Surely Tom has been off-hook w.r.t. Cocoon for much too long, and we all should blame his bosses for that, but I find it hard to believe "Tom is a Cocoon dumb-ass":http://blogs.cocoondev.org/tomk/archives/000889.html since he wrote "some":http://xml.apache.org/cocoon/developing/httprequest.html "original":http://xml.apache.org/cocoon/developing/extending.html "documentation":http://xml.apache.org/cocoon/developing/avalon.html for it long time ago, while we were still working in the same company, and he, the "Xume":http://www.xume.be/ boys and Bruno were about the only five developers (amongst 30 Java dev'ers) being slightly enthusiastic about Cocoon.

Heh. Saw some pre-announcements for what "Carsten":http://radio.weblogs.com/0107211/2003/04/28.html#a110 has promised to do tomorrow morning. After a slightly turbulent period (we aren't really much used to such things), the Cocoon developers are big-time resonating again, resulting in Carsten planning to do the first Milestone release of "Cocoon 2.1":http://cocoon.apache.org/. Coincidentally, I've just sent out the resolution proposal to the Apache board stating I will replace Stefano as the Cocoon PMC chair, but that's about the last thing you'll hear about it on this blog. I'm happy to be relieving him so that he has more arm and leg space for RTs, and I also hope to free up time and energy for the real coders so that they can have a ball, while I am trying to do the administrative work which a PMC chair is supposed to do. No fluff, just stuff. Funnily enough, I was also voted President of the "XML Users Group of Belgium & Luxembourg":http://www.xmlbelux.be/ on the 7th. Two chair positions. I guess I should take golf courses now. Does anyone know some good service club? Not.

Marc & I will be in London at XML Europe on May 5-7th, and the idea of meeting up with local Cocoonies has already been mentioned on the mailing list. Locals are hereby cordially invited to suggest meeting places for a snack, drink and some GettingTogether - we'd propose to meet in the "hotel lobby":http://www.xmleurope.com/2003/hotel.asp on the 6th around 8PM local time, and see where we get from there. Please leave a comment if you're interested. *Update*: meeting location has changed to central London, the "Slug and Lettuce":http://www.slugandlettuce.co.uk/directions/soho.htm in Soho. 8PM, May 6th, be there. :-)

In finishing my little overview presentation on Cocoon for XML Europe, I realized that already, there's way too much stuff inside Cocoon to condense into a 45' presentation. I tried to stick to a maximum of 30 slides, which already is a lot for 45', and could barely glance over the most important and popular components and frameworks inside Cocoon. As much as I'm strong believer in manufactured serendipity, it was still a nice surprise to see "Tony":http://manero.org/weblog/ adding a nice "Getting Started with Flow":http://wiki.cocoondev.org/Wiki.jsp?page=GettingStartedWithFlow to the Wiki. Lots of shared neurons in the air. I like that. There was me cleaning up and sysadmining the Wiki, while Tony adding and editing this great concise sample, immediately followed by Chris Oliver -- Cocoon's current Flow Master -- check-reading, commenting and correcting, and "Bertrand":http://codeconsult.ch/blogs/bertrand/ cleaning up: a thoroughly peer-reviewed document in "only 10 hours":http://wiki.cocoondev.org/PageInfo.jsp?page=GettingStartedWithFlow of internet time.

Stuff I gathered lately: * "logging":http://www.red-dove.com/python_logging.html, a JSR47/log4j-inspired Python logging package * "Cheetah":http://www.cheetahtemplate.org/, a templating engine for Python Stuff I'm missing out: * a webapp development framework that integrates well with Apache, without requiring mod_python or the installation of a standalone service, while still offering reasonable speed, higher-level constructs like session persistence and hopefully some Py/db mapping * Komodo depending on a standalone Python install, and not on his own copy

Via Brian (via various others): Hello Jesus!

"Stefano":http://www.betaversion.org/~stefano/ now has a "weblog":http://www.betaversion.org/~stefano/linotype/. Cool! Even cooler: "live updates of blog entries on a worldmap":http://www.brainoff.com/geoblog/.

