following is chap. 8 of
> http://www.citt-online.com/downloads/Diplomarbeit_BaGu_Final.pdf
>
> Tim Bass already asked us some months ago to write an article on the
> basis of this, but I had no time the last months, perhaps I'll do it
> together with David/Hans-Martin within the next weeks.
>
> Perhaps it could contribute to your discussion.
>
>
>
> Cheers,
>
> /rainer
>
>
>
>
> 1          Event Processing Languages
>
>
> High thoughts must have high language.
>
> Aristophanes, Frogs, 405 B.C.
>
>
>
> This chapter gives an overview about the different event processing
> language approaches. A variety of CEP platform vendors will be
> classified depending on their programming language for event processing.
> The statements and results of the vendor's representatives and business
> analysts will be evaluated and summarized. A clearly classification of
> all vendors to a particular approach is not possible, because some
> vendors provide different approaches within their platforms or the EPL
> is embedded into another language approach. In this thesis the EPLs are
> classified as "pseudo" SQL approach, special rule language or script
> language approach, Java approach, XML approach and GUI approach.
>
>
> 1.1         "Pseudo" SQL approaches
>
>
> The main intention of vendors providing "SQL-based" approaches is that
> developers of relational databases can apply their knowledge based on
> SQL. These vendors offer language constructs of the SQL standard, which
> are well known in database expert groups. Common keywords as INSERT,
> SELECT, FROM and WHERE are used to filter and manipulate event streams
> and insert data from one stream into another one. They compare an event
> correlation with a join of database tables, whereby the join in the
> event stream platform takes place by joining streams [Coral8 2007a],
> [Coral8 2007b], [StreamBase 2007e], [Esper 2006a], [Esper 2006b], [Aleri
> 2006a]. The EPL of Coral8 is the continuous computing language CCL.
> Coral8 calls it "SQL-based", because it provides these common keywords
> to process events. Additionally they offer a variety of non SQL
> expressions to accomplish the new paradigm of event processing.
> Therefore CCL has expressions to define windows CREATE WINDOW or KEEP 10
> SECONDS, sequences or patterns MATCHING [10 Seconds: A, !B] or access on
> previous events PREV() [Coral8 2007a], [Coral8 2007b]. Figure 102 shows
> a processing statement.
>
>
>
> Figure 102: Coral8 code example
>
> The language StreamSQL of StreamBase is another proprietary approach,
> which is based on expressions, defined in the SQL standard. StreamSQL
> includes expressions like SELECT, FROM and WHERE, which have basically
> the same semantic as the corresponding SQL expressions. A complete
> statement needs the CEP enhancements like CREATE STREAM to define a
> stream or INTO to route events into a defined stream [StreamBase 2007e]:
>
> CREATE STREAM completedApplication;
>
> CREATE STREAM cancelledApplication;
>
> SELECT * FROM read_all_rows
>
> WHERE time-entry_creationTime>minutes(30) AND isnull(exit_creationTime)
> INTO cancelledApplication
>
> WHERE notnull(entry_creationTime) AND notnull(exit_creationTime) INTO
> completedApplication;
>
>
>
> The event query language EQL of the open source project Esper has a
> similar syntax as SQL. In addition, EQL offers capabilities for event
> stream analysis, especially the usage of event streams and a concept
> called views. The appendix .win:time(30 min) on a stream identifier
> creates such a view. Similar to tables in a SQL statement, views define
> the data of a stream which is available for querying and filtering.
> Other non SQL expressions are used to define scenarios of events such as
> every a=EventX -> every b=EventY(objectID=a.objectID). These stream
> analysis capabilities and scenario definitions can be combined into a
> statement like the following [Esper 2006a], [Esper 2006b]:
>
>
>
> select a.id, count(*)
>
> from pattern [ every a=Status -> (timer:interval(10 sec) and not
> Status(id=a.id)]
>
> group by id
>
>
>
> The stream processing platform provider Aleri also features an EPL,
> which uses SQL keywords. The Aleri platform provides the implementation
> of CEP applications with the SQL based EPL Aleri SQL. Keywords like
> SELECT, WHERE, HAVING or GROUP BY are used to leverage the
> implementation of CEP applications for database developers [Aleri
> 2006a], [Aleri 2006b], [Yahoo Tech Group 2006a]. Many keywords defined
> in the SQL:2003 standard are used in the particular EPLs [T�rker 2004,
> P. 181ff]. The semantic difference in the usage of SQL as a query
> language on a given set of data and an EPL as a query language on
> flowing event streams implies different usage of the particular keywords
> [Yahoo Tech Group 2006]. In order to use aggregate functions like SUM or
> AVG or grouping functions like GROUP BY, the individual events must be
> collected by window techniques to get a set of data. Additionally one of
> the core functionalities of CEP, the recognition of event scenarios or
> patterns, is not covered by the SQL standard. These functionalities are
> managed by proprietary expressions of the respective vendors. The
> resulting queries or statements are implemented in a different syntax
> than the defined in the SQL standard. The commonness of the particular
> EPL query to a SQL query is usually only the SELECT FROM WHERE clause.
