mercredi 27 novembre 2002
... with this new toy. Until I found
Anthony Eden free weblog, I was poking around trying to choose what to
install (PHP preferred) to do my weblog. This is an endless trip when
you have no strong goal. Every implementation has its pros and cons, and
eventually, there is no clear cut winner. I've to say that in parallel,
I was thinking about a Java/SWT implementation of a BloggerAPI or
MetaWebLogAPI client. This is the kind of thing that leads nowhere, ideas
go round in circles and nothing can break that. Start to code, start to
code... and release asap: this is the only thing I've seen to release
something else than vapour.
By the way I took the decision to start again to contribute to the
CinemAsie source code. We're slowly going mutlilingual, with a customized
XTemplate. I've added a compilation phase which does the static part of
the translation job, and then the dynamic template takes place. This way
we can keep a manageable repository of translation, with no extra cost
at "runtime". Once this task has been done... the real work begins,
refactor all the site. Too lazy for that, I've to grab some force
somewhere.(Tout venant, 2002/11/27 18:26)
lien permanent
... but with some fun nonetheless. I went to a local store and bought the 3 last albums of Sakamoto Ryuichi, and it's really expensive. I still can't realize that I've spent so much money for those three plastic discs.
Another part of the day was spent trying to find out some standard software classification scheme. After some site, I've found Eric S. Raymond's Trove, which does too much for my little goal: organize our departement software repository, and this is far from being a large-scale archive. By the way, appart from the SourceForge Trove Software Map, whose source code doesn't seem to be available, I can't see any implementation of this thing. Do we need really something like a software categorization thesaurus? I think so. There is a real effort to produce the necessary semantics, and widespread them. Maybe it could be a part of the Semantic Web development, like a proof of concept on an "easy thing". This whole Semantic Web/RDF seems to bother Russell Beattie quite a lot currently. I hope his researchs and reflexions will abbreviate my own ones, dear lazyness 
(Cyberpunk, 2002/11/27 18:26)
lien permanent
Ugo Cei mentionned Spaces today, it can be found here. Obviously, he's not satisfied with the news aggregrator he tried, neither do I. Why has nobody coded a simple, effective and free news aggregator ?(Cyberpunk, 2002/11/27 18:23)
lien permanent
Why do advertisements rule the world? Why .net has got that momentum? Is it revolutionnary? No. Is it free? No. Just because it's marketed by the first software company, not known for its innovations, we have to endure all this advertising, FUD campaign, books, public proofs of love, and soon immolation by fire?
There are plenty of things about which companies should take care of. The rapid development is the norm among Python users. Python is mature, cross-platform, reliable. Your company is devoted to clean development, believe in reusability, readability and maintainability. Did it consider Eiffel?
And the others: Ruby, Rebol, Lisp, Perl... and I'm sure I forget many languages. Each of them gains to be known, and there is documentation, presentation available: the big difference is bucks. M$ has bucks, M$ owns companies by the balls. They have locked their documents, locked their messaging, locked their planning. They have $22,956,000,000 of total annual revenue, and they spend a lot on lobbying.
What should we do toward a company who buys judges and government? Just ignore it, but is it enough? If we talk, it can slip too fast on the FUD side of the Force. It's not worth it. Break the chain of advertisement, keep quiet.(Cyberpunk, 2002/11/27 18:22)
lien permanent
Today a cousin caught a virus on her PC. What more needs to be said? The PC runs Windows 98, the mail client is Outlook Express, period. Thanks M$, you did a great job for all the virii developpers on this small planet. Thanks for providing so much funds for the antivirii companies. Even if they try to convince me I need one for my Palm, I'm not at all: there is no trace of M$ on my Palm, so I don't need an antivirii.
On the other side, I don't think Symantec, Norton, Network Associates make that much money with this kind of software. Why? Simply because if that market was that much profitable, M$ would have crushed those products for a long time already.(Cyberpunk, 2002/11/27 18:07)
lien permanent
I don't know why, today I checked PostNuke's status. Some things amazed me:
- the number of developers on sourceforge went from more than 100 to 13
- I did recognize only 2 names on the developpers list
- the most active developers have vanished
- public CVS browsing has been removed
To be true, it's not a really big surprise. When I left PostNuke (around the end of March), I was angry because I didn't contribute much (neither from an external point of view nor from my personal point of view). There are several reasons for that:
- general atmosphere
- code moving too fast for me
- hypocritical decisions
- I can't concentrate on one thing
I'll treat those points one by one.
