On Sat, 18 Jun 2005 13:59:38 +0200, Ludovic Dubost <ludovic@xwiki.com> wrote:
Currently we are maintaining compatibility with IntelliJ and Eclipse + building with ant for the automated build. We are looking at maven or maven2 for future automated build. We have an open source licence for XWiki developpement for IntelliJ. I still prefer IntelliJ because of productivity reasons (I even payed my licence before they created the open source program).
How do we download an open source version of IntelliJ?
If we were round-trip engineering then there would be a question to discuss about diagram import/export between tools.
Stupid me. We have the source code and that can squirt the format of any modelling tool. We don't need to import or export diagrams from different modelling tools which should (joke) all produce the same result from the same source. We just have to decide how to share and use them. Do you have a preference between publishing intellij and or eclipse models or exporting them to xwiki as graphviz or jpeg or what? __DevUC__1 Publish UML models If we want to import other people's models then we would have to put both intellij and eclipse models in subversion. The common source would keep them in synch. If round-trip is adopted then all developers
You're current class diagrams are jpegs. We have Graphviz as a built in tool.
WDYT?
I've very open to that.. I don't have much experience on this so I would need help setting up things like that. Do you have a preference between publishing intellij and or eclipse models or exporting them to xwiki as graphviz or jpeg or what?
I'm currently pushing the javadoc and junit results on the automated build platform:
+1
In this case we probably need some "automated" copy when we release a offline wiki database. Thanks to Jeremi's work, we might be able to package the documentation as an XWiki Application which would mean we could also upgrade them.
Now I think we should move the documentation to the xwiki.org web site in a specific area. Using some XWiki classes, we can mark the documentation articles that are specific to XWiki.com. +1 We can definitively open the default wiki to developer's editing so that people can contribute to it's improvement. We can then have an automated export to XML so that we can release it on ObjectWeb regularly. +1 Documentation is a weak part of XWiki. We are definitively looking for people that can help in this area. I'll continue with use cases then.
I have started listing some of the use cases. They appear to start something like:
__UC__1. Render XWiki Documents __UC__1. Verify Authorisation and Rights __UC__1. Read Documents __UC__1. Access and Modify Documents, Objects __UC__1. Search Documents and Objects __UC__1. Read User and Rights __UC__1. Read/Write Object/Classes
Is that for user doc or developer doc ? It could actually apply to both.. Indeed but it just you package relations from your engine diagram. How it fits with road map, faq, jira etc I'll spend some time working out.
For users then the next step is: comments, attachements, versioning, interact with specific applications
For developers then the next step is: Create forms, create templates, access priviledged API, create macros & functions, write plugins Will integrate these into user, dev and common threads.
There are plenty turning up nearly every day on Jira but they are technical.
IMHO The current (largely implicit) business use case hierarchy doesn't quite seem abstract enough or explicitly and correctly nested. It just seems that way, I have no examples cos I'm talking about a missing level or lack of separation of levels.
Are you talking about the FAQ, Jira, the web site ?
As before just your top level engine diagram.
For the FAQ we start to have a good number of documents. Agreed Indeed we can start reorganizing, nesting, etc..
+1
It's part of the plan to redo the web site (both XWiki.com which should show the business value of XWiki and xwiki.org which should help user's and developer's find everything they need and allow them to interact.
Great.
IMHO The classes seem too large to me and seems to mix model content and model management functions in the same class. Is that unfair?
It's probably not that good to mix that.. That will probably need some refactoring for XWiki 2.0. It's going to be tough to fix that for the first release.
For anyone who might want to work on the text use cases, Scott Ambler has standard templates free for use and download for use cases, test case, business rules, change request etc on his site at: http://www.ronin-intl.com/downloads/ Cheers Jim