Mircea Bardac /random_thoughts

I spent a large part of this day updating the Psi Windows Installer, getting it ready for Psi 0.10 release.

If you don’t know, the Windows Installer has been moved to the Psi Darcs Repository (since Psi also switched from CVS to Darcs). The sources can be found here.

For a quick preview on how the new installer will look like, have a look over Psi Wiki: Windows Installer Guide. Graphics were created by our Psi web designer - MRAY - lots of thanks once again for the wonderful contribution.

The installer reached Version 2. The version bump was caused by some implementation improvements I’ve added recently, easing up the work required to package future versions. These 2 improvements include:

• script for building the file list with everything that needs to be installed - NSIS has its own way of defining what files should be installed where; this script takes a Psi archive, unzips it and creates install/uninstall file lists to be used directly by the installer
• script for creating language pack sections - as you probably know, you can include Psi language packs with the installer. The Windows Installer has a feature which makes it particularly interesting: if you select a language for the installer interface, it will be automatically selected from the language pack list - not an exquisite feature, but useful. This script looks in the directory with Psi language packs and automatically generates sections for the language packs in the installer (install/uninstall file lists) and setup functions (for autoselecting the language packs depending on the installer language)

One thing to notice: I wasn’t able to compile the installer script properly with the latest NSIS compiller(v2.10) and I used an older version (v2.03). There a strange incompatibility between the two versions, when it comes to displaying variables put in the Installer section names.

The new version of the Windows Installer will be used by Psi when 0.10 is released.

I had today a nice seminar of Systems’ Theory. This subject is described by our student-ancestors as “a unique experience” (and that’s it, nothing more).

This subject has been known for years to as the killing subject of our faculty - the last killing subject, the last subject stopping you from starting a lean life researching only computer science.

To put is short, this subject is about Mathematics. Mathematics all the way. We’ve done tons of Mathematics, more or less important - this seems to gather everything and amplify it a few times. It really looks like a unique experience. It’s goes through your body from head to toes, activating tons of pointers to names like Laplace, Fourier, Dirac etc. This combines Analysis with matrix computations. It is the dream of any mathematician.

Ok, now.. we shouldn’t get desperate about it. We’re far from loving this subject. Analyzing the facs in an objective manner, we notice the following:

• tons of Maths and we were happy that were finally over (last year) - everything we do gives us a very weird feeling - each Mathematics we learn didn’t seem to depend on something else; it was only needed by itself; nobody else required it - as older students tell us, same thing applies for Systems’ Theory
• Theory - now, this is a recent shock for me - I realized today that this is theory; in a very weird way, this didn’t occur to me until now; this makes me remember about how students ask for more practical stuff to learn during University.

Subjects depending on themselves are not helping the development of the computer science student - we need a clear connection between each subject we learn and its practical implementation in Computer Science.

Input Data:
* two semi-groups of 14 students each (some of them unexperienced with computers in general)
* one Linux lab for the Using Operating Systems course
* one TA (teaching assistant) (me)

Output Data:
* students understanding the way Operating Systems work, step by step, as this can’t be taught in one night/week

Problem: the best way to get from the input to the output.

The situation is like this: I am keeping some lab classes which value 50% from the the final grade. 40%/50% is given after test examinations from the lab classes work. That leaves the TAs with 10% for activity during classes, participation etc.

Other TAs and me have discovered that the students come at the lab classes without previous preparation. They don’t even have the papers describing the lab activities. This lab is like no other lab I have had (attended), since the TA (me) has to describe lots of functionalities all the time, practically holding a course on practical usage of Linux, taking into consideration the inner workings of the Operating Systems. During other lab classes I had, the only thing we had to do was to complete a set of tasks using knowledge gathered from the course. These lab classes are not like that.

We, the TAs, kept telling the students that they need to prepare in advance. I’ve personally asked them to bring questions. I have received no questions though. Everything suspiciously seemed clear, given that many are newcomers in this field.

Under these circumstances, we weren’t seeing any implication from the students. The thought that they might not be ready at all for the future official examinations appeared in our minds. The course also introduced lots of new ideas. Therefore, it has been decided to give them a flash test - I think this is the “standard motivation technique” . The students on Tuesday were the first ones taking “the hit”.

I was both sad and happy to discover on Wednesday, the following day, that students came prepared. The effect of an unannounced test was incredible. I was happy because they were finally looking over the materials we have provided, sedimenting the new knowledge. I was sad because we had to use this, kind of “barbaric”, method to determine the students to look over the materials.

I do understand that people learn only when they like it. That brings great motivation under normal conditions. Something strange happens in the first year though. Given what I’ve noticed, it seems that students tend to give more attention to the threatening subjects - subjects like the ones eating all their time, subjects which are often considered responsabile for them not learning enough about the usefull stuff.

I personally don’t like that… and I hope this will change in the future.

The teacher holding the Using Operating Systems course asked the students of the first group in the series he teaches for to remain after the classes to get aquainted. I have attended this meeting, since I am a Teaching Assistant.

This was a good priviledge for gathering input on their feelings as new students. It was interesting to see the diversity of the environments students were coming from. It remembered me about my first year, when I was surprised by many of my colleagues.

Some of the students felt like expressing their ideas about the school, after only 2 weeks of studying. It was partially a shock for me to hear what they said.

Not to get into lots of details, the most expressed ideas were on the implementation of the new teaching system. As you probably don’t know, the teaching system has been revamped in .ro starting this year. Students will do 4 years instead of 5 (+other years for BSc etc.). Everything looks great until now. Older students can only imagine how is university without lots of useless subjects.

Now comes the “fun” part. Some subjects were provided less classes. Surprisingly, some teachers admitted this AND they said that, because they have less classes, they have to compress the content - do more stuff in less time. Now, where’s the improvement there?!? Students were supposed to make 4 years learning other (&less &addapted better to the specialization) stuff. Students were puzzled.

Some of them complained that they don’t have time to learn usefull subjects because of the other ones. It kind of remembered me of the 2nd university year, where we so caught up with other subjects that we had to leave behind Assembly Programming.

Well, it was sad to hear that… I really thought the system could do it right from the start. Unfortunately, there’s too much suff to be changed, starting with the mentality of some teachers…

A happy migration day in my life of a Linux user.

I easily managed to upgrade my entire system from GCC 3.4.3 to GCC 4.0.2, using the wonderful ArchLinux distribution.

Steps:

1. list self-compiled packages using a script
   

   #! /bin/bash
   installed=`pacman -Q | sed -e “s|\(.*[^ ]*\) .*|\1|g” `
   repos=`pacman -Sl | sed -e “s|[^ ]* \(.*[^ ]*\) .*|\1|g”`

   for pkg in $installed
   do
   found=`echo $repos | grep $pkg`
   if [ "$found" == "" ]; then
   echo $pkg
   fi
   done

2. disable/uninstall all the packages listed above, in order for the system to remain stable after a bootup with new GCC & libraries
3. upgraded everything with “pacman -Syu” - followed close all the changes being done in the system, mainly backuped files
4. rerun “lilo” - new kernels got installed
5. rebooted
6. I could feel fresh GCC 4 air
7. reinstalled “nvidia” - the installer has a setup which must be run after a kernel upgrade
8. recompiled own packages previously disabled/uninstalled
9. started X
10. enjoyed KDE 3.4.2… an updated version compared to my previous one

Now, I’m getting ready for KDE 3.5

P.S. Yep, I do enjoy being an ArchLinux user.