1. General atmosphere. Unfriendly. I think that is the word. I don't speak about the individuals, but about this combination of 'hype with PostNuke' + 'war against php-nuke' + "I have the truth". This doesn't invite you to participate in conversations, and thus, you don't feel implicated. And if you don't participate, there are great chances that you don't fully agree with the decisions, and you don't feel like ranting, because the time for arguing is over.
2. Code moving too fast for me. When you come back from your full-time job, you have to read all the dev mailing lists, follow a bit the support forums, read the CVS notices and read some of the changes. After that... well, there is not that much time left. When things calm down, you don't feel like contributing, because you realize you spend almost all your free time on that, and you think it's good to let your brain breathe.
3. Hypocritical decisions. It's the result of the other factors. Only the few people able to work almost full-time on the project can argue efficiently. And given the atmosphere, the reason is lost, and the first to write the code wins. This is not always a bad thing, it can cut endless discussions and allows people to move forward. But this works only if there is a natural hierarchy which is respected, otherwise, grief becomes the norm.
4. I can't concentrate on one thing. I put it at the last place, but after having taken the time to think, it's the main one why I failed to contribute. PostNuke is a big project, with blurry borders. If you don't define straight border, the mind will escape. When you debute in the Bazaar model, it works only if you're able to focus yourself on a tiny point and ignore the surrounding noise. I'm not that kind of people, I like when the boundaries are well defined, and when I'm assigned a specific task. I should have resigned sooner.
Adding all this little things make contributing difficult. What's the situation today? PostNuke has lost all of the core developers and the original Project Manager. Those people have taken the development tree and created Xaraya. The fork is done, essentially because there was no natural agreement on the "patching/enhancing policy", just like when PHP-Nuke was forked to create PostNuke. Work goes on with PostNuke, with a new set of core developers. Xaraya is being improved, probably with less pressure than before. Work has never ceased on PHP-Nuke. Meanwhile Xoops gained momentum. Spip, daCode and Drupal are alive as well.
The situation is in a sense worse than it was two years ago. More incompatible CMS, more choice, more unfinished products. The only thing it advises you is to not choose one and either buy a commercial product or roll your own. I've followed the evolution of PostNuke since its creation, back in 2000, and the goals are simply not met (yet). The product is not finished, and from an external point of view, it has not left PHP-Nuke far behind. Meanwhile, the "nuke community" has grown, fought and chosen several camps, nothing to really worry about. For all the developers involved, it is great that you can find a project that motivates you. For the end-user... well... come back in two years, or resignate yourself to the idea that an open source CMS is bleeding edge.(Cyberpunk, 2002/11/27 18:06)
lien permanent
A thing I always loved in Dan Simmons' Hyperion, is the accessibility of information in the infosphere. A space where the information you need is there and accessible. Currently, the web is far from being that, but it's changing, slowly. I would like to be able to track those changes, and I'd like in 50 years (if I'm alive) to tell that story. I'm not that interested in the underlying protocols, machines, connections, routing, but on what we put on top of that... (2 hours go by)
While my primary intent was to speak about W3C RDF and how it could be integrated into our habits, I've taken a look back on Gopher. Gopher was already almost abandoned when I first knew what a modem was, but today I wanted to check from my eyes what it is. So I downloaded WSGopher, and installed it on my company provided laptop. First surprise: it works. I'm so much used to beta/bleeding-edge/"in the hype" software that this has a very special taste. According to the about box, this version (1.2) is from 1994, Microsoft managed to keep a good backward compatibility (the PC runs Windows 2000). Once you have chosen your home server (I took gopher.quux.org), it's intuitive and very easy to use.
