Symphonious

I’ve been looking into creating an editor for the JDNC XML descriptor file using EditLive! for XML partly as a test case for ELX using a big and complex schema, partly as a way to contribute to that community and partly as a marketing effort (it would make for a good example of what you can do with ELX). Sadly, the JDNC schema is completely invalid so it causes a whole heap of errors to be output by Xerces and finally causes Xerces to throw an ArrayOutOfBoundsException.

Since we use the schema heavily and use Xerces to parse it, that makes the JDNC schema pretty much a no-go for ELX right now (not to mention making it completely useless for anything else). I’ll have to check a few things before I can be completely sure it’s the JDNC schema that’s causing the problems (obviously the exception from Xerces should never happen either) but I’m pretty sure it’s missing some imports for schemas it references.

The schema also seems to override types in a way that isn’t supported (attempting to change mixed content types to element only).

What the hell is a kiss and ride? I’ve heard of park and rides where you can drive a short distance from home then park and catch a train the rest of the way but a kiss and ride? I guess it’s a drop off point where you get your husband/wife (or in San Francisco your boyfriend) to drive you a short distance from home and then you catch the train and they drive home again. Still wouldn’t it be easier to call it a drop off point? Is it really that inconceivable that someone might be dropped off by someone they don’t want to kiss?

Anyway, I’ve got to go find a good looking girl to give me a lift to the train station tomorrow…..

Chris DeBona argues that the incompatibilities between Linux distributions don’t matter because the majority of people use RedHat or SUSE so that’s all people bother supporting.

Of particular note is the quote:

In the windows world, people don’t feel the need to support every version of windows, either.

In the Java world we do try to support everything. We want everything to be compatible, we raise the bar so that code should run everywhere. Sure it doesn’t always work out but that’s the aim and it’s simply not good enough to lower the standards for compatibility to what the linux distributions have - it’s not even good enough to lower standards for compatibility to what Windows offers.

So Chris, you obviously need to spend a lot more time in the Java community rather than the Linux community to grasp this argument.

This morning after James Gosling’s keynote, was a panel discussion on the future of Java centering around whether or not it should be opensourced. Mostly the discussion was just a whole bunch of useless hand-waving with IBM saying we want it opensourced, Sun saying we want it to be compatible, Laurance Lessig saying don’t stuff with the open source licenses there are other legal means to ensure compatibility (he never mentioned what any of them were) and the users saying “we don’t give a damn either way”.

The one stand out in the argument would have to be our very own Brian Behlendorf who pointed out that the key thing is more about allowing open source implementations of the standards rather than whether Sun’s implementation is opensource itself. He stressed a few times that this did require being able to make available implementations that weren’t yet 100% compliant but *not* claim that they are compliant (essentially you don’t do a 1.0 release until it’s compatible but you can have nightly builds and/or CVS available). I may be slightly twisting what he was saying there but that’s what I took away from it and it’s got to be pretty close to what he meant.

So good work Brian!

Very incomplete notes from the XML Security using Apache session. The slides cover things better than my notes do but I thought I should save what I wrote down anyway. I gave up taking notes half way through once I realized I wasn’t really adding anything to the slides that should be available online anyway.
Read the rest of this entry »