(sorry if this message appears multiple times, I had some mail problems)
Hi,
like (probably) many others, I was surprised by the number and significance
of the changes from PFD1 to PFD2. And I also have to admit that I'm not very
happy with the changes themselves, nor with the process that led to them.
Perhaps I am alone in this judgement, but perhaps I'm not; so I'm expressing
my concerns here to find out what the community's thoughts are.
First of all, I consider it a violation of the Community Process to
introduce major changes like the ones in PFD2 in this phase, because - at
least as I read it - the *public* draft is released in several iterations to
give the public a chance to influence it, and after it has reached a pretty
stable state, it's finalized. Basically, I think everybody expects only
minor changes once a spec has reached the "proposed final" state. I'm not an
expert on the rules of the Community Process, so I may be wrong in
a formal sense. But I think in spirit this is true, anyway.
I know that a lot of companies doing consulting started to advise and coach
customers about the possibilities of EJB 2.0, even though it is not (and
wasn't) finalized (at least I know that my company did that). Anyway, a
usable CMP spec (instead of the 1.1 mess) is desperately needed, and for big
projects that are expected to be executed over several years, there seemed
to be little value in selecting an O/R mapping tool and/or to go for BMP if
EJB 2.0 CMP is "just around the corner".
>From a first reading of the new spec, it seems that things have not only
changed from an implementation perspective (I have no problem with that),
but also the functionality has changed (see e.g. the thread on remote
references in this group). How can functionality be excluded from the spec
in a new revision? I just don't get it.
In another related thread, BEA's comments on dependent objects and the
problems (or impossibility) of implementing them correctly have been
discussed. (If it's so hard or even impossible to implement the PFD1 CMP,
how come that several implementations exist?) To me, it looks like BEA and
perhaps other vendors have been able to convince the expert group that -
contrary to what I and probably most expect - the route that is the easiest
for the vendors has been taken, instead of focussing on what developers
of the technology need.
What's more, I wasn't able to find any documented discussion about the
reasons why these changes have been made. I wasn't able to find a public
archive of the comments sent to the expert group and the responses to them.
Does something like this exist? If so, where?
To summarize: My confidence in the standard has been very much shaken, since
I don't know what to do with the current version. Will we see a similar
number of changes from PFD2 to PDF3? And PFD4? At the very least, I would
expect a member of the expert group to comment on the reasons and rationale
for the current changes and finally (!) give some information about when the
spec will be finalized.
Regards,
Stefan Tilkov
stefan dot tilkov at innoQ dot com
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
