Archive for December, 2007

It’s been another fun weekend. Including some rumours which rather pleased me but I won’t be discussing on a public blog (don’t get excited, it isn’t juicy gossip). It has sparked an interesting series of thoughts, though.

Turns out Eric Atwell really does live this side of Leeds. He came in tonight while I was working.

Speaking of which I Googled his name to check I had the spelling correct and thought “hmm, what if I had blog to the end of this search query.” It wasn’t long before I was fishing around to see what SoC staff had blogs. While my search was admittedly limited to a narrow selection of the staff, one man, of course, failed to disappoint.

The blog is called Pythoneer’s Journal, so I’ll leave you to guess which member of staff it is :D.

Which then leads me onto the question, is it a good thing if you know that people read your blog? (Or perhaps another way to phrase it, how bad it is if you know people read your blog?). Because if you know people do, how much do you start censoring it?

I mean, would I be so harsh on coursework specs if I knew the author of it read my blog? Would I be so harsh about my PD31 group if I knew they read it? I like to think so, particularly because I knew several of my PD31 group did read my blog but it makes you wonder…

There is something magical about all-nighters.

I’m going to miss them when I graduate. Although I will hopefully graduate into an environment where there are still plenty of them. I can’t see it disappearing if I’m working at a software company beit my own or even if I sell out. Hopefully the professionalism of a company will give us the organisational skills to actually get some chair races going too.

Helen commented “you lot love all nighters.” Who is us lot? That is a very sweeping generalisation which really annoy me. But what could I really expect, after all, everyone from Wales is small minded. Eh, see what I did there ;). The point is, I kind of do love them. They suck in having to do work and I would much rather have the work done beforehand with plenty of time to spare. But if the work is out of the way and it’s just adding extra bits or polishing things off then they can be a lot of fun too.

You get a wonderful community atmosphere that you don’t really get during the day. The desks are piled high with energy drinks and cans of coke (unopened of course, nobody drinks in labs :p) and the bins in the corridor outside smell of the late night run to Fortune Cookie which last night saw us order so much food for the people in DEC-10 that we got a box to carry it all in.

In the end, while I wouldn’t miss the stress of late night coding sessions just before deadlines, I will miss the atmosphere and experience. What we should really start having are all-nighters as soon as the coursework comes out :p.

We’re currently working our way through the second SY33 coursework but like the first, it’s a pointless exercise. It doesn’t test your understanding of anything to do with SY as we have to do things like consume the Yahoo Search web service – all the code for which is provided by Yahoo in their developer’s library. Your actual task in this coursework is trying to work out what the coursework wants you to do. That is a task which turns out, is very difficult.

Yet again the author of the coursework spec is unable to grasp numbering (it’s worth noting I believe the specs had different authors so no one person is to blame). The first coursework had two sets of concurrent numbering systems while this numbering system starts at 0 instead of 1.

The real issue though is how unclear the coursework spec is. It randomly jumps between instructions, setup information, hints, etc. You have to spend a lot of your time picking out the actual spec from the spec itself. Once you have, most of it doesn’t work. You spend more time trying to get the setup instructions working so you can do the work than doing the actual work itself.

To demonstrate my point, I started writing my coursework in gedit but got a load of errors. The solutions didn’t help, the only other solution I got, and I got this repeatedly is, “oh, use Eclipse.” So much for the school’s policy of avoiding a focus on specific software. I did but this just produced more errors. There are loads of extra classes you needed, Dan kindly email them to me but it didn’t help. I then added all the ones Si told me I needed, still didn’t solve the problem though.

We ended up concluding that I had identical code to Si and Dan’s so it couldn’t be my code. Kieran eventually traced the error down to the WSDL2Java tool not producing all the files I needed. So I spent some time looking into that but in the end found that I did have all the classes.

Eventually, at around 2:30, it just started working randomly. This left me way behind even the people that had just turned up that day to do the coursework. I eventually managed to get everything else sorted and moved onto servlets around 3:30ish.

That was another mission. 5 hours I spent trying to get them to work. Kieran spent a good few hours on it too. It never did. Because of lack of time in technical coding? Nope, of course not, but because setup was yet again the mission it shouldn’t be. We were having to compile everything in Java 1.4 for some reason. This produced an endless string of errors that resulted in me spending all night in the lab and still not getting servlets working.