"Matthew":http://radio.weblogs.com/0103021/2003/04/20.html#a968 is wondering why I bought an Xbox instead of a Nintendo Gamecube or Sony PS2. Being a proper geek when it comes to gadgetery, I spent some thinking about this, but not too much. I own and have owned a decent array of MS peripherals, like gamepads, mouses and the like, and I must say the quality of these is generally good. Also, the Xbox fits in a larger strategy of an important and successful digital home entertainment provider, with some *real* knowledge about P2P networking, identity services and the like. Given the "success of broadband internet services in Belgium":http://www.strategyanalytics.com/press/PR00035.htm, I believe online console gaming will become big, and you need a company like MS to pull through. While they have been slow in adopting internetworking in their service offer, for sure they have much more experience already then Nintendo and Sony. Besides, the advent of acquiring "DOA":http://www.tecmoinc.com/games/doax.asp surely appeals to me. ;-)

"Brian":http://brian.behlendorf.com/ "seems surprised":http://blog.neoteny.com/brian/archives/004965.html about the amount of reaction when people started discovering about "his blog":http://blog.neoteny.com/brian/. Let's see if I can spread the word further.

Without much further ado, and probably against dr. Spock's advise, we acquired an Xbox yesterday, together with the DVD playback dongle/remote, a second controller, and three games: MechAssault, MicroMachines and Splinter Cell. I'm still on the lookout for a game for my youngest boy, age almost 5, who finds the controls of MicroMachines a bit too difficult. Anyone an idea? Other cool things one should do with a newly acquired Xbox?

As much as my fingers were itching to stop trying and write one of my own, I succeeded in installing a quite buggy (still) version of a Python rewrite of the ugliest mailing list archive in the world. Enjoy.
Lessons learnt:

• It is fairly easy to mess up signature dependencies of required modules if there's no clear indication what version of Python is exactly required.
• Everybody seems to invent his own templating tricks when using Python.
• ezmlm might be powerfull and trouble-free stuff, but it is really hard to find a decent web mail archiver which runs directly on the Maildir archives. Too hard.
• I'm normally not used to work with open source code with no community around it. It is scary to work with code coming out of some obscure SF CVS module, with a mailing list archive of only 15 messages over 1.5 years. Commit messages, that is: I found no trace of real, human communication. :-)
  

This is so very much Cluetrain. A Sun peep providing an entire on-line J2EE training curriculum for free, certificate, exercises and handouts included. Kudos!

Yeah.

Today will be a sunny day, since new life is beaming rays of light into our hearts. Many, many congratulations to Katleen and Rik, the proud and fresh parents. Rik already had that familiar sleep-deprived tone in his voice this morning on the phone. :-)
Let the first weeks of nights and days who fade into each other, into a stream of consciousness filled with diapers, weird sleeping habits and lots'o emotional moments begin! Enjoy!!
(Does it show I'm really excited about this? It must be because we are gradually leaving the baby period behind, with our youngest offspring having her first anniversary next month ... ;-)

... And the winner for this and some future years of overhyped, underspecified and hypermarketed technoglut is 'CMS'. You do have a CMS in-place, I hope? And it is industrial-strength and future-proof, I hope? In the wonderful world of CMSes, people seem to have forgotten what really defines a framework and a product.
In researching a free or low-cost solution for some website I host but don't want to spend too much time on (by sharing content administration effort with some others), I encountered way too many so-called CMS tools which, while looking nice on first sight, all seem very much targeted towards solving a single goal, i.e. rendering the website they've orginally been designed for. Also, most if not all of them post requirements which make setup in a shared hosting environment insecure at best, and impossible in most cases.
To be continued.

