[Moin-user] Wiki for document management

Tim Bird tim.bird at am.sony.com
Tue Oct 7 14:37:17 EDT 2008

Thomas Waldmann wrote:
>>   * editing a large document with MoinMoin is awkward -- we're all
>>     jealous of MediaWiki's "edit a single section" feature.
> Yes, that is a quite often wanted feature and we are aware of that.
> But I could hold myself back from implementing some dirty hack for this
> because we'll likely have a much cleaner way in the foreseeable future:
> Bastian Blank did some parser/formatter DOM refactoring project in
> summer of code 2008. This basically gives a parsed moin document some
> structured representation in memory - and I think this will be a good
> base for implementing section editing and similar features.
> Merging this stuff into main branch will take a while, though, because
> we first need to merge the work of other SOC projects:
>  * FCKeditor upgrade (merge done, release as 1.8)
>  * WSGI refactoring (merge done, release as 1.9)
>  * Storage refactoring (merge + quite some related work todo, release as
> 2.0)
>  * DOM refactoring (merge + much related work todo, release as ?.?)
>>   * no concept of review/approval
> There are quite some people dealing with such stuff / thinking about
> using moin in such an application.
> Pity that nobody of them managed to start a discussion/facts page about
> this on the moin wiki yet.

In addition to being a long-time MoinMoin user, I have written
my own wiki software (hasn't everyone written their own wiki
engine at some point? ;-)   I implemented section editing, and
seriously, it only took about 2 hours.  I always thought I should add
it to MoinMoin, but never got around to it.  Admittedly, it was probably
in the category of what Thomas would describe as a "dirty hack", but
it works fine, and could serve as a stopgap until some more comprehensive
solution emerges.  My wiki doesn't support WYSIWIG editing, but the
algorithm should work well for most cases. (The problem
would come with attributes with effects that cross section headings.
In most cases, I think those would be errors anyways.)

IMHO a 95% solution would still be useful here, but I don't want to
work on something that would be thrown away soon.
Would it be worth posting the algorithm for review/discussion?
 -- Tim

Tim Bird
Architecture Group Chair, CE Linux Forum
Senior Staff Engineer, Sony Corporation of America

More information about the Moin-user mailing list