At work I’ve been asked to write a (very small) SWT app, being the only member of the development team with any Java GUI experience at all (I’ve written some Swing apps before).

Of course, there are hundreds of SWT resources out there on the web, notably on dev.eclipse.org web site and the eclipse wiki, but by far the best I found were notes written for Computer Science students (in pdf) – if only my lecturers had been able to make things this clear, simple and understandable!

It’s interesting to note that Eugene Belyaev has also started to look at SWT, and in whose comments, someone has also posted a link to the resources I mention above.

Serendipity is clearly afoot though, because his previous weblog entry is about how to
Launch the default browser on Windows
with this bit of code:


String url = "http://www.eugenebelyaev.com";
Runtime.getRuntime().exec("rundll32 url.dll,FileProtocolHandler " + url);

which I was also looking for!

I hate crashes in the middle of blog posts. But anyway..

JabRSS has been playing up all today, so I only got the news via the old-school method of visiting Danny’s site, but Peter Saint-Andre (aka King of Jabber) has posted on the rdfweb-dev mailing list that they’re thinking about using FOAF to identify Jabber entities. Currently Jabber uses vCard to identify users, and a hacked vCard-temp namespace to identify Jabber Components (IIRC); the use of FOAF would be a great, great move. These are currently the two technologies I’m most interested in, so perhaps I’m biased, but I can’t see of anything but good coming from this.

On a more downbeat note, I can’t quite get my head around what the simplest way of linking a Jabber account with a pre-exisiting FOAF file could be, except for some kind of smushing based on the jabberID property.

Interestingly, Peter points out that It [FOAF] defines data elements for things like projectHomepage and geek code, but not for things like marital status or sex. This hadn’t even occured to me before. The file says who I am, who I know, where I went to school and what my interests are, but not if I’m a boy or a girl. Is this an oversight, or defined elsewhere? I’ve not seen it in the relationship vocab or biographical vocab, which seemed the only likely candidates.

Bah, the bugtracker post was supposed to be a complete rant, but work is clearly exhausting me too much to build up a full head of steam.

Anyway, the original xurble bugtracker is actually available here.

I’ve removed comments temporarily. Mainly because they were taking forever to appear. Replacements soon. You can still contact me via Jabber as specified in my FOAF file </raises the bar>

And so finally, as promised – bug trackers.

Only free bug/issue trackers, and only ones I’ve tried, natch. If you don’t have a Live Demo, I’m not coming in.

And why not start at the top? The daddy of Open Source bugtrackers? Bugzilla.

I can only talk about Bugzilla as a user, I’ve tried to install it to manage my own collaborative projects, but never managed it successfully. To do so is a feat in and of itself.

So how did I manage with it as a user? Trying to see if a bug had already been logged? Failed. Simple search to return all bugs? Failed.

I don’t know, maybe I’m just a dolt, but even when I bring up a page detailing a single bug, there’s so much going on on the screen that I can only just about tell what the bug even is, let alone who opened it, what its status is etc. The only thing I really like is the comments section after it which shows progress and direction the bug solution is headed in.

The functionality it provides may be great but the awful experience of actually a) installation and b) use means Bugzilla is a big no no.

IssueTracker, powered by Zope, is great. Really simple, easy to use. No real complaints, although I’m sure I could come up with some if I really tried. Like the use of Zope. ew. 😉

The sourceforge bugtracker – pretty decent, it works, but stores comments in the opposite order to the way I’d expect. Pretty good otherwise, and probably the most solid of all the ones I tried in terms of a balance between usability and functionality.

I’ve looked at loads of free bug and issue trackers over months, and I’ve been really disappointed. What makes it so hard to make a decent bugtracker? Do I ask too much? Do I even know what I’m asking for? Is it because bug trackers are knocked out by people who need them for purely practical reasons and so don’t spend much time on the interface? Surely this can’t be the case? All in all fairly depressing.

So why was I even looking at issue trackers? Well, the developers at work don’t have one, and I thought it might be kind of handy (duh! how did everyone cope without one?!), but in the end I wound up splitting the bugtracker code (originally written by Gareth Simpson) out of a much larger project I’d worked on called Xurble into its own entity. Sadly the original xurble bugtracker isn’t running anywhere for a compare and contrast at the moment, but the one we’re now using at work is an improved version of this. Which seems to be easy enough, and is in daily use. It’s Java and runs within Tomcat, Jetty and JBoss (and most other web containers I’d imagine). That’s right, this post is really a shameless plug and taster for when it’s finally released. 😉

p.s.

DEVIATION! PAID ISSUE TRACKER AHEAD!

FogBugz from Fog Creek Software is basically the nemesis of bugzilla, it’s very powerful, but with a far simpler interface (although not as good as I’d expect from Joel Spoelsky). It does pretty much everything you’d want it to, is better than IssueTracker, has more functionality than my version and you don’t have to have used an issue tracker before to understand what’s going on. Nice job, now release it for free. 😉