Wrap-up consisted of having to answer the few questions and also print out a copy of my readme file to include in the physical hand in – why? What purpose does that serve? This detailed how to get it running like they had any chance given pretty much everyone got beaten down into using an IDE because it was such a mission to get it working from the command line (although I did eventually manage this in the late hours of the morning).

Seriously, what is the point of this coursework? To test my patience?

Got back my GI31 coursework 2 marks today. They aren’t too pretty. While it isn’t the mess that my first coursework was (the coursework itself was OK I thought but the mark it got wasn’t), given how much time I invested in it and what standard I thought it reached I was disappointed with the mark. I got 69 and a half. Doesn’t sound too bad but then this is one of the good modules in terms of actually examining you on useful stuff (rather than BS modules like DB). Further more this kind of thing needs to pull my grade up. How can it pull my grade up when it isn’t even a first itself?

There gets to be a point where you just have to conclude, “I’m not just not putting the work in. I’m just too stupid.”

It now being past 8am I have admitted defeat with regards to servlets. I got mine compiling but it still wouldn’t run so I’ve included it in my source code submission for some kind of consideration however small that may be. All in all though this has been a joke of a coursework and while I will be making a separate blog post about it I feel the need to rant a bit here too. My mysterious error which plagued me for hour after hour just randomly started working at 2:30. I had gone on and coded further stuff by then but this essentially placed me back behind everyone else at 2:30am!

Anyway, we’re through it now, term is over! At least for me. Sort of. I still technically have TKD today but I doubt I will go even on the off chance I am awake. I also have to sort out a few meetings with people next week with regards to my FYP and such not to mention all the work I have to catch up on for things like my websites and of course there is always FYP and lots of revision for exams that need to rescue my spiraling grade.

All in all, this isn’t the joyous end of term I was hoping for.

• The University of Toronto Press publishes a journal named the Canadian Journal of Linguistics
• Journal papers require a needless amount of special formatting
• Weka isn’t a particularly useful tool
• Don’t be fooled into thinking because you’re taking a DB module, the software you will be using will support databases
• Don’t assume that because you are on a computing degree your modules will be anything to do with computing
• Don’t assume that because you are on a computing degree your assessment will be based on technical knowledge
• Don’t assume that because you are a third year computing student they won’t try and teach you how to use Google
• If a module leader explicitly tells you not to use CRISP-DM, don’t believe them
• Data mining happens, and I quote, “by magic”
• Data mining has no value and isn’t useful in language research
• Data mining has no value and isn’t useful, period

Last night saw my first TKD grading. It was slightly tiring to say the least lol. Lesson one, make sure you eat something that day. Luckily I had had breakfast but not everybody had. These people were easy to spot collapsed at the back of the hall. In the end I passed though with a less than perfect grade but it was a pass none of the less.

Earlier today A-Soc held a support rally for the British teacher Gillian Gibbons. We had quite a decent turn out in that people actually turned up which is always a victory :D. While we were only around for an hour we got quite a bit of interest and got someone new interested in the society so all in all it was a rather successful event. Go A-Soc!

Well you can’t have any. Because yesterday Colton Mill ran out :D. And somebody had to go chasing all the way over to Oakwood to get some. Fun times.

Before I start I’m going to state a few things. I’m going to say something that’s a bit out there, a bit controversial and a bit, at first glace, counter-intuitive. But never the less I am going to put the proposition forward.

Private profiles and turning Facebook into MySpace.

There, I said it. It seems crazy at first, after all the point of Facebook is that you are connecting with your existing real world connections and that it isn’t just an open directory of internet aliases.

But the fact is, this recurrent doom that people keep talking about has failed to materialise. They let Leeds Met onto Facebook, nothing really went wrong. They let the whole world onto Facebook, I’m not going to pretend I am pleased about that but it didn’t turn Facebook into MySpace. They added applications and yes they are very annoying, but you choose to list the people as friends that send you the application invites and you can say no. I do. It keeps my profile nice and clean, it’s still Facebook.

Private profiles however, I would argue are doing some damage. What I truely loved about Facebook was that each university had it’s own network and you could pretty much move freely around that network. You could see new people’s profiles, see what mutual friends you have in common, see the connections. Private profiles prevent you from doing this.

This hysteria about keeping your profile private is ruining the Facebook experience and turning Facebook into the jungle of uncommon ground and paranoia that is MySpace. It’s a slippery slope my friends, a slippery slope.