Older blog entries for mrcsparker (starting at number 86)

company I think that the biggest hurdle GNOME has is the poor state of developer documentation creates a high level of entry. There is not one place that I can go to find current documentation.

By the way, I am not trolling. This has been confirmed - for me at least - when I did the KDE KDevelop tutorial this morning. The KDE docs are fantastic - second only to the PHP docs. I avoided KDE for years, not liking Qt's pseudo C++, but it is too well written and well-integrated to pass up.

Yea, I know that I am not a big-time developer around here. I have posted patches here and there where I have found problems. I still like Free Software development and I really wish that GNOME would stop worrying about C#/Java and just write good documentation. The GNOME language bindings are great, but there is no consistant way to develop Python/C/C++/Perl GNOME programs. Anjuta is very cool, but every time I developed applications with Anjuta/libGlade I would find out that Glade was using deprecated widgets.

Please, someone correct me. I would like to know that I am wrong.

5 Apr 2004 (updated 5 Apr 2004 at 19:58 UTC) »

After using GNOME for two years, I decided to emerge KDE on my system this weekend. The reason being is I have been looking through the KDE source code for various projects here and there and I have been very impressed by the thought that went into designing such a large system. Everything is so well integrated and "just works" the way that it should. 3.2 is also hella fast.

There are some little inconsistancies in KDE, like some of the settings tools marked "Configuration" and some marked "Properties". The dialogs also have way too many options IMHO and when I open most apps it is a bit overwhelming to have so many options in the default dialog. I also miss the Ximian icons, and the beautiful Industrial theme.

KDE is excellent, though, and I plan on sticking to it. After two days I can honestly say that my productivity has gone up by using it.

Also, the documentation is excellent. The GNOME documentation leaves a whole lot to be desired and much of it is out of date.

The HP Way

Just got my new Itanium 2 server in from HP last week. Of course, it is not being used and there is no operating system even loaded on it. Right now it is just collecting dust in the server room.

I ordered a small, dual-processor, 2 gigs of ram, raid pci card (etc...) box to run Red Hat Advanced Server 3. It is pretty much to evalutate running GNU/Linux with our current applications on Itanium rather than RISC. Good enough.

So, I find out a week after I order it that HP is offering Opterons for a lot less. A HP guy came in and told the Unix Administrator that the Itanium 2 is "less proprietary" than the Opteron and that the Opteron is not as good as the might Itanium2. Okay, I figure, we ordered the Itanium 2, I will let this slide knowing that all sales people are pretty much full of shit. By the way, I heard from someone inside of HP that the Opteron announcement was a bit of a mistake.

We get the box in - no Red Hat box, PCI in a box, and missing a couple of parts. Yea! We call HP and find out that they are going to send in an "Installer" to set up the box. It has been about 4 working days and we are yet to hear from the "Installer" after leaving multiple messages.

What crap. And I know that HP is really trying to get more servers in here. Guess what - it is not going to happen. IBM has been much more attentive and we already purchase Thinkpads in the hundreds. Right now we are running our critical systems on HP-UX/PA-RISC and they are going to slowly fade out. HP-UX has been great but it seems like it is in limbo with updates, and Linux is not only getting better every week but it is a hell of a lot more easy to use for developers and users.

So, I have a new system sitting just collecting dust. No word from HP yet and I alot of people that need some processes waiting to be moved to that machine.

26 Mar 2004 (updated 26 Mar 2004 at 20:41 UTC) »

Been writing my shell scripts in C. Not sure if they are shell scripts, but they do the same thing as a normal shell script would do. Anyone else do this?

I was writing them in PHP, Perl, or Bash, but I write pretty much the same app in C in the same time. Plus, I don't have to have a manual open in front of me ("How do I do this in BASH... google for it... hmmm") so that I am doing it the "correct" way. I would make a crappy sysadmin.

By the way Muine really rocks hard.

Yeah! Java is not going to be free software

Smart move, Sun. No ANSI, no ISO, and now no code. Sure, but you have to JCP, which is almost like a standard - if you look at it sideways and you have no clue what makes a real standard. Dum dum dum dum dum.

Gnome 2.8 Roadmap

Yea. Just read that all of the GUI stuff in GNOME is moving out of GNOME libs and into GTK. Smart move. So no more libgnomeui?

Downtown Houston: 700+ people are getting visas a day to work in Iraq for Haliburton. There are lines going around the block with people waiting so that they can go over. No matter what you feel about this, it is still odd that this is not being considered news-worthy. I haven't seen anything about this on the news, but the lines are there.

In other news, working on focaza. Creating more of an object-like system.

I guess that sometimes forking can be good

Well, when you have no choice. The developer said that he wouldn't accept outside patches and we already had the software installed on a dozen servers. So, I emailed him begging - please, please - the administrators want me to change the program around a bit and I did not want to maintain my own patches. After a few days, a few emails later, I asked him if I could fork the project and he agreed and all is fine.

So, I am a forker.

Really, I get the best of both worlds - I get great piece of software to build on and any changes I make go back under the GPL. Really, though, I wanted to work with the developer.

So, I added the GNU build system, broke out a bunch of code into libraries. I can pretty much do what I want now, you know. Though it is written in C, I gave it a quasi-OOP system so I can just pass structs between interfaces. And I gave it on of those really annoying names that means absolutely nothing: focaza.

GNU Arch

is really fucking cool. Eveything that I ever wanted in a source control system. I don't understand why people think that CVS is easier than GNU Arch - have any of you guys/girls who think this ever tried to delete a directory in CVS, change a file name or move one large section of your code to another section. Arch makes all of this simple.

tla commit

Had some great conversations via email on Friday with a couple of Free Software developers. One project I wanted to contribute code to did not want any outside code so the guy was really cool and said he wouldn't mind if I forked the project (yea, I know - fork bad). The other guy gave me updates on what he thought was wrong with a really big Free Software project from the inside. {} Not being specific here for a reason - for the first project I do not want to announce that I am forking because the project is excellent and I do not want to cause inner conflicts (but the code will be released under a different name). On the second topic, I would hate to skew anyone's opinion and the guy who had the complaints has not voiced all of them in public yet (though I think that he will).

In other news, I read Havoc Pennington's blog about choosing OpenOffic over GNOME Office and it was a bit depressing (http://log.ometer.com - 2004-03-06: Office suites). I am not a GNOME developers myself, but I thought that with things like bonobo having to pick a specific application as an office suite would not be necessary. I know that OpenOffice is not able to embed gnumeric or AbiWord documents, but I really hope this is where GNOME Office is eventually heading. While I am at it - where does bonobo fit in to the picture? From my days as a Windows developer, I really miss being able to easily embed one application into another - or use a component from one application in another.

Forking Code and Releasing Early

Reading my email this morning I see that a project that I worked on is forking for what seems the hundredth time. I left it a few months ago waiting for the 2.6 kernel to start working on it again and it seems like about 3-4 developers have decided to start their own little 2.6 forks. You see, the problem is that the original project has not released often enough and the developers who actually need the code have gotten impatient. Myself, I have given up. The code looks like spaghetti now (with all of its #defines for 2.2, 2.3, 2.4, 2.5, and 2.6 kernels) and it really needs some serious cleaning up.

There is also another project that I started working with where the maintainer released the code under the GPL but did not want any patches from any external developers. No patches? The project does 75% of what my company needs and it seems a waste to start a new project all by myself. The Unix administrator has started installing the project on a few machines and it works really well but has some apparent shortcomings. So, do I fork? I emailed the maintainer and he does not seem keen to respond back.

77 older 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!