Sevrin Fixed, Built, and Ready to Roll

The problem with the Sevrin Core build script was that I upgraded Ant and it decided to compile all of the classes twice. A simple classpath fix.

 

Note for those of you who are interested in running Sevrin Web Stack: it is not well documented. It's not hard to understand, but it may take some work.

 

The Web Stack itself will attempt to auto-configure itself. If you install it in Tomcat, for instance, and then load it in your browser, it will ask for configuration data. For security purposes, you will be required to enter a randomly generated password located in a protected file on the server. The absolute path will be given to you on the screen to find this password file.

 

Just be sure to create a database in MySQL and table creation rights to a specific user for the Sevrin database. It should create everything for you. After it is complete, restart the application (or the server itself) and it should load fine. (And yes, you can use this same mechanism to auto-configure your own applications based on the Web Stack. :)

[=] posted by jason
Oh God... is this still here?

Low-down: I'm alive. 4 more months and I'll finally have my CS-degree.

 

And another version of Sevrin Core is out, with a completely re-vamped Web-Stack. Cancel that... my Ant script is old and refuses to compile and package everything correctly. *sigh* I'll re-post when it finally gets updated.

 

Now I just need to finish writing a CMS for this site... time is hard to find these days.. :(

[=] posted by jason
Ranting

Scenario: Write a C server socket application for a made-up protocol and connect to it using telnet.

 

Problems:

  1. Where *ptrSize = np->valSize would work perfectly on Fedora Core 1, assigning the int value of np->valSize to the int* ptrSize declared earlier, SunOS 5.8 does not agree.  Apparently, it requires about 10 hours of debugging before you realize that the C code you learned last semester actually gets completely #%$@#! when Sun Microsystems gets involved.
  2. When working with sockets in C, using this code:
    setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
    (void *) &val, sizeof(val));
    should technically allow the socket to be reused after your done with it.  Even after calling close(sock);, I have to wait 5 minutes for SunOS to allow me to run my program again because it keeps the goddam port open.  Ever try debugging code in 5 minute increments?  Not fun.

 

[=] posted by jason
Once again...

... I'm busy.  Christmas break turned into worktime, trying to move from one job to a new one at the same time.  Lots of hours, very little 'whatever' time.

 

On the plate this term: Computer Organization with Assembly, Network Computing, Macro-economics, and Ethical Issues in Information Technology (Philosophy course, essentially).  All of it is good stuff.  And a great schedule, too, with a few days off each week (and one really annoying mandatory tutorial session... bleh!).

 

Work continues on Sevrin (finally! - link to the right!).  And my sanity has finally returned, after years of having been lost in Saint John, NB.  Wow... it is possible... and in Halifax at that...

[=] posted by jason
Busy

For the few that wander these digital paths, I've been very busy for the past 3 months.  Mainly with school.  I've returned to Dalhousie University as a Computer Science student (fulltime) while working 32 hours/week at a "part-time" job.  In between school and work I only have time to study, so that leaves very little time to think about the site.

 

Sevrin has been put on temporary hold until Christmas exams are over with.  I should (hopefully!) be able to get v0.6 out at that time.

 

On the plus side, I've been learning some cool stuff about number theory, matrix theory, and some really useless and abhorent probability and stats.  Oh, I'm being "forced" to learn C as well, which is a bit of a change from my mainly Java/Pascal background. :)

 

Anyway, off to sleep.  Have to work in 3 hours. :(

[=] posted by jason
IBM crystal ball says your future holds...

Oblomovka reported a long time ago about an IBM projection of storage costs for the next twenty years.  They projected that by 2004 a GB would cost a mere $0.77 USD.  Every year the price drops by 45%/GB.

 

According to his calculations, IBM was almost right on the dot.  And that's in US dollars.  So I did some conversion for Canadian geeks and found that we're doing pretty good ourselves.

 

TigerDirect.ca offers a Western Digital 200GB ATA-100 (w/ 8MB buffer) for $184.99 CDN.  That works out to: $184.99 CDN = $140.92 USD / 200GB = $0.71/GB (scaled up).  That's even less than IBM projected, but still close enough to call them right.

 

Of course, if you hook up on the $40 rebate on that same HDD that they're now offering then your cost lowers: $144.99 CDN = $110.45 USD / 200GB = $0.55/GB.

[=] posted by jason
Cereal and Music

I once downloaded an MP3 compilation called "Laidback Niceness", which was made up of a good selection of tunes that were nice to sit just sit back and enjoy listening to.  Now I find out that there's a Web site called Soulshine with all kinds of goodies.

 

While listening to some of that music, take a gander at what's new with cereal.  And what Cereality has up it's sleeves.

[=] posted by jason
New Yay! Yay!

Whoa!

 

YayHooray! has completely changed suddenly.  As of yesterday, it seems.  The whole experience has changed.  Personal blogs, advanced filters, improved search, no more point system (you can sign up without anyone voting you in; no more 500+ people in a small "waiting list"!).

 

Very cool.  But the color scheme definitely has to change.  My previous one was very dark, which I loved.  This one is too bright.  Oh well... in due time.

 

They even have easier links to direct profiles (like mine)!

[=] posted by jason
RDD: Is it for you?

John Topley writes about RDD, an interesting new approach in J2EE project development, also hinting at a .NET port.

 

Dry humor.  I love it.

[=] posted by jason