Sometimes, you wonder why you shouldn't be better off just doing your own thing, personally and business-wise, and leave that community thing aside. That's what you get when seeing several things happen at the same time:

• Some freeloader offloading all Cocoon Wiki content using HTTrack with at least 8 threads working in parallel, reducing the accessibility for anyone else because of JSPWiki trying to feed that guy as fast as possible. Thanks, 211.72.53.138. I hope you are happy with all the old revisions of every page, too, since your stupid offline reader also fetched those. I assume I should now spend half a day trying to exclude people like you from downloading the Wiki, configuring Apache and such. Tell you what? Even though I try to do my fair share of offering the community with physical resources, I still have to make a living. It's nice to see you think bandwidth and server space are free.
  
• When trying to be a better community citizen, being told to piss off and RTM.
• Seeing no interest in the community you love to live in to solve outstanding organisational issues, without anyone stepping up the plate and telling you what exactly the problem is.

Of course, some people care enough to make my day. Thank you, you-know-who-you-are, for being persistent in reminding me how to be a good citizen, and to be polite in your enquiries.
Update: some other fellow Cocoon committer asked to host his blog alongside the others. My mood is slightly better now. :-)

I've been messing around with w.bloggar after hearing success stories from Andy. A new release should be there in a few days, hopefully adding WYSIWYG/rich text editing. Much more to my joy however, I also found out Mozblog is now again happily coexisting with Mozilla 1.3 and Movable Type. Happy happy me. :-)
Still, the URL drag and drop I was so keen of is not working ATM. But at least I now can use Mozblog again to complain about this not working. ;-)

Sam is considering remaking Gump using Python, and refers to Cheetah as a nice templating system. Stefano is on a quest for a non-XML variant of XSLT. Bruno is progressing on his form framework, and was questioning what the glue language should be between a form instance and its visualization. Come to think of it, I'm wondering why XSLT, XML and templating approaches cannot coexist. XSLT has its role for out-of-band, declarative transformation descriptions of a mostly static XML document into some other markup. As such, it has only access to the document content itself, and not per se to the transformation context. This means one needs to add an eventual transformation context into the document stream if its context needs to be considered upon transformation. OTOH, when looking at Cheetah, I'm charmed with its simplicity and unobstructive syntax:

#for $client in $clients

#end for

$client.surname, $client.firstname
$client.email

I can imagine myself feeding an XHTML form snippet through some FormTemplateTransformer:

  #for $color in $myForm.colors[]
    
  #end for

which translates this source document into:

This document can then be piped through a skinning XSLT stylesheet to provide you with the look & feel you want. Of course, all of this could be done with XSLT too, but a template approach might be much more intuistic here since all of the thinking required to build up that nice form is on one screen. An XSLT-like template language might provide us with a familiar syntax, and when freed of its side-effect-freeness, might also perhaps provide us with access to the transformation context. From there, we can easily merge context data (the form instance field values) into the document stream.

Netcraft has posted an interesting new report on the usage of servlet engines. Interesting because it confirms my impression that Tomcat is used in production more often than one would suspect, and also that Bea, who used to claim developer's top-of-mind as the container for industrial-strength serving, comes in at a measily 4.6%, after IBM and Oracle.
It's good to see the small-scale, open source or low-cost containers heading the pack: Tomcat, Resin, Orion in its Oracle outfit, and Jetty are all there.

