xml-soap-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Glen Daniels <gdani...@allaire.com>
Subject RE: Potential F2F design session for 3.0?
Date Fri, 10 Nov 2000 16:58:34 GMT

Comments inline:

> -----Original Message-----
> From: Sam Ruby/Raleigh/IBM [mailto:rubys@us.ibm.com]
> 
> James Snell wrote:
> >
> > Mandate given. Seems like this project is now about
> > SOAP & XP, or even more broadly, it is about
> > providing a Web Services development architecture
> > that is capable of supporting multiple XML-based
> > protocols.  Would everyone agree?
> 
> If XP is the successor to SOAP, then there isn't a strong 
> requirement for
> the next code base to continue to support SOAP.  Don't get me 
> wrong - It is
> a darn good idea, but not a mandate - hopefully this 
> distinction is clear?

Actually, I disagree with this.  SOAP exists now, and more importantly, it
(i.e. SOAP 1.1 technology) is going to be the base for a lot of stuff in the
upcoming .NET platform, which like it or not is going to be installed on
some huge percentage of the installed base of machines.  Although I do
believe Microsoft is committed to the W3C effort to standardize XP and thus
improve upon what SOAP has to offer, I also think it would be ill-advised to
undervalue SOAP support in our engine.

I think (as will hopefully be validated in the coming weeks) that if we have
a good abstract model for web services, we should be able to pretty easily
plug-and-play with different wire-formats, XML protocols, and data
encodings.
 
> I am a bit uncomfortable with expanding the mission as 
> broadly as you state
> (and previously as Sanjiva described) at this time.  What you 
> describe is
> where I see this project likely being in 6 to 12 months time, 
> but I don't
> believe that now is the time.  This is for two reasons.  The first is
> Apache politics, and the second is the quite valid (IMHO) 
> rationalle behind
> this.
> 
> Essentially, this project is still in an "incubator" status.  Open
> development isn't about documenting intentions or 
> contributing large and
> complete code bases, but about continuous and incremental 
> improvement by a
> diverse set of contributors.  Commits over the last two 
> months have been
> rather anemic.

I think a good deal of that has been a result of the steep learning curve to
the system, which we hope to flatten with a cleaner architecture and some
good documentation for the upcoming versions.  Once it feels to people like
there's a solid, flexible framework on which to build, I'm hoping that the
commits (and the to-do's, and the good project management ideas) will start
to flow a lot smoother.

> If that comes to pass, one could image a point in time where 
> there is an
> Apache Project Management committee for Web Services at par with the
> current ones for Jakarta and XML.  The members of this committee would
> largely be selected from the committers on this project who have
> demonstrated through their actions an understanding of what open
> development is all about.

Sure, I could see that, but the space isn't so big that I think we need an
infrastructure like you describe to get good stuff accomplished in the near
term.  And truthfully, I think to make a really usable system in this space
necessitates having at *least* a well-defined and thought-out plan for
integrating client tools, service descriptions, and flexible deployment of
extensions, if not the vastly preferable route of supplying real working
code for these services.

The current interoperability problems between the SOAP engine and the
Microsoft tools are a good indicator of the need for the kind of stuff we're
talking about.

> - Sam Ruby
> 
> P.S.  I'm getting really bummed that I can't make it to the 
> design session.
> Sound like a good meeting.

Hopefully! :)  I wish you could attend as well, Sam.

--Glen

Mime
View raw message