Writing as a
neutral non-technical observer, I have to state that your points here do
resonate, Sanjiva. Certain people in this Group obviously know a lot
about REST and equally obviously are keen devotees of the approach.
However when someone as evidently experienced and bright as Steven has such
difficulty in clarifying basic issues, it does give the impression that the
RESTafarians still have a bit of work to do in at least explaining the use of
REST in terms that are as clear and simple as the claimed nature of REST.
There is obviously far more vendor support for WS-*, and this could arguably be
explained in partly politico-commercial terms, but is this widespread support
entirely without technical merit??
A question for
the RESTafarians: what could the vendors usefully provide (apart from some
clear, simple-to-understand whitepapers!) to assist the task of REST
developers?
Gervas
From:service-orientated-architecture@yahoogroups.com
[mailto:service-orientated-architecture@yahoogroups.com]
On Behalf Of Sanjiva Weerawarana Sent: 09 December 2006 11:24 To:service-orientated-architecture@yahoogroups.com Subject: Re:
[service-orientated-architecture] REST FAQ
On Fri, 2006-12-08 at 17:51 -0500, Steve Vinoski
wrote:
> > Its a bit surprising how such a simple architectural pattern
> > requires so
> > much of explanation eh? ;-)
>
> Umm, remind me again how many pages of WS-* specs are there?
A lot- because WS-* has standardized a lot more stuff that REST has.
However, my point was that even the teeny bit that REST has standardized
is sooooo poorly understood as a platform for application integration.
Don't believe that? Go read the thread on a RESTful lightbulb on this
list and see how much it took to get some degree of consensus. And of
course the current ROA/SOA thread.
Neither is a silver bullet. I don't know of any WS-* folks who seriously
think that WS-* is a silver bullet. Its *amazing* to me how many REST
people (especially on this list) seem to really believe that REST is the
silver bullet and that it has been a silver bullet for 15 years and
people just have not known about it. Wow, it must be really comfy in
that world .. are there even 70 virgins that come with getting it? ;-)
... A lot- because WS-* has standardized a lot more stuff that REST has. However, my point was that even the teeny bit that REST has standardized is sooooo...
Writing as a neutral non-technical observer, I have to state that your points here do resonate, Sanjiva. Certain people on this Group obviously know a lot...
... It's really hard. Take my latest InfoQ article about separation of concerns, for example (motivated by the lengthy "pizza order" discussions we had here...
So, Mark, are you in effect asserting that REST as a disruptive technology could cause, shall we say, career and economic problems for some of the less agile...
... No, you're bang on, Gervas. ... Thanks. But in one very important sense REST is doing just fine; the Web remains the largest and most successful...
... for some of the less agile members of the WS-* community, or am I reading far too much into this? ... to bring REST's virtues to the attention of the...
... Jan's more old school, like me 8-). Our arguments are probably getting old by now. ... Insofar as mashups are "hot" and WS-* is irrelevant to them, I...
... That is the impression most of the times, yes. FWIW, I always hope for lurkers being out there that take something out of these discussions; hopefully to...
... That's why we're doing it, right? ... Not only that, but the fundamental viewpoint is slightly different from the alternatives. You can make REST *look* a...
<MarkB> What is hard about understanding REST, IMO, is that you really need to build up a lot of foundational knowledge in the area of distributed/networked...
... If the words above make others feel like "poor souls", too I should propably take the responsibility - it was me who said that, not Mark :-) Anyhow, making...
<Jan> If the words above make others feel like "poor souls", too I should propably take the responsibility - it was me who said that, not Mark :-) </Jan> My...
... Trust me, having done distributed multi-vendor systems for pretty much my entire IT career, it doesn't help knowing about distributed & networked systems...
... Yepp, exactly. And (that is my personall approach) whatever you need to have the orders participate in processes you stuff into the Atom entries (via an...
... Now you have to explain to me why using ATOM as an envelope is actually different to using SOAP as an envelope, given that the meaning of the transaction...
... Putting my naive hat on and giving my two cents on what I assume and therefore what I'd like to see (noting that I've worked with WS-* much more than...
... Its been part of the problem , data semantics are a challenge, but function semantics are a much bigger one. Agreeing on what data pieces mean is trivial...
... Sounds reasonable. ... Without getting into a type system argument, I completely agree that *strong* typing is a very good thing. I think there's room...
Gah, hit send too quickly... ... Should read... Anyway, Gervas' question was about what can vendors do to assist REST developers, so I included it. Many...
... REST people do not think that. What they (at least I) think is that SOA/WS-* is just Corba/DCOM by another name; that you get the exact same architectural ...
... Sigh. IMO that was a fair characterization of WS-* when it started but that's not what it is any more. If you'd like to see my perspective of how this has...
... It's poorly understood, at least by many. So what? In contrast to WS- *, it is radically different from CORBA or DCOM or RPC. ... I find it amazing that...
... Yes, but recent arguments about REST and WS give the impression of having generated more confusion and conversions. I suspect that what the RESTafarian...
Ah, but, Ronald, you sell something far more valuable as analysts - advice! So, don't you fancy sponsoring such a REST Foundation? Gervas ... software? Or ......
We're just pundits ;) But sure, if the market demands REST, then why not? Right now, the market seems to be favoring Web Services. This is just an observation....
I dare say similar observations could be made about Jini, JavaSpaces and other what I call Orphan Java technologies, which Sun in its wisdom never chose to...
I'm still not entirely sure why WS-* gets lumped into this. Many current specs are based on SOAP/WSDL but conceptually there is no reason we couldn't...
Having problems with message search?
Offline 
Send Email