Hi again,

Here, I'm forwarding you an e-mail that I have received earlier today from Sanjaya Joshi of Userspace Corp.

Now, before you read the attached e-mail I would like to provide you with a little bit of background info. It recently came to my attention that the ASTM LECIS standard is coming up for a revision within the ASTM soon. Within the ASTM, there now seems to be a push towards using standards like XML, Webservices, etc. So one question that needs to be answered in the near future is whether we want/need to XMLify the ASTM LECIS standard during that revision. There was also talk about coming up with an LECIS API (Application Programming Interface) to simplify the implementation and foster 3rd party library development.
Of course, this upcoming ASTM revision will also provide us with an opportunity to fix all the shortcomings of the current standard that we have encountered over the years. So please get your change requests ready. I'll keep you posted on the ASTM revision process.

Having said all this, here is what Sanjaya Joshi had to say: (BTW, the attachment Sanjaya is referring to in this e-mail is now available for download at http://www.lecis.org/downloads.htm)

Subject: RE: ASTM is preparing for LECIS revision (with LECIS commentary attachment in PDF)
Date: Thu, 20 Feb 2003 16:14:04 -0800
Organization: Userspace Corproration

All,

(This is Sanjaya Joshi from Userspace)
SOAP (Simple Object Access Protocol) died a horrible death during our
testing...
Web services is a good marketing ploy, but the actual implementation is
not quite there yet. WSDL is CORBA-Like, but support and applications
are not industrial strength.
(our first XML project was in 1998)

What would be interesting is continuing on the path of XML-RPC (we moved
to this years ago, and based our thought process on the IML or the
Instrument Markup Language which was originally a NASA project for
Remote IR Telescopy).
There are rumbling in the process community of not sticking with RPC,
but standardization is very simple if the heart of SOAP without the fat
layers (ie XML-RPC) is retained.

Thinking of LECIS as a Process Protocol is instructive...
In my discussions with IBM and Microsoft (the benefactors of SOAP) and
the Workflow initiative, they openly admitted that SOAP was not really
meant to do very fast transactions, let alone real-time instrument
communications.
Look at the Workflow Management Coalitions eXtensible Process Definition
Language. (XPDL)
http://xml.coverpages.org/ni2002-12-10-b.html
Also NIST's process description:
http://ats.nist.gov/psl/xml/process-descriptions.html
There is also the Petri-Net Markup Language, but the node paradigm for
automation has so far been too complex a solution for automation...
(maybe a better one for data).

A rectangular LIMS model (in my opinion) should move into a dynamic
Process model...
We have looked at NetCDF and AnDI (the child of NetCDF for Analytical
Data Interchange), but there is no real threading between the major
instrument groups: LC, MS, NMR...
Also the Thermo-Galactic GAML (which needs to pass muster as a tested
DTD...)
There is SpectroML too:
http://www.mel.nist.gov/div826/msid/sima/02_instrmnt_chemrefdata.html

Our focus is on MS and NMR which are very data intensive.
Specific points from scientists is the inability of AnDI or GAML to
handle MS-MS data (and MS2+)
Separation of data from process is not going to work out, especially for
nascent protocols that keep changing till they become an "Assay"...

I would urge us to follow the route of DICOM (Digital Imaging and
Communications in Medicine) -- my previous incarnation was in Radiology
-- where the acceptance was much better with the user and vendor
community. DICOM is now an XML standard and merged with HL7 (Health
Level 7), the healthcare industry patient data standard.

My 50,000 ft 2 cent drop...

Attached is our overview (from Userspace) on LECIS (as a PDF
whitepaper)...
Note that this is not a "techie" paper, but more an "executive summary".
Also, the architecture diagram in the paper is dated...

You can contact me if you need to respond to my comments (or bash on
them!)

Best,

Sanjaya Joshi
President and CTO
Userspace Corporation