We'll make a true component framework out of netWindows yet, by golly.
I was out talking with Patrick the other night and we started talking shop (security). He's working on an interesting project which we talked about at length, trying to find mental holes in the encryption implementation that kind of thing. It's always good to exercise those neurons.
One thing we stated talking about that I can't seem to get out of my head though was some kind of replacement for IDMEF. Patrick noted that Snort's XML output might make a very good starting place to work from in designing an intrusion detection data standard that doesn't require you to read tea leaves to implement. IDMEF is/was well intentioned, but as we all know, the road to COBOL is paved with good intentions. What we need is something more lightweight that is more domain-specific (say, network IDS only). My thought was that if we did it right, we could write a definition for a network IDS data exchange format and then write a definition or XSLT conversion for it that would turn it into valid IDMEF markup. Not that anyone uses IDMEF, but at least we wouldn't be throwing away years of work in getting a spec built, and it leaves our domain-specific thinger with a migration path to a more all-inclusive language for those orginizations that need it.
Not that I have spare time to do it in.
I fixed a couple of nagging problems with my linux install today: I got accelerated X working for my GeForce 2, I fixed KDM so it uses a background that I acutally like (as opposed to the default SuSE one), and I got KDE 3 (final) installed. Having 3D capable X makes me feel better about the box as a whole, not that I use 3D apps a lot, but the idea that I didn't have control over my own box annoys me. I'm happy to report that glxgears runs like a bat out of hell on a 1.4GHz processor =)
I think I've also turned Matt into a KDE convert, which can't be bad. It's not just eye-candy, it's good software, and that's what I like best about it.
Now I just have to scrape togeather enough dough to replace this bad DIMM...
So I borrowed my roomate's copy of "Python, Essential Reference" by David Beazley to see how it stackes up against O'Reillys "Python Standard Library", and I can say without a doubt that Essential Reference is the book that Standard Library should have been but wasn't.
Those that know me have probably been witness to my loathing of O'Reillys aweful Python books (with the exception of "Learning Python"). Standard Library covered it's bases, but I couldn't ever really think of it as a step up from the online documentation for Python. Essential Reference on the other hand is exactly what I want. It has COMPLETE attribute and fuction listings, reference tables that make sense, and sample code that illustrates things in a useful way. I may keep the Standard Library book, but I intend to buy myself a copy of Essential Reference ASAP.
Last night I checked in support for attributes in netWindow's generic parser routine. I've since updated almost all of the widgets and sample pages. While not exactly enthralling, this improvement should both make inline widgets much more legible and easier to use. A substantial side benefit of this change is that for widgets like the tree viewer the use of attributes in place of sub elements can speed up processing times by an order of magnitude by reducing the depth of recursion. All in all, a very good thing for the project.