Douglas' Blog

Welcome to Typo. After three or so years using Roller as my blogging server software, I’ve decided to move to something a little more lightweight and RESTful. And in keeping with my terrible habit of basing choices on the language I prefer to use, I’ve landed on Typo. I don’t remember where I read it, but there was a great article on why you shouldn’t choose a piece of software based on anything but it’s functionality. This is mostly true. But that said, when you are administering a server, you tend to want to consolidate the pieces that you look after. So, if possible you choose software that can run on your database, middleware, web server and platform of choice. So for me, that means if I can find a blog system that runs on Linux using Ruby, Apache and PostgreSQL, and it has most of the features I want then it’s the software for me. And Type fits all those criteria.

I’m still going to have to administer an Apache Tomcat system for the remaining Java Servlet-based applications I run, but I expect Typo to be much less of a hassle to maintain and update so in this case I don’t mind that the apps aren’t consolidated. As it turns out, Typo has way more features and themes than Roller does and the Typo community seems much more active too so overall it’s a good move.

For anyone who regularly reads this blog or subscribes to the feed, you will want to update your bookmarks. In the meantime (and probably forever), I’ve set up my webserver to forward any incoming requests as best as it can. I’ve also managed to import all my old posts reasonably well, complete with comments.

I finally got around to installing the new version of Apache Roller (formerly called Roller Weblogger), which has now been accepted as an Apache Software Foundation project. The new version has a bunch of new features but the most noticeable of these features is the ability to tag entries. Tagging is a feature available to most other bloggers for quite a while now so it is nice to be able to join the masses. I’m not sure if I will bother going back in time and tagging all my past entries but perhaps I will tag some of the better entries I’ve written.

Update: Of course, none of the themes (either the defaults or the extra, downloadable ones) have support for tags so in order to display the tags for an entry, a tag cloud or a search box for finding entries by tag I will have to manually modify my current theme. Sigh.

A new version of Gallery (version 2.0.3) was stabilized in Gentoo Portage yesterday, so I installed it on my test server. You may recall that I recently ranted about how great Gallery was to install (I guess that’s not a rant per se), so I thought a quick followup on the upgrade process.

The upgrade process was even slicker than the install. I copied the new version of the files overtop of the old files. When I reloaded the Gallery main page, the app was smart enough to sense that I had upgraded the files and asked me for the administrator password. Then, a wizard guided me through a handful of automated steps that ensured that all the required files and directories were in place and that various modules were upgraded. And that was it. So slick.

I tried the Quake 4 demo, conveniently available for Linux, earlier this week using my poor Dell Latitude C840, with a 2.0 GHz Pentium 4, 768MB of RAM and a 64MB NVidia GForce4 Go. Unfortunately, the game was barely playable, even at the lowest resolution and with various special efforts turned off. Of course, even had the game been playable, I’m not sure I’d be all that interested in playing it like that. When you play any game associated with id Software you want to do so at high resolution with all the bell and whistles. Otherwise, you’re missing all the reasons to play the game in the first place.

So, I guess Quake 4 will have to wait until I get around to dropping some cash on a new home machine. And given that my laptop is under warranty for another year still, I can’t see it happening anytime soon.

The KDE and Mozilla Firefox groups have both released new version of their software today. From KDE we get KDE 3.5, the last of the 3.x releases on the way to 4.0. And from Mozilla we get the long awaited Firefox 1.5, which is the halfway mark on the way to Firefox 2.0. I have already started building KDE and expect to build Firefox at some point tomorrow (once KDE is finished building). Of course it’ll likely be next week and beyond before the packages are marked as stable within Gentoo portage. Sigh.

When I first heard about Nav Canada’s plans to make filing a flight plan available on the net I was quite excited. Then the project was delayed such that the application wasn’t released until well after the promised date. I don’t remember the length of the delay but I recall that it was several months at the very least. I also have no idea why there was a delay but I can only assume it had to do with the code that they likely had to build to integrate the web application with the existing system. Why? Because the web application is terrible.

