Re: Discussion About Community Contributions to NetBeans

Here are my bits of experience from various angles ...

- The need for an "IDE platform" separately from "platform".
   By this I don't mean an ide cluster, but a grander concept including maybe tiered
   documentation, tiered books and tutorials.

   I've had to build a standalone debugger based on NB/CND and it was tough. There is a
   good separation of API's between "platform" and what's above it. But if you want to build
   something that requires editing debugging etc there's no such thing as "IDE platform".
   I got into the hell, well purgatory, of dependency management.
   Other standlone tools hearabouts have had similar issues.

   Ultimately I think this is what the recent posts by [email protected] and
   [email protected]'s boil down to as well.

- Does API compatibility work in practice?
   This is a genuine question to all and sundry.
   NB has a very well thought out API compatibility approach. There's Yarda's excellent
   "Confessions ..."  book to go with it. Yet I've had coffee-stand conversations
   with people who aren't %100 trusting. I also see Ernie put out new jVi's for
   each release, partly for new features but partly because the old jVi doesn't work
   anymore.

   I suppose the API compatibility promise is solid for the platform level so maybe
   this is really the same issue as "The need for an 'IDE platform' above" where
   if the concept of "IDE platform" were to be committed to seriously then it's
   API's would fulfill the compatibility promise better.

- Motivations ... scratch the itch problem.
   The basic terminal widget is one of the first, if not _the_ first, external contribution
   to NB, so it's very old.
	I know it's not a bona-fide plugin. Also my role is ambiguous. Am I an
         independent contributor or this is part of my job description?
   Yet after all these years I haven't been able to improve it that much.
   It started to scratch the itch of lack of "real" terminals in the JDK. I needed it for
   the dbx cmdline, it needed to be good enough for ksh cmdline editing, more, an occasional
   vi. Playing rogue, fancy multi-colored prompts and running midnight commander are all
   considered extra. So the original itch is scratched and even though I have a lot of desire to
   improve Term I simply can't find the time.

   The next logical thing to do is to create a testsuite to enable others to fearlessly
   work on it. I have a strawman at home but haven't had time to work on it.

- Fear of dropped support - succession planning
   I wouldn't have touched using the NB IDE if it weren't for Ernie's jVi.
   Therefore with every new release I start worrying ... will Ernie come through?
   What if he gets run over by a beer-truck? I'm doomed! No?

   In the business world this is called succession planning.

   Other than jVi (which has consistently been at the top of the popularity
   and/or download charts) what other plugins are curcial?

   Is having a link from NB to whatever repository the plugin is currently in
   sufficient?  Might it not be better for NB to host plugins? Will anyone come?