Similar Linux Desktop Install
Here is a link to a guy who had a similar linux desktop install.
Archive for November, 2005
Awesome article by Mark on a Sony DRM Rootkit
Real quick: Great article by Mark Russinovich on how he finds a Sony DRM root-kit on his system. That guy is a great hacker. It would have taken me 10x longer to pull off what he did, and I have most of those tools (except for the Kernel debugger/windbg harness).
Note: Try to pull this off in Windows without Sysinternal’s tools. Try to pull off this kind of introspection in any other OS. The message: other OS’s better get their tools up to snuff or the DRM storm will overcome them.
Note 2: Wouldn’t it be cool to do that kind of detective work for a living wage?
RedHat Linux in 2005
I wanted to play around with some Unix again, and I have been wanting to get into OS X. However, economics tells me to wait a little bit. Macs are priced a little too high for me to just play around with right now and I’m not going to put together an OS X x86 box.
So, I tried out one of the recent desktop unix distros, just to see what it would be like before I get a Mac Mini.
I downloaded Red Hat Enterprise Linux and installed it on a spare PC I had for Windows 2000. It is a decent Athlon 900mhz box with 1 gig of RAM and a 25 gig drive. The video card is an older Geforce MX AGP card.
The install took quite a long time, but when it was over, it all just worked. The only problem I had was getting the correct video mode for my display (a 24″ LCD). That took a while, but I eventually got it working. Here is what I’ve found:
They have made significant progress. Overall, I’m impressed. This is a usable developer system. Even the sound came up on the first try!
What did they get right: RPM is still a great package distribution system. It was easy to get ‘yum’ to install all the software I needed. They have also, finally significantly improved the configurability of the system with a GUI. It’s not equivalent to Windows yet, but it is gettting there. I even got rid of my CAPS_LOCK all through the GUI ! This is a first for me on Unix. The GNOME evnironment has improved and even uses the Windows keyboard shortcuts, which is smart. It is also looking much better than previous releases.
Where they are being held up: Graphics. The system known as X Windows is still hanging like an anchor on their neck. It is too hard to configure, and has visible screen-tearing on animation. The font system and graphics just have a rougher edge to them. Who knows if my graphics card is using acceleration, I have no easy way to find out. Put GNome on top of that, and you are just doing OK. You get the feeling that it isn’t worth learning too much about the system since it will all get re-written again in a year.
The bundled apps/open source is also below par. If they didn’t have FireFox and a few other polished apps, it would be back to 1990 style development. Don’t expect a painless media experience. For example, I tried to start the display manager and nothing would happen. I had to try and run it in a terminal to find out that the python script that was that program was throwing an exception.
Overall: I still think OS X (NextSTEP 
) is the best desktop Unix around and probably the most developer friendly. This is a decent second place. I will keep using it until I get a Mac Mini or Mac Laptop.
(Also, note, if I didn’t have the previous Linux experience, I would have been screwed trying to get the display working, and given up.)
OS X for $200
Here is the article describing the technique.
I already saw this running on a more compatible system a while back. If you read the article, you will see that this system can outperform a Mac mini in most operations.
$200 is the right price point. I wouldn’t put any more money than that into a system since OS X will probably change hardware when it ships.
IMHO, a Mac mini is still worthwhile as a system since you get so much software, good hardware integration, competitive performance, and great industrial design.
People Just Don’t Get It
There are programmers in the world. Then there are also those prorgammers, CTO’s with large megaphones.
Here is one of them: Making programmers more productive
It is 2005, and I love what this guy finally says. “The sheen of Java has worn off”.
Kudos for finally saying it.
Now I direct this to the vocal programming community at large:
Why does it take people so damn long to figure out these huge mistakes?
I think there is this hubris or pride in the programmer community. If you aren’t doing something really complicated like that other guy, then you aren’t smart. They say to themselves, “I’d better show that I know this stuff in order to be the uber geek! Damn the long hours and hair pulling. I love long compiles!”
Here is a tip: change the word complicated to productive and see if you don’t get on the right track.
Now to be fair, that isn’t the only reason. People are risk averse and usually blind to the differences between C and Java and XYZ. They will pick what everybody agrees on. Unfortunately, this system called ‘the bandwagon’ fails a lot.
Another fair point, ‘what about very weird technologies, Dru?’. Good point. There is the risk of lock-in on a language or system that may not make it. Would you want a system written in a dialect of Forth or Haskell?
Ok, but still. If you expose yourself to lots of technologies and you are a CS grad, you should know how to balance this and choose the correct technology. C’mon - when you see a ton of sites being built faster in PHP by non CS people … much faster than you can even design anything in Java… it is time to take a look.
Side note: the guy who wrote that article worked at Excite@Home. First, that company failed. Second, I worked with a Sr. guy from Excite at Danger (I think he is a founder at Oodle, from what I hear) who took the CEO’s position that you couldn’t run a service on Linux and PCs and insisted on Solaris and Oracle on Sparc. This was in 2000 and there were obvious examples to my point like Google and eGroups. Nice guy, but way behind the curve. After they spent a ton, I think they finally got it, but I was gone before that. It is important to record that Danger slipped over 1 year from their target ship date and had serious quality issues on launch. I am glad I didn’t attend those board meetings.
Side note 2: At Danger, they also insisted on doing things in C/C++. I tried to get people into Python. Last I heard, they actually started going in that direction.
‘Ok, Dru, why the angry man voice?’
When you can see a clear path, and everybody takes the harder one… it can get a little frustrating and disheartening… right?