Professionally built web applications these days, as a whole, are generally good and most of the older apps have been updated as time goes on so as to take advantage of the features offered in modern browsers. The Royal Bank’s Online Banking is one such example. I’ve been using their web banking since it was made available and though the years they’ve added more features and made things easier. Their design hasn’t changed much, though I think that is a reflection of a good design in the first place. The applications built by Solium Capital are also very good, though since I was a part of that, I might be somewhat biased. Nav Canada’s online flight plan filing, on the other hand, is a step backwards. Maybe two or three in fact.

The first impression of the app is frightening. There are 32 sections that are required to be filled in, all listed on one long page. To Nav Canada’s credit every field has a link which popups a reference within the user manual so you can read about what each field means or how the field is to be filled in. But, that said, there is no per field validation when storing a flight plan. For example, the cruising speed field expects N0140 for 140 knots or M082 for Mach 0.82. That’s fine except that I’m not a flight service specialist nor have I ever flown an aircraft that would require me to file the cruising speed as a mach number. So my first ever stored flight plan has 105 in that field.

Perhaps if I used this system every day I would eventually remember this but given that I fly a few times a month, it would be much easier to have a help icon of sorts that popped up a small window. Inside that window could then be a text field and a drop down. The drop down would have two choices: knots and mach. In the text field I would enter 105 and then select ‘knots’. Then when I pressed the submit button, my choice would be converted to N0105 and filled into the proper field on the original page. The same could be done for most of the other fields. And while the documentation is fairly thorough, it would be nice if the gave a couple of complete examples for the route field. They have a samples of single points but not of a complete route.

Don’t get me wrong. I think it’s great that Nav Canada is using the Interenet to try and make it easier for me to file a flight plan. But I’ve gone on perhaps a dozen cross country trips since I registered for the online flight plan filing. But I haven’t actually used it once because it takes me 15 minutes to make sure I’ve filled in the fields properly but only 2 minutes to call up the 1-866 number and talk to a specialist. I actually tried to use the system once but when I went to submit the flight plan I found that I needed to enter my PIN, which would be fine except that I couldn’t find it and it wan’t anything I normally use. That’s because the PIN (which is different from your passowrd needed to login in the first place, is a randomly generated four letter password per se that is assigned when you register for the system. There’s no way to reset or change the PIN and if you ever lose or forget it, Nav Canada can’t reset it either. They have to delete your account (thus losing any saved flight plans) and you must re-register. Not even the online banking apps are that archaic or poorly set up.

After all this, I’m curious as to what others think of the web app? Am I just being snobbish because of my past experience building web apps or are others hesitant to use the system also?

Following up on my past post on Apple’s QA woes, it appears that some other pieces of Tiger weren’t exactly thought out fully. Sound like any other large companies you might know that make an operating system (read: Microsoft Windows) full of holes?

Ars Technica has a brief article about how Tiger breaks Java for some users and wonders whether or not Apple will ever fix it’s QA problems. I’m not sure it’s entirely fair to say that given that testing entire operating systems is a big job, to say the least. But, that said, it has been many, many months since the last Mac OS X update, so one does wonder how something as large as the Java VM core dumping could get missed? Then again, it’s not like Apple has a great track record on this.

Update: added link to other Mac QA/testing woes.

The best part about the open source development movement is that basically anyone, regardless of experience or training can participate. The worst part about the open source development movement is that basically anyone, regardless of experience or training can participate.

If you don’t know what I mean, then you are most likely just a regular user of open-source software. Not that there’s anything wrong with that. But if you ever get involved, you’ll run into examples of childish behaviour all over. There seems to be a lot of it going around, like some contagious disease.

Though, that said, I will admit that the code in question was quite clever.

Sun has released JDK 1.4.2-08. It’s simply a bug fix release, of course, since JDK 1.5.0 has been out for quite some time now. That said, there are still quite a few changes, as noted in the release notes.

I actually installed 1.5.0 back in October but have yet to fully move to it completely because: (a) Gentoo still hasn’t stablized the package; (b) Gentoo’s current setup makes it virtually impossible to have both 1.4.x and 1.5.x on your machine at the same time; and (c) I am still working with multiple projects which still require 1.4.2, which in most cases is because they run on Gentoo servers limited by points (a) and (b). Nice. That said, I am making progress in my transition to JDK 1.5. I anticipate by June that I will no longer have a need to even have 1.4.2 installed on my laptop.