On 24/03/2003 14:23 Stefano Mazzocchi wrote:
I don't expect 2.0 to live long after 2.1 is out. There is no reason to.
To be really honest, I'd like to see some facts backing this statement. Not technical facts, but truly compelling reasons to make the switch. If people don't go with the flow, or with xmlform, why should there be a reason to switch?
Tons of it:
1) modularity: you can include in your cocoon, only what you really need, keeping memory lower and lowering the chance of bugs and potential security holes
2) interpreted sitemap: reduces your try fail cycle
3) tunable pipelines: you can have caching/non-caching granularity
4) better proxy friendlyness: improves your speed for static resources.
These are the most evident but there are tons of them if you look at changes.xml.
Anyway, we can't force people to do anything: if they won't migrate from 2.0, we have failed and we should start reconsidering our architectural strategies because our user base is not following us.
Still, given the feedback i received on 2.1, I don't think this will happen, as it didn't happen between cocoon 1.x and 2.x even if the changes were so radical.
The 2.0 -> 2.1 transition will be piece of cake compared to that, trust me.
If there is something back-incompatible, it's a bug and it will be fixed. Nobody should have problems in switching to 2.1 and if they did, we fix their problems because we (and them) *expect* an easy migration.
At that point, there will be only *one* repository for docs and it will live close to the code it belongs to.
Sigh. I don't understand, and perhaps will never understand, what this obsession is with keeping docs close to code.
There is no 'obsession', Steven. I'm wide open to arguments that tell me why keeping the docs with the code is wrong.
Diana's arguments about duplication of effort are good ones, but I pointed out that they are only short-term ones until the transition is finished and should not influence our longer-term vision.
I see three types of documentation involved in a generic software project:
* code-based documentation, aka Javadocs & comments
* reference documentation, which could possibly be partly generated out of the above, and states the exact input/output requirements and behavior of components
Javadocs has one huge drawback: they aren't semantic. If we had semantic javadocs, we could integrate more meaningful javadocs into our own documentation and this would be *so* nicer than what we have today.
javadocs are cool and useful, but they are too developer oriented. Still, if we had semantic capabilities, we could write skins to make it more 'forrest-friendly' for example, and provide more coherent visual semantics. Worth thinking for forrest, IMO.
* 'user' documentation, like 'building new Cocoon components', or 'installing Cocoon on Znorbaf appserver'
I agree the first two are somewhat different from the last one.
I agree blocks & modular builds kinda spoil my picture, but the reason why I think all of these are wildly different, has to do with flow and access trails. Ultimately, one might think docs will be generated using some Javadoc++ process, as some novel language features in c# and Java, and tools like xdoclet are trying to suggest. I'm still one of these firm (silly?) believers that good user documentation is created with a blank editor screen in front of you, and that a decently written document of several screens long might be more helpful than the technically correct, hyperlinked-to-the-max, autogenerated alternative.
Oh, I can't agree more. Believe me.
For some examples, please see:
- http://xml.apache.org/forrest/your-project.html
- http://xml.apache.org/forrest/linking.html
- http://wiki.cocoondev.org/Wiki.jsp?page=DevelopingComponents
- http://httpd.apache.org/docs-2.0/mod/mod_include.html
- http://www.zope.org/Documentation/Books/ZopeBook/current/SimpleExamples.stx
Nicola might be right in me being wrong, i.e. that I'm focusing too much on the Cocoon website. If that is the case, I'm sorry, my view must be warped then.
In the future, i would like block-specific documentation to remain inside the blocks. Everything should remain as close as possible to the code: javadocs, tests, metadata in general even documentation, configurations, avalon roles, block metainfo and what not.
creating a single docu repository is, IMO, a very dangerous road because:
1) it gives a perception that documents are maintained by somebody else. This perception is already here and it's probably my fault and this is causing pain and trouble to many people. I want this to be fixed in order for the process to be more scalable.
I don't see the point in addressing a perception which surely isn't generalized. Some people like to work on docs, and they will, and the entrance barrier should be 'low enough' for them. Some people believe plenty of Javadoc will do the job, alas to be read only by co-developers IMHO. But we can't force anyone of them to become the other - we should support both styles.
I agree.
Making Cocoon user documentation independent of Cocoon itself might be a good first step, that's why the Forrest transitioning is really, really good.
yes
I agree I said something like 'we the doco people'. What I meant was the 'people usually contributing to documentation'. Some of them are developers, some of them not.
I haven't contributed to documentation because I love xml but I bloody hate writing it and I love writing structured text but I hate the fact that wikis are so damn idiotic to navigate.
I want a serious CMS, damn it! with a dead-simple (not xopus!) inline editor on top! and using CVS as a repository! and transforming structured text in xdocs with perfect roundtripping!
And there's exactly where I want to push to go.
My own belly tells me that people will write more and better user documentation if they get some proper playground. Having to worry about code sitting right beside their documents will not bring peace in their minds.
Ok, please, explain to me why the cocoon CVS module is not a proper playground for people writing docs because I don't understand.
Stefano.
