[dna-dev] Proposal: new sandbox area for DNA

[Randall Hauch]

I've been wondering recently how to distinguish the (Maven)  
subprojects under trunk that are ready for use vs. those that are  
still under development.  In fact, it's not even only the source  
organization, as our release binaries have included all projects  
(regardless of their state).  I've tried to note the status in the  
documentation, but I think that is insufficient and we need something  
that is more formal, obvious, and intuitive for users.

The 0.3 release included several projects that were not really "ready  
for use":
dna-sequencer-cnd
dna-sequencer-esbMessage
dna-sequencer-jbpm-jpdl
dna-connector-svn

And since the 0.3 release we have started (or will so be starting)  
several new subprojects:
dna-connector-store-jpa
dna-connector-jdbc-metadata (soon to be created)

So, here's my proposal:  we'd create in SVN a new "sandbox" directory  
under the "trunk" directory, and in this area is generally where  
subprojects would live until they are "mature enough" to move outside  
of the "sandbox".  In some cases, "mature enough" means that they are  
indeed complete, while in other cases they may be far enough along and  
there's commitment to finish them before the next release.  I'd  
anticipate that we'd be flexible on what "mature enough" means and  
decide on a case-by-case basis as a community when a sandbox project  
is "mature enough" on a case-by-case basis (although ultimately the  
project lead would have the final say).  However, sandbox projects  
would be excluded from the binary release assembly (but maybe included  
in the source assembly?) and would not be mentioned in the Getting  
Started document, and all mentions in the Reference Guide would be  
moved to a new chapter covering the sandbox.

Also, with this proposal, I'd think we would still keep all of the  
sandbox projects in our development environments (unless other wise  
requested by a project's developers).  And these sandbox projects  
would still included in the Maven builds and Hudson CI jobs.

Here's a strawman list of existing extensions that could be moved into  
the "sandbox", based purely upon my understanding of where these  
projects are and the plans to complete them by 0.4:
dna-classloader-maven (complete, but to be of any use it really needs  
additional components, like perhaps a Maven or SVN connector)
dna-sequencer-cnd
dna-sequencer-esbMessage
dna-sequencer-jbpm-jpdl
dna-connector-jdbc-metadata (soon to be created)
dna-sequencer-jdbc-metadata (actually, this is to be removed)

These projects would stay in "trunk", since they are planned to be  
completed for 0.4:
dna-connector-svn
dna-connector-store-jpa

Forgive me if my understanding of the state of these projects is  
wrong.  If that's the case, please speak up and correct me.

So, what do you think of this proposal?  Please weigh in with your  
thoughts and opinions about the proposal and the suggested list of  
"sandbox" subprojects.

Best regards,

Randall
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/dna-dev/attachments/20081123/20577752/attachment.html