RE: [stellation-res] Database Package Audit.

["Mark C. Chu-Carroll" <mcc@xxxxxxxxxxxxxxxxxxxxx>]

On Sat, 2002-08-31 at 21:35, Jonathan Gossage wrote:
> 
> I would be happy to take this on as soon as we have a solid Windows
> implementation and a solid MySQL implementation. Incidentally, I have had
> absolutely no problems with MySQL. All my problems have been with Windows
> specififc gotchas in the workspace package. I am in the process of doing a
> file by file audit of this package as well and should have the results
> either this evening or tomorrow.
> 
> Incidentally, while I believe you will be able to implement the fine grained
> artifact support without too much problem, I suspect you may hit some
> significant usability problems which will derive directly from the current
> level of integration you can achieve with development environments such as
> Eclipse. My belief is that you will need much tighter coupling than
> currently exists and a high level of cooperation with tools like the Java
> editor and the compiler. But we can discuss this later and I could well be
> wrong.

You are absolutely correct. One of the necessary steps for the november 
deadline is tying in to the Eclipse resource model and Java internal
representation. Getting the basic fine-grained artifact support is not
going to be that difficult. To the repository, it's almost a no-op. To
the workspace, it's basically just using the Eclipse Java parser to
identify fragment boundaries and identity. But for the Eclipse IDE
integration, it's a huge change. That's where Jim Wright and I will be
spending most of our time for the next year or so.

For the November milestone, we're going to support a somewhat weaker
integration with the editor than we'd eventually like. It'll work, but 
it won't be nearly as smooth as what we really want. The current Eclipse
editor framework isn't well-suited to  Stellation fine-grained stuff at
all, and we will need to do a huge amount of work to properly integrate
complete fine-grained support. I'm not sure, but I suspect that we'll
need to implement a new editor. (That's a new editor in the Eclipse
sense of the word... so that doesn't mean that we'll be starting from
scratch, but we won't be able to do what we need just by adding
extensions to the current Java editor.)

The current eclipse generation is very weak. It's basically the work
that Jim did when he joined the project to learn how both Eclipse and
Stellation work. Most of it will get substantially rewritten in the
coming months in order to make the fine-grained support work properly.
You'll be seeing some pretty nice improvements to the Eclipse
integration within the next month, as Jim transitions the code so
that it doesn't go through the command-line tools, but talks directly
to the Stellation repository. 

	-Mark

-- 
Mark Craig Chu-Carroll,  IBM T.J. Watson Research Center  
*** The Stellation project: Advanced SCM for Collaboration
***		http://www.eclipse.org/stellation
*** Work Email: mcc@xxxxxxxxxxxxxx  ------- Personal Email:
markcc@xxxxxxxxxxx