... so what have I done today? * brought my little not-very-but-persistently-ill baby-daughter to the hospital for a session of thorax X-rays * went to the drugstore afterwards for medication and aerosol machinerie * dropped a letter at the postoffice with a reimbursement claim for one specific expensive medication (Belgium has an amazingly inexpensive public healthcare system, still you need to fill in some paperwork sometimes) * drove to the office, discussed with Bruno whether the Cocoon form framework he's working on should return plain strings or seriously-typed objects into the environment after validation (the latter) * had a hamburger meal with a particular nice and long-lasting customer discussing some interesting nice new project (CMS) * returned to the office and sifted through emails and blogs (infoglut! infoglut!) * endeavoured into some long-postponed Forrest hacking (making the search widget configurable for intranet applications) * got sidetracked by Leo who was happy with Forrest building nicely with Gump * got sidetracked even more with Sam explaining we were cheating * ... trying to be a friend of Gump, I tried to correct the Forrest Gump configuration * finally learned how Gump's descriptors work, with Stefan helping along * messed around with cocoondev's Gump install and unearthed some permission problems * did some more blog reading * laughed our asses off with the particular nasty but sincere "customer reviews":http://www.amazon.com/exec/obidos/tg/detail/-/0971661901/ref=cm_rev_all_1/103-7863522-9227803?v=glance&s=books&vi=customer-reviews of some very bad Struts book we acquired two (!) copies of more than a year ago * ... and now, I'm departing for a meeting of a group of friends supporting some great people who have been living abroad for many years, helping poor farmers in El Salvador - they've come home after 6 years and it will be the first time I'll be meeting them * later this evening, drive home and hopefully find my family fast asleep While it was a packed day, already it seems like on of these days drowned in nothingness. Do have a nice weekend, I'll try and do, too. Things I haven't done: * making the Forrest search box configurable * do a long-postponed update of the Forrest DTDs * find time to chat with my beloved wife

We are discussing forms ATM, and here's my continuing list of bookmarks: * "http://www.pengoworks.com/qforms/":http://www.pengoworks.com/qforms/ * "http://www.brics.dk/~ricky/powerforms/":http://www.brics.dk/~ricky/powerforms/ * "http://www.peterbailey.net/fValidate/site.php?page=index":http://www.peterbailey.net/fValidate/site.php?page=index * "http://www.dithered.com/javascript/form_validation/index.html":http://www.dithered.com/javascript/form_validation/index.html * "http://sarissa.sourceforge.net/":http://sarissa.sourceforge.net/

Besides "Trijug":http://www.trijug.org/, there seems to be also some cool "TriXML":http://www.trixml.org/ group over there, which hosts even "cooler events":http://www.trixml.org/confindex.html - I'm missing out on some "bearded guru":http://www.intertwingly.net/blog/ however.

I'm posting this from inside the newsclient of Mozilla using the powers of nntp//rss. I'm using the Blogger API, the Metablog XMLRPC API didn't seem to work. So far, I'm quite happy with nntp//rss, although it focuses a bit too much on a single user setup. I'm running Hep already for four users, and while the CVS version of Hep doesn't offer the same out-of-the-box feeling nntp//rss does, it sure would be nice if server-based RSS aggregators start to support more than one user, with a centralized control of how often a feed is fetched. nttp//rss does support ETAG as far as I see, which is of course cool.

It has been "blog-talked":http://www.intertwingly.net/blog/1316.html about already in extenso, so this is just my quick referrer to a "must-read":http://cyber.law.harvard.edu/people/jmoore/secondsuperpower.html for those who haven't found out about it already.

Let's hope tomorrow will be better. * I spend the longest time ever (3.5h) on composing a single mail which I should have not sent * I had a fight with Stefano over stuff we shouldn't fight about * The Cocoon PMC chair has suddenly become vacant * In order to fill the vacuum, I self-proposed myself * I "criticized HEP":http://blogs.cocoondev.org/stevenn/archives/000835.html but "Abe's reply":http://www.fettig.net/index.cgi/2003/04/#Hep_tour_2 got me drooling I'm off for a smoke and a good night of rest.

"nntp//rss":http://www.methodize.org/nntprss seems like a candidate for ditching HEP which is a bit undermaintained.

Reminder to __self__: "Unison":http://www.cis.upenn.edu/~bcpierce/unison/index.html

I just pushed a "new version of our company website":http://outerthought.org/ live. Right now, I deeply admire the skills of a copywriter: being able to write on a continuous basis, without major writer's blocks, must be really, really hard. I've been composing the content of the new site over a many-weeks-long ordeal of lack of inspiration mixed with bright moments where text was composing itself, and I found out I generally prefer to work on one document at the time. Anyway, I'm glad it's over and I hope the underlying message of services, who we are and what we do is much more clear in the "new website":http://outerthought.org/ than with the "old one":http://outerthought.org/cocoon/.

Cheers, "Marcus":http://blogs.cocoondev.org/crafterm/!