Month: May 2003

Girls. Women. Females. I swear that, for as long as I live, I will probably never completely understand them, but that’s probably the way it’s meant to be. I’m not going to say too much because, well, I don’t really want to. Let’s just say that I’m unsure what has changed about me in the last six months or so. Something must have changed, because clubbing is very different for me now than it was before. I think it’s fair to say it’s much better now than it was before, but I have absolutely no clue why this should be the case. That’s me: Completely F. Clueless, Esq.

Anyway, I went out clubbing tonight with my brother, my cousin Terry and his two friends Treung (sp?) and Colin. It was Terry’s birthday on Thursday, so we were celebrating it in the typical way. We had a great time. I went to a club called Level 5 for the first time. It’s on level 5 (of course) of the Hilton Hotel on the Queen Street Mall. It’s basically a hip hop club. I didn’t think I’d enjoy it, but I did. It was great! I’ve decided the reason I go clubbing is simply because I actually enjoy dancing. As soon as it starts getting too freaky for my taste, I pull back. I think tonight is the first time I’ve really and truly realised, I don’t want to meet a girl in a night club. No matter who the girl is, it just doesn’t seem right to be attempting to pick up a girl in a night club. People have been telling me this for ages, but I just haven’t been listening I guess. I’ve either got to be comfortable with meeting girls in clubs, or I’ve got to realise that when I go clubbing, I’m doing it because I enjoy dancing. If it starts getting freaky, get the hell out of there because you’re not going to pursue it, idiot!

So this begs a question: If I’m not going to pursue girls’ advances in clubs, where am I going to meet them? Well uni seems like the obvious place, but it’s easier said than done, I think. It’s not like you can just go and sit down next to some stranger and start talking to them, and appear as if you’re a perfectly normal human being. Or is it? Maybe that’s just the thing. Maybe they’re not looking for completely normal. Maybe they want something out of the ordinary to happen. I don’t know. It’s late and I’m getting tired.

I can’t believe I’m committing this to my blog. I really need to add that option which will allow me to post private entries. But I needed to put this somewhere. I needed to record the fact that I’ve come to a decision about clubbing. And I just need some help to enforce this decision, I suppose. At the very least, I hope my humiliation has provided some interesting reading for those who bother to read my blog from time to time. Honestly, this is worse than Big Brother!

For the latter part of this week I’ve been taking a crash course in Perl programming so that I can modify the scripts Nathan wrote to automate the marking process for coms3200 so that they work for comp4301. I’ve finished the modifications in the script I think (I hope). The thing is I need to modify my solution to the assignment so that it writes files which the script can then check. Otherwise there is no way to tell whether an operation ran correctly or whether the correct exception was thrown at the correct time and so forth. It’s ugly, but it should work and it should save me some time. It also guarantees that everybody’s solution is subjected to exactly the same set of tests.

Perl is kind of cryptic, especially when it comes to specifying the parameters to procedures. To specify that a procedure, let’s call it foo, takes a string as argument, you declare it as

To declare that it should take two strings it’s just $$. Lists are specified with an @ symbol. There are some other things that are a bit strange to me as well. What I do like is the way it does regular expressions. String matching and manipulation is really easy. Anyway, enough of this talk, because I’m not a nerd, remember.

Looks like Nigel and I will be headed to Caloundra this weekend since Mum kind of needs us to take some stuff up there. I hope it doesn’t rain this weekend!

This article has attracted more comments from readers in the short time that it’s been up than any other Slashdot article I’ve seen recently. The initial post asks the Slashdot community for some good cyberpunk and sci-fi novels. The Slashdot community obliged with a torrent of suggestions. Worth checking out if you’re into William Gibson and Neal Stephenson type stuff.

My alarm goes off at 4am in the morning. WTF?!!! Then I remembered: The Champions League final. I was dead tired, having gone to bed around midnight. So it was time to do a bit of reasoning. Sleepy reasoning, but reasoning nonetheless. Who was playing? Two Italian teams. AC Milan and Juventus. That probably means a defensive game of football. Nedved, who would have been one of the best players on the field, was suspended. Would there be a reply on TV later tonight? Probably. Conclusion: stuff it, I’m going back to sleep!

As it happens, this turned out to be a pretty good decision. A nil-all draw after extra time, so the game was decided on penalties. AC Milan won 3-2. By most accounts, it was a pretty dour game, but AC Milan was the more enterprising and deserved to win. I’m glad I stayed in bed!

