For what its worth. I like maven sites because they are all (mostly) the same. You know where to find the javadoc, ci, scm, reports, mailing lists, etc.
On Wed, May 26, 2010 at 3:34 PM, Les Hazlewood <lhazlew...@apache.org>wrote: > Gotcha - thanks! > > On Wed, May 26, 2010 at 12:21 PM, Kalle Korhonen > <kalle.o.korho...@gmail.com> wrote: > > Right, we had already agreed to create the symlinks. Fully agree we > > shouldn't direct people to the Maven site but just link to selected > > content from there. The current Maven site infrastructure doesn't play > > well with other site technologies such as wikis so you see various > > solutions to the problem when people are trying to cope with it. The > > Maven site infrastructure is targeted for a rewrite in 3.x. > > > > Kalle > > > > > > On Wed, May 26, 2010 at 12:01 PM, Les Hazlewood <lhazlew...@apache.org> > wrote: > >> I would (hopefully) change it a bit the next time around. Ant changed > >> his vote to +1, so we're still on for a release - I'd hate to have to > >> start over again if not absolutely mandatory. > >> > >> As to the static site, I think generating it is still ok, but maybe we > >> just have symlinks to the reports that people care about. We did this > >> on the old JSecurity website, and it worked really well. That is: > >> > >> /www/incubator.apache.org/shiro/api (symlink) --> > >> /www/incubator.apache.org/shiro/static/1.0.0-incubating/apidocs/ > >> > >> Then our website navigation only points to the symlinks. That way, > >> the static site never needs to be referenced anywhere - it is used > >> only as a report generation and hosting 'implementation detail'. > >> > >> This does require the extra work of maintaining a download page like > >> we did previously (http://www.jsecurity.org/download), but I think it > >> is well worth it - it is extremely easy for people to get the > >> information they care about, and the static site doesn't have to be > >> referenced again. > >> > >> I just haven't set this up yet because we don't have an official > >> release yet. Once that is (hopefully) approved, I wanted to add these > >> links. > >> > >> Thoughts? > >> > >> Les > >> > >> On Wed, May 26, 2010 at 11:42 AM, Kalle Korhonen > >> <kalle.o.korho...@gmail.com> wrote: > >>> On Wed, May 26, 2010 at 11:08 AM, Alan D. Cabrera < > l...@toolazydogs.com> wrote: > >>>> Web site documentation evolves. It should be decoupled, vote-wise, > from our > >>>> release process of artifacts. > >>> > >>> Fundamentally, I don't care too much about it either way, but the > >>> point is that specifically the site artifacts (and it's less about > >>> documentation than the reports - dependencies, quality etc.) do not > >>> evolve since the reports are relevant for that specific revision of > >>> code only. Les asked me to configure the pom so that the the site can > >>> be archived so I did. The site can be seen as supporting documentation > >>> for the release and it may be easier to view the produced web pages > >>> rather than browse the raw pom file, especially for people not > >>> familiar with Maven. Since the site's not supposed to evolve and any > >>> desired change in the site would require changes in the pom and the > >>> tag, at least the given meta-data for that release should be correct > >>> and reviewed and thus subject to a vote, don't you think? Your > >>> statement that it should be decoupled contradicts with the process > >>> that Maven (the project) has put forth so I assume your statement is > >>> your opinion rather than Apache's official position on it. Is that > >>> correct? Finally, do you think we should vote on 1.0.0 again, > >>> excluding the site or just perhaps word the release vote email > >>> differently the next time around? > >>> > >>> Kalle > >>> > >>> > >>>> On May 26, 2010, at 11:05 AM, Kalle Korhonen wrote: > >>>> > >>>>> On Wed, May 26, 2010 at 10:57 AM, Alan D. Cabrera < > l...@toolazydogs.com> > >>>>> wrote: > >>>>>> > >>>>>> I think it was a mistake to generate it and include it in the > release > >>>>>> vote. > >>>>>> People will focus on this non-release artifact. > >>>>> > >>>>> Perhaps. But considering that the site will be versioned and > archived, > >>>>> it's a secondary artifact of the release and deploying the site is > >>>>> according to the Apache/Maven release best practices. I don't mind if > >>>>> we need to make minor adjustments to the content to make everybody > >>>>> happy; I'd rather pay now than later. If nobody votes against, we can > >>>>> just gather the notes and fix the remaining issues in the next > >>>>> release. > >>>>> > >>>>> Kalle > >>>>> > >>>>> > >>>>>> On May 26, 2010, at 9:48 AM, Kalle Korhonen wrote: > >>>>>> > >>>>>>> Additionally, the static sites will be versioned and archived > unlike > >>>>>>> the wiki, where there's in principle just one version of it. > >>>>>>> > >>>>>>> Kalle > >>>>>>> > >>>>>>> > >>>>>>> On Wed, May 26, 2010 at 8:38 AM, Les Hazlewood < > lhazlew...@apache.org> > >>>>>>> wrote: > >>>>>>>> > >>>>>>>> Yeah, its mainly just for the auto-generated reports - much easier > to > >>>>>>>> let Maven generate and upload the site automagically than us > having to > >>>>>>>> piecemeal it and do each one individually. > >>>>>>>> > >>>>>>>> On Wed, May 26, 2010 at 7:46 AM, Kalle Korhonen > >>>>>>>> <kalle.o.korho...@gmail.com> wrote: > >>>>>>>>> > >>>>>>>>> On Wed, May 26, 2010 at 7:06 AM, Alan D. Cabrera > >>>>>>>>> <l...@toolazydogs.com> > >>>>>>>>> wrote: > >>>>>>>>>> > >>>>>>>>>> Why do we generate a static maven site if we have a perfectly > good > >>>>>>>>>> one > >>>>>>>>>> driven by the wiki? > >>>>>>>>> > >>>>>>>>> For javadocs, info & quality reports and since it's simple. > >>>>>>>>> > >>>>>>>>> Kalle > >>>>>>>>> > >>>>>>>> > >>>>>> > >>>>>> > >>>> > >>>> > >>> > >> > > >
