[Rets-dev] On Reference Implementations

[Matt Lavallee]

Well, Steve, you say that, but during the call the spoken goal was to
have standards _accepted_ during the April 18th meeting.  Again per the
call, the indication was that there were currently two implementations
in progress (only one of which was NAR-backed and would be public) and
they shared a common codebase from previous work.  No one but those
working on them has seen them.

The payloads are just getting their first public review, and there are
so many test cases to be done (schema validation, payload validation,
transport validation) that I don't see how it's possible for them to be
ready in one month.  Not to mention the ironing-out of so many type
conversion issues (which I also raised) that I don't foresee two
_different_ reference implementations performing an actual transaction
any time soon.

-Matt

> -----Original Message-----
> From: Steve Verba [mailto:sverba at adelphia.net]
> Sent: Friday, March 16, 2007 7:57 AM
> To: 'Matt Lavallee'
> Cc: rets-dev at rets.org
> Subject: RE: [Rets-dev] On Reference Implementations
> 
> Currently we have been operating for several years under exactly that
> position; no major change proposal or new version of the spec can get
> a
> final approval unless there are reference implementations - two in
> fact.
> 
> There has been debate as to whether to relax this or modify it, but
> nonetheless that has been standard operating procedure for some time.
> 
> Also worth noting for this thread is the fact that of the three real
> estate
> standards (RETS, MISMO, OSCRE/PISCES), RETS is the only one with a
> large
> body of open source tools, reference implementations, compliance
> checking
> code, POC's, utilities, etc. To be more precise, the other two
> standards
> have virtually none - a testimony to this community and to CRT and
> NAR,
> given that the MISMO and OSCRE communities are larger.
> 
> Steve
>