I reiterate my belief that umbrellas are complete useless inventions. I suppose they come in handy when trying to ward off aggressive magpies or the next door neighbour, but when it comes to keeping one dry, they’re as adept at keeping out the rain as a toothpick is. Yet everybody uses them, even though they all get soaked from the waist down. The umbrella industry must be making a killing! A stick with a flap of canvas on top, sell it for $20. Why am I bothering with a Ph.D when I could make a comfortable living selling umbrellas to people? I’ll be contemplating that very question today while I sit here at my desk with wet jeans and wet shoes.

The G-G has finally resigned. Probably a good decision in the long run. The issue had rather consumed the nation for the past little while, and the Governor General shouldn’t be attracting so much attention, especially attention of the negative kind.

I’ve started implementing a scale-free version of a DHT. I should probably do some design first, but this thing is just supposed to be a prototype which I get can some performance statistics out of. It’s good to be doing some programming again!

I’m reading The Bourne Identity at the moment. It’s keeping me fairly riveted. Over 300 pages read so far, which means about another 200 pages to go. I should zip through the rest of the book, because the pace is quickening, the body count is rising and the thrills arriving with greater frequency. I enjoy Robert Ludlum’s style of writing. I’ve noticed that for me, the writer’s style is the most important determinant as to whether I will enjoy a book or not. For example, Douglas Adams could have written about absolutely anything and I’m sure I would have enjoyed it.

I’m generally going to bed earlier and waking up earlier these days. In my opinion this is a Good Thing™. However, as evidenced by the timestamp that will appear on this post, I’m going to bed a bit late tonight. I bet this stuffs up my biological clock for the next week. D’oh!

As I look out my office window, I see a plume of dark gray smoke billowing from behind a church in the Tarragindi area. It’s been burning for quite some time now. I hope it’s nothing serious.

One thing I’ve been overlooking so far with respect to DHTs and so forth is that it may be difficult to solve the problem of heterogeneity. When a peer joins the network, its finger table is initialised to contain the addresses of m other nodes, where m is the number of bits in the identifier or key. But what happens when node A cannot communicate directly with node B because they are on completely different networks? Perhaps node A is a Bluetooth device and node B is on a TCP/IP network. As long as some physical path exists between the two nodes (via a multi-interfaced device or whatnot) the situation is not completely hopeless. My DHT extension must incorporate some notion of proxying directly. That is, proxying must be inherent to the protocol rather than an optional extra added at a later time. The heterogeneity issue is a large one, and one that I cannot afford to ignore from the outset.

About the nerd thing. I just remembered that the title of my blog is Segmentation Fault (Core Dumped)! But I’m still not a nerd, really I’m not. OK, maybe I’m sort of a nerd, but definitely less than half way to being a full nerd. I think I’d better stop analysing myself. Right.

Yay, I’m officially not a nerd! A couple of years ago I’d have been devastated at this news. But I’m actually very comfortable with this at the moment. :) (Says he who has created his own automated blogging system. Hmm…)

I’m trying to get myself motivated about work again. I’m feeling less gloomy right now because I’m toying with an idea to make distributed hashtable protocols like Chord more scalable by using what we know about scale-free networks. If I could crystalise my idea, my service discovery protocol will benefit from this better scalability. The idea hinges on allowing very capable nodes (i.e. those nodes connected via high-bandwidth links and with abundant computational and storage resources) to be able to store more than one entry for each interval in the finger table. This enables these nodes to become highly connected, meaning that the average number of nodes contacted to deliver a message decreases, because there is a high chance that the highly connected nodes can deliver the message directly or to a node which is closer to the target than would otherwise have been possible. In other words, it is possible to more than halve the distance to the target on each hop. The problem then becomes a question of allowing these highly connected nodes to find nodes within a particular identifier interval in a scalable manner, and enabling other nodes to determine which are the highly connected nodes whose IDs are less than or equal to the target key/identifier. I understand that this probably all sounds like gibberish to anyone not familiar with DHTs, but it makes sense to me!

So is there a scalable solution to those two problems? Aside note: Holy crap! My ICON paper just got accepted! Do I seem surprised? I kind of am. I never expected my publication record to be 3 from 3! Yay! So, scalable DHTs. Yes. Um… Capable nodes can simply cache the addresses of nodes that it contacts during the routing process. Eventually, a node will cache a sizeable proportion of node addresses within each identifier interval. This is scalable in terms of bandwidth consumption because it requires no further messages than are already used. But how do other nodes find this highly connected node? Perhaps, in response to lookup operations, the responding node can return the address of the most highly connected node that it is aware of within the relevant interval as well as the node whose ID is closest to the target. If the node knows of no such peer, it returns only the address of the next hop node. It’s hard to get an idea of just how quickly this information will be propagated through the network. It might be very slow. On the upside though, this solution requires just a few more bytes in each lookup response message.