How Gopher is different from HTTP? A Gopher server has a hierarchic organisation for serving documents, and it manages server-side symbolic links across servers. On the contrary, the directories structure on an HTTP server has no meaning. It was primarly used to reflect the filesystem where the documents reside, and now with dynamic content everywhere, few servers are really accessible through URL. Look at this three urls :
http://www.ietf.org/rfc/rfc3118.txt
gopher://gopher.quux.org/0/Computers/Standards%20and%20Specs/RFC/docs/rfc3118.txt
http://www.ixus.net/modules.php?name=Rfc&rfc=3118
The third one is an abomination, it uses the search part of the url scheme to represent a hierarchical navigation and gives us a detail of implementation. This happens because the HTTP servers have been twisted very much since the introduction of CGI, and human-friendly URLs are too often the least concern. On the other side, HTTP was primarily the transport layer for HTML documents, and those documents are built around an hyper-linking facility (navigation, and after, images). The hyper-linking is the core of the relationship HTML/HTTP, with HTTP as a dumb server, whereas Gopher stands for itself. It is better explained in the gopher:// manifesto(alternative link):
Gopher is an infoserver which can deliver text, graphics, audio, and multimedia to clients. Keeping documents "link clean", making linking a function of the server info-tree and not in the doc, layout is kept to its most frugal minimum, and is standard across all docs. No graphic design means its the ideal navigable interface, a hypertext Eden. It gives simplified usage for sight-impaired users, same contents for
wired/wireless, and requires no capital investments in layout and
"design". Gopher is real -- and it was fully functional in 1992, even
without advertisements!
The fact is I felt immediately comfortable gophering, it's so straightforward, like weblogging. The www has become a difficult place to crawl today. I remember (helped by my mail client) Dave Winer saying:
I heard someone say the other day that they don't use bookmarks anymore, they just go to Google and type the name of the site and it takes them there. (And gives them a few more ideas on the way.)
Driving in the car the other day I found myself wishing I had Google so I could complete a thought I was working on. Had to wait till I got to my friend's house. On arrival I asked if I could use his browser. A funny new social construct.
There is something terrible in those paragraphs, the idea that we have no more points of reference on the web, just a big indexer and a dozen of weblogs. Of course Google don't prevent us from having bookmarks and maintaining them, which is... time consuming. We have known three terrible years for personal pages, where some majors free webhosters have disappeared, the survivors putting more and more limits on resource consumption, and the raising of the advertisement. Most of my bookmarks from this time went to the dust bin. Beside me, there is still my faithful books, true references.
That brings me back to the infosphere, a space directly accessible, with easy access to its content. Gopher was an answer, the current www is not, but that doesn't mean it's no good. Simply, we need accessible reference information. Gopher brought us that via its hierarchical structure, the www could possibly bring us that via a better categorization, perhaps with RDF.
side note: once again I had a pretty clear idea of what I wanted to say before I started to write, the result is a mess.(Cyberpunk, 2002/11/27 18:05)
lien permanent
It crashes more and more often, and its a memory hog. But I've other things to code before I try to do a news aggregator.(Tout venant, 2002/11/27 18:03)
lien permanent
Today on Andy Oliver weblog, JSR-14 blows and .NET rocks, Windows Still blows and other happy thoughts :
Engage a feverant Java zealot in a discussion about the short comings of EJBs and they'll tout the advantages of Appservers. Correct them: "I'm not talking about the concept of app servers or CTM, etc etc, its the implementation provided by EJBs that I'm criticizing."
Why talk to ignorant zealots? Do you make your point clearly?
Talk about the RemoteException and the disgusting descriptors[...]
Nothing more to say on the RemoteException? So don't talk about it.
[...]the ridiculousness of the EJB stub interfaces, etc etc. They'll start touting XDoclet and various patterns and workarounds. However its kind of like arguing the logic of darwin to religious fundementalist. Workarounds don't make a crappy technology good.
I'm wondering who is the fundamentalist... And furthermore, is EJB a technology or a workaround?
Regarding Generics:
So back to Java Generics and JSR-014. Well it IS just syntax candy. In rabidly trying to keep the VM at 2.0 they are going to create this disgusting half-assed thing that will produce a bunch of Bla$1.class files.
The main critic here is that the VM is going to stay compatible, that is you can put new code on an old VM without too much rewriting.
What .Net does is currently nothing, what it will do is to modify the CLI and the CLR (see this paper). .Net can do this because it's still a young, not much used platform, it must take the opportunity to do it now.(Cyberpunk, 2002/11/27 08:39)
lien permanent
![[]](http://m3.moostik.net/img/?pseudo=kame&cpt=NuAccueil&option=invisible)