14 Sep 2002 Mondragon   » (Observer)

mascot: I wouldn't worry too much about the LCD backlight life unless you're used to having your CRTs last more than 3 or 4 years. I've had my Samsung 170T for over a year now, and it still stacks up nicely sitting next to the brand new 171v that I bought. Realistically, the backlight is eventually going to die, since that's what lights do... :-) That being said, my Powerbook G3 - wallstreet - still works great after 4+ years, so I would expect an LCD bought today to last on average as long as a CRT (and when you go to replace it, it'll be a heck of a lot cheaper).

GUI Development

I'm really looking forward to the new O'Reilly 'Creating Applications with Mozilla' and hoping it turns out to be a good text. I'm pretty down on doing anything with a UI at this point because it's well-nigh impossible to work in a cross-platform environment. wxWindows claims to be cross-platform, and it builds, but realistically, many GUI elements are messed up on different platforms because the library doesn't support native control sizing, or dialog resizing. I'm giving up on resource files from wx at this point - it'll have to be something like .nib files for MacOS X or what Galaxy (from Visix) had before it'll take the pain out of building a cross-platform GUI.

Which, of course, brings us full-circle back to Mozilla, since XUL does appear to do a very good job. I've been told that Mozilla is rendering, though, and not using native controls, which could be a problem. Galaxy rendered, and they did a pretty good job, but it's always a little strange when a new OS release is available (mostly for win32), or when trying to handle i18n support. Supposedly a company called Ambiencia bought Galaxy from Visix, but I don't really know what the deal is with that - like, what it costs...


Speaking of O'Reilly books, I think they need to be a bit more cautious in their mad rush to put out more books for the community. I'm still pleased with the company as a whole, but it used to be that if O'Reilly had a book on the subject, it was the book you bought, no questions asked. They were hands-down the best books you could buy on the subject matter. Trouble is, that isn't so true anymore. Both 'Virtual Private Networks' and 'Practical Internet Groupware' were pretty soft on real useable content, and I've perused through a few on the shelf recently that didn't look so good either.

There's also the matter of certain author's letting their perceived celebrity go to their heads. 'Internet Core Protocols' is a decent book, and it covers the basic to intermediate subject matter pretty well, but every time Eric Hall makes a comment on any networking mailing list, it reminds me how much he really doesn't understand about how real networks work. That's fair - you certainly don't need to have run a large network to be able to cover things like TCP and higher layer protocols - surely Stevens hadn't - but you shouldn't think that you have the core knowledge to comment on operational problems either.

Of course, in retrospect, it seems that their coverage of software development and unix administration has been very good, while most of the books covering networks and protocols haven't necessarily been top notch. Which, of course, makes me a bit nervous about 'BGP', by Iljitsch van Beijnum, but I certainly am not going to pass judgement before I see it. He seems to have the background experience, and the world certainly needs a good book on BGP (Halabi is great, but not necessarily for the uninitiated), so I have hopes, but I'm hedging my bets a bit.

Software Development

In an attempt to avoid being overly critical of other people (too late?), it's time to change the topic. I'm still in the hunt for a project that needs some core development help that I can get in on, or at least something where I don't have to fight with the UI code. Data structures and networking are my thing, so anything along those lines would be great. I tried to get in on a Python SMTP server at sourceforge, and it seems like an interesting project, but it appears to have died from lack of attention by the lead developer. Maybe I should just pick up the torch and run with it, but frankly I'd rather leave the project management up to someone else in a large project like that.

Obviously Python doesn't lend itself to any real data structures work, since it's fairly meaningless to implement complex data structures in Python itself, given the high overhead of the primitives used in those structures, but I find it an interesting language that I'd like to do more with. And besides, you can always implement data structures in C/C++ and expose them to Python as extension modules, which is probably worth doing for some specialized data structures, since the built-in ones are a bit too general for some tasks. General is great for the native language, since there's really no telling what people are going to do with it, but if you want to be efficient, it's probably best to roll your own. Of course, the question there would be that if you wanted to be efficient, why are using Python, and I do think that there is some kind of balance to be struck between development time and runtime efficiency, but I also think that Python makes great glue for C++ components, and that performance doesn't necessarily need to be lacking.

But anyhow, it's late, and I'm riding the Civil War Century tomorrow, leaving in about 3.5 hours to drive up there, so I should probably get some sleep.

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!