> Thus these EPLs are classified as "pseudo" SQL languages [Ammon 2006a].
>
>
> 1.2         Special rule languages or script languages
>
>
> Another declarative language approach origins from the area of rules
> engines. Some vendors use similar syntaxes as business rules management
> systems, whereby they extend the common IF THEN blocks which describe
> causal dependencies, with additional language constructs in order to
> describe temporal relationships of events. So e.g. Apama extends the IF
> THEN block with the FOLLOWED-BY clause to describe the temporal order of
> events and the ALL WITHIN clause to define time windows [CEP-Interest
> 2007], [Yahoo Tech Group 2006d].
>
>
> Figure 103: Apamas Rule language [CEP-Interest 2007]
>
> IBM's solution Active Middleware Technology (AMiT) uses a rule language
> which allows exact description of situations. The life span of the
> situation can be described exactly with the initiator and terminator.
> Additionally events can be classified as "relevant" or be "rejected".
> The relevant events can be defined exactly by their attributes like e.g.
> second operand = event: "E3" threshold: "X > 7" or third operand =
> event: "E4" threshold: "X > 7". Correlation conditions can be described
> like condition = "E3.X = E4.X". So, situations can be described like the
> following:
>
> operator = "sequence"
>
>   detection mode = "differed"
>
>   lifespan = "trading_day"
>
>   first operand = event: "stock-quote"  as: "first-quote"
>
>     threshold: "change > 0"
>
>     quantifier: "each"
>
>     consumption condition: "false"
>
>   second operand = event: "stock-quote" as: "second-quote"
>
>     threshold: "change > 0"
>
>     quantifier: "last"
>
>     consumption condition: "true"
>
>
>
> The Rules Manager, which is included in the Oracle 10g database
> management system, implements ECA rules with the IF THEN rule language
> terminology. Action procedures can be implemented by SQL procedures.
> These rules and corresponding actions are stored in data base tables.
> Figure 104 shows the definition of events, rules and actions [Oracle
> 2006]. Press releases of Oracle also introduced CEP capabilities of the
> Oracle Fusion middleware platform, but a nearer look on the EPL of this
> platform could not be done within this thesis [Oracle 2007].
>
>
>
> Figure 104: Oracle Rule Manager architecture and code example [Oracle
> 2006]
>
>
> 1.3         Java
>
>
> The implementation in BusinessEvents of Tibco can be done via Java.
> Therefore Tibco provides the TibcoDesigner, which allows a design of an
> event driven UML model. The UML-based "state model" describes how
> applications and services interact as part of activities and processes.
> The creation, destruction or state change of these objects is done by
> events, which match to the rule condition implemented in Java. A
> creation, destruction or state change of an object is thereby a part of
> an action and represents a new event and may influence other objects.
> The implementation of these rules is done in Java. Describing an event
> driven UML model and writing rules in Java which can filter, correlate
> and aggregate events by applying constraints and threshold boundaries is
> the main part of implementing CEP functionalities. Thus the Tibco EPL is
> classified as Java based [Tibco 2007].
>
>
>
>
>
> Figure 105: Tibco rule for the creation of a new CreditApplication
> object
>
>
> 1.4         XML
>
>
> Some platforms like Aleri, Coral8 or StreamBase store their applications
> in XML files, for example Aleri applications are stored in Aleri XML
> [Aleri 2006b]. The modules developed in Coral8 are stored in .ccl files,
> which are structured in XML. Thereby the CCL statements are embedded
> into Query tags [Coral8 2006a]. The applications built with StreamBase,
> are stored in .ssql and .sbapp files with a XML structure [StreamBase
> 2007a]. These files can also be manipulated with other applications
> [Aleri 2006b].
>
>
> 1.5         Graphical User Interface
>
>
> Several application areas of CEP like BAM, exchange surveillance or
> algorithmic trading affect users without any programming knowledge, e.g.
> business analysts, sales [CEP-Interest 2007], [Luckham 2005a]. Vendors
> like AptSoft, StreamBase, Apama or Aleri try to bridge this gap by
> providing a GUI [Yahoo Tech Group 2006d]. Therewith CEP applications can
> be built for example by dragging and dropping graphical components into
> a diagram and connecting them [StreamBase 2007b], [Aleri 2006b]. The
> event driven object models of Tibco can also be built within graphical
> environments [Tibco 2006b]. AptSoft delivers graphical components, which
> can be reused in other components in order to build more complex queries
> [AptSoft 2006a], [AptSoft 2006b]. These graphical components can be
> configured or parameterized in order to perform a desired function like
> for example a filter function based on the configured condition. These
> parameters are subject of a proprietary syntax, so for example a filter
> condition in StreamBase looks like TradeInformation.Volume > 300000 and
> in AptSoft a filter condition is implemented by applying conditions with
> text boxes and dropdown boxes as shown in the Figure below.
