Fast

So I just recompiled the latest trunk build of Firefox (Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9b5pre) Gecko/2008031221 Minefield/3.0b5pre) with -O3 and -mtune... WOW
this thing flys, my total sunspider time was less than 5s! (sunspider results). It only took 22 min to compile too. Whatever those guys a Mozilla have been doing the last few weeks is paying off big.

RAID

I love Ubuntu, it is really a first class Linux distro, and I could never see my self going back to Windows. However, there are some rough edges. Yesterday I got all new guys for my desktop/server box, and spend 2 hours carefully assembling it. I grab an Ubuntu CD and head off to install. I now had 4 80GB SATA 2 drives that intended on putting together in a RAID 5 array, for a blazing fast 220GB of fault-protected storage. 14 hours later at 4:05 am, I headed to bed while my system updated. Why did it take 14 hours to install Ubuntu? Well, it seems that Linux doesn't play nice with the fake raid on nforce mother boards, (yeah i know about, dmraid, but have you ever tried to install it?) Some how the first time I installed it, it crashed pretty hard, and the system seg-faulted on boot. So I went to format and reinstall, and I couldn't get linux to forget about the old array, which was now corrupt. I even went so far as to use a Windows boot disk to format and stat a windows install on each drive individually. I finally resorted to plugging in the drives one at a time, installing ubuntu (ok, so I'd cut the power after it started downloading packages). That finally did the trick. I will say however that there needs to be a good tool to completely wipe a set of drives, as in factory condition nothing at all on the disk.

liftoff

wow, so I published some code to launchpad, e-mail the programming team and 36 hours later I'm running a real project. Google 'code score', 8 of the first 10 hits are for our project, that's just amazing to me. We have a wiki and have a great start on outlining the feature set. Hopefully over spring break me and a few other people working on it will really be able to get a lot done. If anyone has any interest in ACM programming team software, please check out the wiki or send me an e-mail.

Code Score

So over the weekend Mercer hosted a spring programming competition for the schools that are in the ACM SE regional during the fall. I was not satisfied with the submission software. So, I am working on my own. I have registered the project at Launchpad as 'Code Score' and it will be licensed under the GPL3. If anyone is interested in working on it, e-mail me: adam@adamcornett.com. I am using ZK and Cayenne again after being very happy with them in my last project.

its been awhile

So my project, which I mentioned in the last post has progressed signifantly (see cct.mercer.edu) and I'm quite pleased with it. Over my Christmas break I ended up re-coding almost the entire thing. I have learned quite a bit about developing a real application, the main thing is to not take short-cuts, writing the class/module/package right the first time is worth the extra time and thought. After I re-wrote all the 'bad' code that has been slopped together at the last minute for demo's last semester, I found that development was quick and painless. Extending a well written framework makes programming enjoyable.
I am also quite impressed with the ZK framework. It really makes AJAX development simple. I chose to code the AJAX UI sections in Java, (ZK also allows for a XUL like UI) which made development quite simple. Object Oriented user interfaces are really neat, code reuse becomes the standard and I can now add a new form and admin console in about 15 minutes.
The Java community as been in quite a huff lately over the idea of closures in Java 7. I have skimmed several articles on them and I can't see the advantage really, just write a method! One of the drawbacks of C++ is the language bloat, and I think Java is headed that direction too. One feature I am looking forward to seeing in Java 7 is type literals and reified generics. In the application I'm working on now, having type information available at runtime could eliminate a good deal of code, and eliminate the need to pass a Class object along with a generic list.
On a personal note, something just bug me. One major thing that I just cannot understand is when people flat lie to your face. There is the person I know, and I asked the person if they were lying to me, and they said no, well I come to find out later that of course they were... Why, did that person lie, it was about something that was not trivial, and easy to verify, I chalk it up to plain stupidity.

code monkey

So finals are done, finally. I am now coding like 16hrs a day to get a project done for a client. It was a class project that I committed to finish, and in the last 48 hours, I've done as much alone, as my team of 4 did over the last semester. Granted I do have an advantage of no other work, and I don't have to worry about keeping in sync w/ 3 other people.
ZK is also a great framework, it took about 30 min from reading the online documentation to moving my old pages to new fancy AJAX pages. I am an API slut, I am using Apache Velocity and ZK (its actually a really awesome combo) and Apache Cayenne for my data back end, which makes development very smooth. Action handling is very smooth and its almost like writing a desktop app, everything is in Java, and the UI elements in ZK look well, and function great!

sum of parts

How tiny and insignificant are we as individuals. A single person is a weak, frail being, with an existence unfathomably short in the cosmic view. Our ability to change the world around us is minuscule, even the greatest works of man will not out live the least of God's works. At first a sense of hopelessness can overwhelm anyone even beginning to grasp how tiny of a spec a human life is, but there is hope. Man has something that is beyond scientific explanation: a soul. What you do with it is what matters, because that is what is beyond the mere dust and shadows of this world.