Year in pictures [DUTCH]

Hilarious

Tech Support

Hilarious cartoon

OpenID & Windows CardSpace

Long time since I've blogged. Just had a look at Windows Cardspace, looks promising but apparently there's
a similar service vendor independant. I've just enabled
mine its at http://koen.serry.myopenid.com

IE is just crap

As happy as I was that I was using AJAX in my latest webapp (http://www.proco.be) as disappointed I was that
the prototype library (Ajax.Update in particular) didn't seem to work on larger datasets in IE since its
header has a certain limit. As usual IE just swallows the error.

Since webwork is
using dojo internally I was given a choice but I found prototype easier to use, so now I was forced using
dojo. OK modifying, debugging couple of hours later...
system error -1072896748 in IE's
usual javascript error box
WTF??
Googling a bit pointed me to some weird
error in the MSXML parser that couldn't handle UTF-8 (mind the dash). OK fixed, another error in IE, what's
next...

Well got an even nicer one.. Seems if you take an element and give it
id="description" and try setting the innerHtml of it it won't work. Setting it to whatever magically fixes
the stuff.

Why is following the standards so hard?

First post on the new blog

Fjew,finally done. Just installed the new server and I have to say it, it's a lean mean killing machine :)

The webapp is completely build via webwork 2 and I have to say it, it really rocks. Validation, I18n and
that combined with the radeox engine (which needed some modifications that I'll need to post back to the
authors still)

OK, more on this later.

.Net collections and the need for speed...

<quote>
A few comments from 'The Other Side'

I've been
reading your blog since I'm doing occasionally stuff in .net, but java is where my heart is. But this entry
gave me the creeps so I couldn't resist commenting on this blog entry of href="http://community.bartdesmet.net/blogs/bart/archive/2006/03/22/3831.aspx">Bart
Desmet.

If you're using JDK5 (and it seems you did) you may have noticed
that new ArrayList().add(1) actually works. So boxing and unboxing has made it into Java too (Although I'm
not sure this is a good thing).
Second the collections framework in java is imho far better than
.net, first of all there is no common interface for (non-)generic collections in .Net. So
System.Collection.Arraylist doesn't implement System.Collection.Collection or whatever, like they do in
java. This would actually allow components to be written that take collections as an interface, regardless
of how they're implemented. Or to have open source alternatives, that could turn out to be
faster/lighter.
On the backwards compatibility, while it's true SUN could have opted for putting
them in a seperate package, they didn't. And quite frankly It doesn't even bother me that the compiler warns
me about it, since I can either disable them or use generics. What this has to do with versioning hell is a
big question mark for me. Java hasn't suffered from versioning hell like .net does. We do however have our
own issues like ClassNotFoundExceptions, ClassLoaders and stuff.
Although things have improved
since the 'COM' days, I don't know about something like class loaders in .net that allows usage of separate
versions of the same package in the same program. And then we didn't discuss about the backwards
compatibility between .Net2.0 and 1.(0|1) which is if you ask me questionable.
As far as startup
time is concerned, well the statement that .net is 2,2 times faster than java based on a simple program like
that, where the JVM has the disadvantage of not being partially loaded into memory already, I'll leave that
to you.
</quote>

Quest for more competition

or how the quest for more competition of the European Commission can have the opposite effect. />

Since more competition usually means lower prices for customers, everyone was in favour of
these rulings.
Now, lower prices means also lower margins, it also means that if companies want
to maintain their level of profits
it would require cost-cuttings as well. Now unfortunately most
effective cost-cuttings are done where the costs are highest.
Till this day this is still labour,
the way companies are doing this is either fire people or do major mergers. So instead of 2 persons />in 1(2) companies doing the same job, ending in only 1 person doing this job. If you see this on a
european scale, this in effect lowers
the wages as well. So now, is more competition better for
the european population or for it's companies in the long run? Strange how good measures can have a twist
and why the gouvernment should thing twice before adding additional legislation.

Re: Usability problems in JSF

A while back I've posted a compairison on
JSF vs Tapestry and even though it was a while back (last year). The points are still valid. If
you have something to add, see href="http://sfjsf.blogspot.com/2006/02/usability-problems-in-jsf.html">here another person
with some issues with JSF

Geronimo & Deployment Descriptors

I'd already checked out Geronimo before when it was still in Milestone stage, but since it got released I
finally got myself to actually deploy a complete earfile in it (so including an EJB module (Session,CMP and
messagedriven beans). Well I can say it's not as easy as it looks. Ok, they have a hot-deployment directory
and a nice "console" web interface, but that's where you can stop the comparison.
/>What I **don't** like at all is the excess of deployment descriptors. Just when everybody agrees that
there are too much DD around, just to deploy a simple war file in geronimo you need something called a
'deployment plan'. This is ok if you want something special for a particular kind of configuration. But what
if you'd settle with the defaults...Even SUN won't make that mistake twice.

New member of the family - Dina

Last saturday we got our new puppy (a jack russel).
On this picture it's really quiet and peaceful, but that's since she's asleep.
Otherwise she's a whole lot more active.