The Designer team has been disconnected from the whole Eclipse CDK plug-in effort until fairly recently (our own fault due to poor communication issues).

Now that I'm aware of the effort, I think we need to make some changes. At a minimum, the CDK efforts needs to be pulled into the radar of the other Eclipse tooling efforts (JBT and possible JBDS).

Since this is an Eclipse tool, it "belongs with" all of our other Eclipse tools in JBT regarding building, testing, packaging, and delivery.

Even from a community standpoint, having this project hosted via SourceForge somewhat hides it from that community,

not to mention the JBT developers concerned with delivering a suite of tools that need to adhere to a number of guidelines to achieve a necessary level of quality and consistency. Then there's also the consideration of what level of integration we may want with other tools without our suite.

Finally, we've also envisioned other possible tools for Teiid that would not fall under the umbrella of either Designer or CDK.

Therefore, I propose that we change several of our Designer components to be more generically "Teiid Tools" components. We'd want the following: * A new URL at www.jboss.org/teiidtools.html

* Either a new JIRA project (something like TEIIDCDK) or changing the existing TEIIDDES project to something like TEIIDTOOLS * As with the JIRA change just described, a similar change to the mailing lists

* A new home for the CDK code, either in its own project that gets pulled into the JBT build/testing/packaging framework, or within the JBT SVN.

* A new Teiid Tools project page that incorporates documentation, downloads, etc., for Designer, CDK, and whatever other Teiid-related tools we invent.

* Sanjay to be made a committer on the appropriate SVN.

* A discussion with the JBT team concerning the CDK, when to include it in a JBT/JBDS release, etc.

On Mar 23, 2010, at 4:05 PM, Ramesh Reddy wrote: > "teiid-dev" mailing list is always used for any communication on this > project. If you are just ignoring the emails that is not poor > communication. That's not blaming you guys for lack of communication, I'm talking about general internal communication issues we had early on, none of which I'm going to discuss further in this forum. > IMO, just because this is tooling means we need lump any/all Teiid > tooling into single project. Connector developers kit is a new effort, > this has nothing do with what Designer does. Yes, this does need to be > under the umbrella of Teiid or Teiid tools and it is. I'm not suggesting it all needs to be under one project, that just one possible option. Considering the use-case and expected user persona, I would expect them to remain separate. But there ARE common concerns for all tooling when it comes to builds, testing, integration, packaging, etc. > Yes. Once this makes into JBT. We are not there yet. Once we request to > be included in JBT then we will do any necessary work. This is an effort we initiated, not just something that happened to appear on its own in some other the community. Waiting until everything is done before you integrate is like waiting until the end of a development cycle before you attempt integration. That always becomes problematic, throws off schedules, etc. We know from experience that, if we know these things will eventually go together in the same environment, we should work with that in mind from the start, thus allowing for things like milestones... > I dis-agree. Fedora is pulled from hundreds of places. Also, I have > mentioned in my emails that this is temporary place holder to get the > basic framework worked out. This will be pulled into main line Teiid > under different "folder" or into a separate svn location. Based on the > facts at the time we felt using source forge was the best option to > manage. I shouldn't have stated it the way I did. I both understand why this decision was made initially, and agree the hosting location itself doesn't matter. I'm really trying to point out that we need to officially acknowledge and link to all of this in the right places within our world, and it all needs to get inserted within our processes for building, testing, etc. > All I can say is Teiid community needs to seek out to JBT community if > this needs to be part of JBT. As part of that integration we have do > that work. I see this as parallel to the Designer tooling rather than > part of it. I was never suggesting this is part of the Designer effort, but should be part of an overall tooling effort. > Please share. I'm talking about efforts around Web/XML services that involve integration with Teiid and other JBoss projects. We don't have anything concrete enough to even throw out to the community yet, so expect "something" will be coming in the future. > Originally Teiid team proposed that, http://teiid.org be the main > umbrella site and others spawn from it. That is still a viable option if > we are willing to put in the work. I agree pulling Designer under your umbrella would work just as well. > Yes, once we have a somewhat working plugin, we do need this. That makes no sense to me. Why not do it now when others might contribute up front towards the direction that plug-in is going? Is there no roadmap? Aren't there design considerations? > Absolutely. IMO, We need to release this in tech preview first. Then we > can talk about integration into JBT/JBDS. I have no idea what this means. How are you going to deliver this tech preview BEFORE talking to the JBT group? Just offer it as a set of plug-ins that you hope works with JBT and doesn't break or conflict conceptually with anything else? I fear you might be overestimating the power and/or simplicity of Eclipse's plug-in architecture... Bottom line is, if we have any plans at all regarding integration with JBT, we should be talking to them now, not later. Thanks, JPAV