>
> Figure 106: AptSoft's event processing GUI [AptSoft 2006b]
>
> The variety of different graphical components and their parameterization
> depends on the particular platform of a vendor [AptSoft 2006a], [AptSoft
> 2006b], [StreamBase 2007b], [Yahoo Tech Group 2006c]. Except of AptSoft,
> generally these platform vendors provide an underlying or separate
> computing language which can be generated from the graphical designed
> applications. These can also be combined with graphical components
> [StreamBase 2007b], [Yahoo Tech Group 2006c].
>
>
> 1.6         Summary and evaluation of the results
>
>
> The opinions about the different EPLs on the present CEP market vary
> very much within in the CEP community, whereby the representatives of
> the individual vendors advocate their own ideologies. So the endeavors
> of StreamBase for SQL as the base for a standardization of EPLs caused
> many discrepancies in the CEP community. It was said that the vendors of
> CEP platforms chose their language for good reasons [IT-Director 2006].
>
> Vendors striving for other goals disagreed that SQL is a good approach
> to process events at all. In their opinion the data set oriented
> processing of SQL does not fit to the event processing paradigm [Yahoo
> Tech Groups 2006e]. They also claim that applications built in SQL based
> approaches are difficult to understand and to maintain. The development
> and maintenance of SQL based CEP applications can just be done by
> developers, not by business persons. In their opinion authoring tools
> with graphical components do not change this fact [Tech Groups 2006f].
> In the author's opinion "pseudo" SQL approaches do not address the
> qualification of clerks, managers or other business related persons,
> whether the other approaches do not address these qualifications, too.
> Usually business related persons are not educated in programming or
> building applications. GUIs or almost natural languages do not change
> the fact, that an understanding of the CEP paradigm and programming and
> design concepts is necessary to build good CEP applications. Study
> courses like business IT were created to bridge these gaps of business
> and IT related persons. Additionally these vendors see problems in
> handling hierarchical event data. SQL based approaches usually use
> tuples as data processing units, which is very restricting. So simple
> events like stock tick events are easy to represent with a tuple. They
> doubt that more complex events with hierarchical structures like an
> "Order Submitted" event which contains multiple order items are
> manageable in tuples; they believe that events usually contain business
> objects as context information [Tech Groups 2006f]. In the author's view
> this problem is manageable with considerations which are related to
> normalizations in data bases (see para. 7.1.2, 7.2.2). So events like
> the CBE which have a hierarchical structure can be handled with stream
> splitting, whereby the attribute identifying an event is the "foreign
> key" for the stream containing the related context information (see
> para. 7.1.2, 7.2.2). To attain more clarity in the topic of how a
> language approach can solve CEP problems Etzion offered two possible CEP
> related problems [Yahoo Tech Group 2006g]. Vendors providing SQL based
> approaches like Hagman of Coral8 and Kulleen of StreamBase offered
> solutions to these problems [Yahoo Tech Group 2006h], [Yahoo Tech Group
> 2006i]. Cameron of AptSoft also offered a solution, which is designed
> graphically [AptSoft 2006b]. Regrettably these endeavours came to no
> results, because not enough different solutions were posted in order to
> evaluate differences, weaknesses and opportunities [Yahoo Tech Group
> 2006j]. It can be said recapitulating about the recent statements that
> there is still no sufficient progress in the evaluation about the
> respective differences in the usability and behaviours of an EPL. Works
> like [Coral8 2007e], [Bry 2006] or this thesis contribute and lead in
> the direction of a more exact rateability of EPLs by describing their
> requirements.
>
>
>
> -----Urspr�ngliche Nachricht-----
> Von: CEP-Interest@yahoogroups.com [mailto:CEP-Interest@yahoogroups.com]
> Im Auftrag von Alexandre Vasseur
> Gesendet: Donnerstag, 20. M�rz 2008 13:48
> An: CEP-Interest
> Betreff: [CEP-Interest] EPL on wikipedia
>
>
>
> I have found this EPL article on wikipedia
> http://en.wikipedia
> <http://en.wikipedia.org/wiki/Event_Programming_Language>
> .org/wiki/Event_Programming_Language
>
> I am calling the group to enhance this article that I personnaly find
> misleading. Before starting the process on wikipedia, I think we
> should iterate some in this expert group.
> Alex
>

I have found this EPL article on wikipedia
http://en.wikipedia.org/wiki/Event_Programming_Language

I am calling the group to enhance this article that I personnaly find
misleading. Before starting the process on wikipedia, I think we
should iterate some in this expert group.
Alex