Evaluating Wikis: A Case Study

I’ll be delivering a talk on Evaluating Wikis: A Case Study at the Exploiting The Potential Of Wikis workshop put on by UKOLN on Friday 3rd November 2006 where I’ll be giving a rundown of how we selected which wiki engine to roll out at the University of Bath.

This means that I’ll have to actually, you know, evaluate some wikis and roll something out before then, but come along! It’ll be a hoot! Past comments on my talks have included “Very interesting”, “Philip is a real live wire!” and “Ran out of time” – how can you resist? 😉

Help, feedreading hits back

I’m currently highly distressed – it turns out that the latest version of FeedReader is really good (that is to say, it’s not a Java or .Net app and so actually runs at an acceptable speed on my 2GHz, 1GB RAM machine), which means the yawning pit of infovorism has opened before me once again. Maybe this time I could keep it to under three hundred feeds? That’s not too many is it?

From the UK, Amazon Unbox is shiny, exciting and cheap

A DVD

It’s pretty funny reading comments like this Wired commentary on the newly launched Amazon Unbox, because despite the onerous DRM issues (which are definitely a problem, don’t get me wrong) – the UK has nothing like this. Even just renting films from a shop is crazily expensive. Netflix? Yeah, we wish. Amazon and Blockbuster have both set up online DVD rental services in the last year or so, and there are now some indie companies doing this too, but there’s nothing (as far as I know) that’s good, cheap and makes you think like you’re sticking it to The Man, like Netflix seems to do in the US. In fact, one of the more well known DVD rental companies in the UK is actually Tesco. Tesco, for those who don’t know are a supermarket who take £1 for every £8 spent in the UK. That’s a hella lotta cash.

I guess the only tenuous comparison I can really draw where the UK audience gets a better deal is, inevitably, with the BBC. The BBC puts a certain amount of their output (specifically that which is produced entirely in-house, I believe) on the web, available to watch for seven days after broadcast and unencumbered with DRM or the need for a special player. It’s not especially high resolution, and if you use something like Net Transport to download the Realmedia stream which they broadcast and pipe it to your TV you’ll see definite blocking and colour loss.

I look forward to the day when such affordable downloads are available (legally) from companies in the UK. They’ll probably be burdened with DRM but I really don’t think that’s going to go away any time in the near future.

Hiring and recruiters

The recent posts by Joel Spolsky on recruiters (you can start here and work backwards) are spot on. Recruiters seem to not only never tell you the name of the company they’re hiring for until you’ve totally confirmed that you will apply and they have to tell you the company name because you need to actually find the building for the interview, but they’re mostly clueless about the job requirements which a) the company is advertising for and b) which would meet your skillset.

I feel quite raw about this issue. I got my new job a year ago (I can still call it new if it doesn’t feel that long, right?) and I’m still getting emails and calls from recruiters asking if I’d like such-and-such a position or am I free for contract work? Oh? You have a job already? We didn’t know, can you tell us more about that so we can update our database? Thanks. Oh, and we’ll call back in a month to see if you’re still happy.

Get lost!

If I want a new job, I will go and look for one. I will notify the agencies and websites etc. etc. and do that on my own without you having to prompt me! Furthermore, when I do get a job, and you next email or call me and I tell you I have a job, write it the hell down! After I’ve started a job, I expect those calls – after all, how do they know I’m now employed? But after I’ve told a different recruiter from the same agency for the tenth time that I’m not interested, seriously, my patience starts to get a little bit tried and honestly, it means I’m not going to actually deal with you next time I am looking for work.

Even worse are the emailers. TOP LONDON JOB NOW! £££!

Yeah, right.

What really gets me about the emailers are the ones which aren’t even mailmerged, they don’t even say “Dear Phil” at the top, it’s just the same mail sent out to hundreds or thousands of people around the country who come up on the “matching skillset” search on their computer. How lazy can you possibly be? I will never, ever reply to one of those mails, and the chances are I’ll write the company name down and never deal with them again either.

In fairness, there are some good ones. There might be an early phonecall after I put my CV up and notify companies, and then the following emails (even if there turn out to be loads of them, which is not often the case with these companies) are always “Dear Phil, Thanks for X. How’s the job at Y going?”, so even if they are mailmerged from their database it’s such a tiny amount more work that I will instantly love the company in comparison and be a million times more likely to use them in the future.

I’ve almost vented my frustration now, but just to conclude: I’m still receiving comments and emails about a blog post I made this time last year, in which (the otherwise good) Computer People make a demand for a photocopy of my passport with no explanation why and then their legal department phone me when I post about the possibility of it being an identity theft scam.

Load fast, help me blog

This blog is currently powered by Blogger. The only two desktop posting tools I’ve used and liked are w.bloggar and Windows Live Writer. The problem with both of them – or at least, my perceived problem with them is that they take too long to load. Quite often I just want to post some idle thoughts on a subject without putting in too much effort, but I know that having to load one of these tools will probably take around ten seconds, by which time I’ll either no longer care, or have expanded the thoughts to the extent that I’ll save the post up for more fuller treatment later (which of course, never happens) and so it goes unblogged and forgotten.

Windows Live Writer I can understand being slow to load – I mean, hey! it’s a .NET application, what do you expect? w.bloggar though just seems to do it to annoy me. Well, not really. It’s a native (non-.NET) Windows application but I would guess that before the window appears and it lets you start typing it performs some checks to make sure the server is up.

I couldn’t care less if the server is up.

All I want to do is write, and if the server’s not running, then tell me in the status bar, save my text when I get to the end, minimise to tray, and publish when it comes up again (or let me close and choose to publish when I start it next time). Just let me get on with hitting the keys.

Ripping last.fm radio to separate mp3 files on Windows

(And if that’s not a title to reel them in from Google, I don’t know what is.)

Step 1) Install and download the last.fm proxy (requires Python)

Step 2) Follow the two-step configuration in the README

Step 3) Start the last.fm proxy and configure a radio station to listen to

Step 3) Download and start StreamRipper

Step 4) Configure the output location for your files

Step 5) Put http://localhost:1881/lastfm.mp3 into the URL box

Step 6) Click “Rip”

You will soon have a directory full of correctly named (but ID3-less) mp3 files.

Storing Atom

So then, the Atom Syndication Format is a great standard for storing data but once you start getting a large number of entries (even if it’s only in the hundreds or thousands), does storing your data as an XML file start to become unwieldly? A small part of my brain starts to cry at the thought, so I’m going to guess “yes” but there must surely be loads of companies and individuals doing this, so how do they manage?

Although I’ve previously harrumphed over native XML support in databases, is this a reasonable use-case? Storing each atom:entry in a single field in a database row would seem to make some sort of sense because you could use the unchanging entry:id and entry:updated as a key for each version of the entry. I imagine the problem then might be one of containing the data in the XML and making sure it didn’t start spilling out into its own fields and tables (things like tags and categories for example).

I know that on previous occassions, Joe Gregorio has dreamed about this and I guess now that Google, WordPress and Movable Type are effectively Atom Stores all of their very own but aren’t designed as such and just provides APIs to their own data stores.

I notice that the MySQL 5.1 beta has its own XML functions so maybe it really is time to start taking a look at this approach, although most of the Java apps I work on use Hibernate at some level, and I’m not really sure how well it supports XML data fields (if at all).