Martin @ Blog

Today Firefox 1.5 has been officially released. This is a major release for the Firefox project, because along with the new version of the browser, a new website is launched (http://www.mozilla.com) which claims to be more userfriendly for less customers with less technical understanding than most of the current Firefox users.
The most important new features of Firefox 1.5 include default support for SVG-images (which is very nice I think, and very important for the adoption of SVG as a general webstandard), improved support for CSS2 and CSS3 and of course the regular bug and security fixes. For Apple users, it is very intersting to know that support for the Mac OS X operating system has been improved.
I have just installed the latest version of Firefox on my Mac mini, and I must say that the improvemts are obvious. First, the rendering speed of websites seems to be dramatically improved. Additionally, there was a problem when filling out webforms. Previous versions of Firefox tend to become very slow when you were typing large texts (such as weblog postings), this seems to be solved. Finally the middle-mouse button now works again when you want to open a website in a new tab.
For Mac users, Firefox 1.5 is definately a big improvement!
I don’t have tried the browser on other platforms yet, because I don’t like to install it on Ubuntu without an official package for it. Because I hardly use Windows, the change that I will try the Windows version of Firefox 1.5 within a few days seems very unlikely.

Today I’ve decided to sell some useless stuff I own. Because we don’t have a very large house, our storage space is limited.
The following items are for sale:

My Canon Powershot Pro1 camera. This camera has 8Mpixels, 7 times optical zoom (28-200mm) and high-resolution LCD. You’ll get an 512MB Lexar 40x CF-card along with it and a LowePro camerabag. Only 500 euro!

My only x86 desktop system . AMD Atlhon 1,33GHz (Thunderbird core), 512MB PC133 SDRAM, MSI K7T TurboII (VIA KT133 chipset) mainboard, 80GB Maxtor 7200rpm harddisk, DVD-ROM drive, CD-RW-drive, Club3D GeForce 4 Ti4200 graphics card, Inwin Q500 full-tower case.
Only 300 euro!

I also have an Microsoft Natural Pro keyboard for sale and a Zipdrive including 3 zipdisc (the case of these disks are heavily damaged, but the discs are fully working and undamaged).

Because I already lost one time my complete weblog and some other databases, and very recently almost destroyed another database which was hosted on my server and required for the unversity project of my girlfriend, I’ve decided to implement a decend backup strategy. Because I have no tapestreamer or something similar in my server, I am restricted to ‘online’ backups.
Last few days I wrote a bash shell script which makes a backup of the production databases which are on my server and copy this tar.bz2-file with the SQL-dumps to a computer in Zeeland (geographically about 150 km away from here), my Mac Mini (which is also almost always on, but is in the same room as the server) and e-mails the tar.bz2-file to a Gmail-account which I have created for this purpose.
I think it will be pretty hard now to lose all date. There is 24 hours between the backups, so in a worst-case scenario I lose at most 24 hours of data, which is not a very big problem I think. Every backup is currently about 400KB (compressed, almost 2MB uncompressed), which means that on the Gmail account can at least about 5000 backups, which means that I’m settled for about 14 years :).

Now, I only have to fix some kind of backup system for my data. But that is less urgent, because all data on my server is stored on a RAID-1 mirror, so if one disc crashes, I will not loose everything.

Coming three days, the 24th, 25th and 26th of November, the HCC-dagen (HCC-days) are happening at the Jaarbeurs in Utrecht. A few years ago, before the world was overwhelmed with online hardware-shops, this was one of the few opportunities to buy cheap computer hardware and see the latest developments on the computerhardware market. Nowadays, with online hardware stores which sells their stuff without any significant margin, the cheap hardware isn’t available at the HCC-dagen. This is one of the main reasons I didn’t visit the HCC-dagen for about three years. This year I’ve decided to visit it again, just to see what it is today.
The HCC-dagen once was one of the biggest computer hardware events in Europe with more than 100.000 visitors. I don’t think they reach this amount of visitors the last few years. Also the ‘magic’ of the HCC-dagen is not there anymore, because on the internet one can find almost everything you will find at the Jaarbeurs. Tweakers.net will be at the fair with two booths. One will be at Cambrium, which is one of the partners of the site (they provide the TweakDSL-service). This booth will mainly used for updating the so-called HCC-pricewatch, where you can find the lowest prices at the HCC-dagen. Additinoally there will be a shared booth with mainboard manufacturer MSI. At this booth it will be possible to buy Tweakers.net merchandise, including casebadges, glasses, t-shirts, etc. While I’m still working for the site, I will not be present at one of these booths 🙂 .
I’m planning to visit the HCC-dagen along with my dad on sunday.

Since I’ve installed CentOS on my server and start using Spamassassin from this distribution for filtering spam out of my e-mail, I experienced a decreased accuracy in comparisation with the Spamassassin-installation which I had when Fedora Core 2 was running on the server. Today I upgraded Spamassassin 3.0.1 to the 3.1.0 version which is included in Fedora Core Development (Rawhide). This required me to rebuild the package, which is very easy. According to Felix this version should work better than the 3.0.1 version of CentOS.

Last week I started my final project in order to finish the Computer Science study I’m currently doing. My final project takes approximately 9 months to finish, which is quite long. The project on which I’m working the coming 9 months is called ‘Sapphire’, which stands for ‘Semantic-enhanced Agent-based Platform for Personalized access of Heterogeneous Information Resources’. In normal language it means I’m trying to improve an existing system for searching and browsing content of a certain information domain and add relevant content from the www to it using external search engines. The idea is that using ontology languages such as OWL and RDF to create meaningful queries for search engines.
The first week I mainly read about ontologies and ontology languages including OWL, RFD, RDFS and so on. The steps for the coming week is to get an overview of search engine which are useable for this project (probably Google or Lucene) and dive into Service Oriented Architectures. Mainly this last subject is extensive and because I don’t have hardly any knowledge about it, this means a lot of reading.

For this project the university of Eindhoven works together with the University of Leeds, which means that maybe in January we have to go to Leeds for a few days.

Yesterday I read about Avahi which is rather easy to fix on Ubuntu, and my Mac mini seems to support it out of the box. But while this worked nice, I decided to improve my DNS-support in order to easy connect to the computers in my LAN. While I wanted to implement it correctly, I decided to fix a split-DNS configuration using the view-construct of Bind 9. It turned out to get it working pretty easy and now I’ve a nice working DNS-setup and can ping ‘laptop’, ‘nebula’, ‘truus’ (computer of my girlfriend) and ‘macmini’. From the ‘outside’ these DNS-entries are invisible.