I spent part of last night messing around with localization issues of Antville. Here's what I found.

• Apache/Tomcat/Helma already handles Unicode nicely. That is, if we set charset to UTF-8 in Helma, the only place where it breaks in our current setup is MySQL, which won't support Unicode for a few years to come, it seems (it's promised for 4.1, and right now 4.0, the complete rewrite of MySQL, is still regarded as unstable). So I got really angry at MySQL and promised myself to investigate into alternatives like PostgreSQL, which has no problems with Unicode. Having one Antville server running capable of handling any character set on earth would be a thing of beauty!
• Antclick will infact support Unicode from the next release on. Hsqldb handles Unicode nicely, and I ripped out a little part of Tomcat parameter parsing and plugged it into our embedded web server. I now have a fully Unicode capable Antville running on my iBook (limited on the display side by what my Mac OS X supports of international charsets e.g. Japanese and Chinese works, but not Arabic. So far the only platform with close to complete charset support out of the box I've seen is Linux)
• I implemented support for setting the timezone for weblogs. That is, by setting a timezone in the weblog preferences, times will not only be displayed in the local language, but also with the right timezone offset, e.g. page that was created at noon Vienna time may be displayed as 8 PM Tokyo time. I still haven't updated the database scheme (WEBLOG table needs a varchar(32) field called timezone/TIMEZONE), but I commited the code into branch called hns_timezone of apps/antville/weblog. It works even without the db patch (although weblogs won't remember the timezone after restart), so feel free to check it out using cvs update -r hns_timezone weblog from within the antville app directory. The reason I checked it into a separate branch was that a) I wasn't sure if this functionality is desired, b) it's still buggy/uncomplete especially in regard to weblog days/dates, and c) it poses some new questions we have to solve before deploying. One is that timezones will currently be used to display weblog lists on the antville root area, which will be confusing, because although weblogs will be listed in the right order, dates will seemingly jump back and forward in random fashion. I think that dates in supra-weblog lists should always be displayed in the same timezone, which could be the server timezone or even a per-user settable time zone, which wouldn't be hard to do. Whatever we do, it is probably a good idea to provide a version of renderTimestamp() which could be called with java.util.Locale and java.util.TimeZone parameters and would thus be independent of the weblog prototype and could thus be placed in global.

Please let me know what you think. I think I'll update this posting later with a few juicy screenshots ;-)

i am currently modifying the skins to output true and validated xhtml 1.0 (transitional, that is). because this is quite an extensive task that affects a bunch of skins and macros (returning html code) i'd like to kinldy ask you to wait a little bit with your next check-ins, at least if skins are concerned.

btw. i'd like to modify the functions that render html output in a way that there is a set of global skins defining the html that will be called from the function with a parameter object instead "res.writing" the whole html data from within the function. is that okay with you?

update: i made a first step with the xhtml migration as well with the new way how html macros (ie. macros returning html tags) are rendered (as described inside) and already checked it into cvs. please be aware that this stuff is still in development and i cannot guarantee whether it's buggy or not (and sorry, i won't be able to continue before tuesday).

I just commited a patch to CVS to get rid of dual skin managing infrastructure in weblogs: "skinmanager" containing the skins and "skins" doing functional management on them. From now on, everything is contained and managed within the skinmgr prototype. The direct benefit is that some code becomes a bit nicer, but the real reason is to make the skin manager less dependent on its surrounding (I was playing with plugging a skinmanager at antville root level, wich is now possible with a few additional hacks, but it's also possible that somebody might want to plug it into a different application altogether.) Let me know if you encounter any problems (in your local installations - it's not on antville.org yet).

PS: An interesting experiment: update to the newest CVS version and then add

skins = mountpoint(skinmgr)

in root/type.properties. Then invoke URL /skins in your antville root. While the skinmanager itself would work, you'll get errors from the security functions and rendering functions which expect a weblog object to be around. If we can solve these errors (not just in a per-case basis by making root look like weblog, but by defining some standard modes of interaction) we'd be able to get true meaningful modularity, I think.

last night i committed a new set of updates into CVS. changes are:

• stories do not change the URL anymore whether they are part of a topic or not; story-URLs will now follow this scheme: [weblog]/stories/[story-ID]. this means that the permalink is now a *real* permalink.
• this problem is solved, which also changed the URL of comments: the new scheme is [weblog]/stories/[story-ID]/comments/[comment-ID]. this allowes unlimited nested comments (i stopped testing after > 100 levels ;-)
• every form now contains an "action"-attribute (i hope i didn't forget one), the action itself is assigned to res.data.action (accessible in skins as response.action) in the appropriate .hac-files.
• changed weblogcount-macro: by default it now only counts the public weblogs, with the optional parameter count="all" it counts all hosted weblogs.
• changed frontpage: now total number of hosted weblogs plus the number of public ones is displayed, the list of last updated weblogs only appears if there are any to show.
• there is now one "delete"-action which allows the creator of a weblog to remove it. i didn't link it somewhere, mainly because i don't know where one would expect this link (any suggestions?)

since private weblogs are not included in the list of weblogs, anymore, i suggest we should change the paragraph stating "until now, xxx weblogs were created..." on the antville frontpage to something like "until now, (xxx - privateWeblogs) weblogs are online..."

i just started from scratch with a new antville installation and it's strange to read "2 weblogs were created" and there is only one in the list.

morever, if there is no public weblog (or no weblog at all), it says "0 weblogs were created and the last updated weblogs are" – and then nothing else...

even if the frontpage says 997 the number 1000 is in the house.

solving this problem will result in a new URL-scheme for comments: every comment will be reachable through [story-ID]/comments/[comment-ID]/action. Now this brought me to the idea that the permalink of every story should be [weblog]/stories/[story-ID] - which could be easy done (btw.: it would make some parts of the code cleaner, and it would solve this problem, too). this means that neither the topic-name nor the day when the story was created is part of its URL. now: what ya think? 's ok?