Tom Morris

BBC News reports that Kingston University has put some pressure on a QAA inspectors to change their verdict to avoid bad publicity. Now they have the reputation of being a university that is so desperate to avoid bad publicty, they harrass QAA inspectors. Way to go. What did the QAA report have to say? Well, students were producing not just barely acceptable but sometimes unacceptable work are attaining passes at Honours level... it is surely important not to over-reward this work and thereby devalue the Degree.

The Guardian are reporting on comments from the head of aforementioned QAA about degree classifications being arbitrary and inconsistent, because universities do not want to do anything to harm the golden egg that is foriegn students. Recently, a student walked out of an MSc Finance coure at the University of Southampton because of concerns over standards of English. The BBC published an interesting report from a lecturer on this topic. It really is quite problematic that we have people getting on to postgraduate level courses who cannot speak English. It is bad enough in the arts and humanities, where reading and writing at a high level are expected - try reading translations of Wittgenstein or Hegel or even some of the current, analytic philosophy written in English, now imagine how difficult that would be if you cannot actually read the language - but as the article points out, it has some very problematic consequences in more practical subjects which have laboratory work. Working in a science lab or art studio with someone who cannot read the safety information is a very good way to get hurt. I have worked in photographic darkrooms - and I think it would be very difficult to make sure people are safe in such an environment unless they can clearly communicate with one another.

And, at the same time, there are concerns about postgraduate plagiarism (although there are also methodological concerns about the study). I've discussed this before - specifically, the essay sites that some people are using. It is unbelievable that people are actually paying huge amounts of money to get someone to write them an essay with the faade that they are not then submitting it as their own work. I had a look at one of the essay services recently, and got a quote. To get a first-rated undergraduate dissertation of the same length as the one I did in the final year of my degree would cost two thousand pounds. The essay sites say that they are only providing a 'research service' - that one would pay two grand, get an essay on exactly the same topic one is writing about for their degree, then use that to help them write their own essay. Pull the other one. Even the essay-writing sites wouldn't hire people who got their degrees by using essay-writing sites.

And, after all that negativity, I didn't get around to announcing that I am starting an MA in Philosophy in the autumn.

There's a whole load of discussion going on over the BBC Microformats drama. Mark Birbeck is using it as a "Huzzah! RDFa for the fucking win!" moment, and John Resig is telling us RDF-heads that we're being unpragmatic and making things too complex (life is simple, after all, and can be represented in a finite number of elements).

I'm not going to get into a religious war about this (contrary to popular belief, epistemology gets me more steamed up and fervent than any microformats-RDF squabbles!): I'm not a huge fan of RDFa - I prefer people follow the same kind of design that microformats has done, but using GRDDL to decentralise microformat development. RDFa aside, I think that the current class-based proposal is the least bad solution - and has some potentially interesting effects, even if it is has some in-built ugliness.

We discussed this all at some length yesterday on #swig - see logs. Basically the class-based proposal is to take data that broadly follows xsd:token, stick "data-" before it and make it into a class name. The XPath to get that out will just be... beautiful.

One of the things that having a decent date-time format in HTML will allow is people to construct documents with different dates for different sections. Imagine an HTML document which contained a quote - either inline or block-level - from some historic source. Being able to tag that particular passage up as having a different author, source and date seems like a very cool thing to be able to do.

I think that Frances Berriman and the microformats community will get their heads around the date-time accessibility problem soon enough. It is a tough problem, and it is something that we need to get right. For those of us who have reached RDF Enlightenment, HTML seems very ugly - but just shouting "Use RDFa!" is not a useful way to help. There is a reason why people use document formats, and one of the challenges of the Semantic Web is coming up with a plurality of different ways of creating semantics that transcend document and data formats. Date-time stamps are just one example of this problem.

The Semantic Web is going to be messy - and that is exactly the thing people don't get. It's going to be RDF, it's going to be more XML formats than you can shake a stick at, and more microformats than you can imagine, with people forking those formats, and forking the parsers. It's the Web, remember. The Web is messy. The Semantic Web people don't deny that - but our detractors seem to think we do.

This entry has had the unproductive (if, for the author, utterly hilarious) snark reduced to trace levels for your convenience.

The BBC reports on Jonathan Zittrain's comments, with frequent misuse of the word 'hacker'. The report describes Zittrain as thinking that the very openness of the Internet is leading more and more people to switch to locked platforms like the iPhone, comparing this risk-free 'sterile' environment with a chaotic world of phishing, botnets, distributed denial of service attacks and illegal file sharing networks.

But I just don't buy Zittrain's argument. He sees the Internet as having only two sides - basically, insecure Windows installs or the iPhone. Experience shows otherwise: Linux, the BSD systems (including OS X and Darwin), Java, Apache, and open source. The very existence of these systems knocks a fatal blow in Zittrain's argument. Windows is an anomaly that distorts the argument. And, to be honest, who even uses Windows anymore? Most of the people I know only have a Windows install so they can see how badly IE 6 renders their standards-compliant (X)HTML and CSS or play DirectX games. Even my parents don't use Windows anymore. I remember back in 1997 when I first discovered FreeBSD, the idea that my parents would be using a computer running an OS I could ssh into was completely alien.

And the iPhone is an anomaly of an equal scale. I have a mobile phone in my pocket that is smaller than an iPhone, about as sexy, and for which I can write my own applications for (even if it requires me to use painfully bloated Java toolkits) - a privilege that I have exercised. The iPhone distorts this too. Most phones will not be getting significantly more closed. And Android, OpenMoko and the array of Linux phones seem to be pointing in the opposite direction. The fact that six million people downloaded Firefox 3 yesterday seems to suggest that people are perfectly at home with open platforms.

It has come to my attention that, in the last few years, a new wave of tolerance and gender consciousness has swept across the geek community. Female attendance at conferences is rising, more women are running events, speaking at events, giving BarCamp sessions, feeling more confident and welcome in the geek community. We can't be doing with this kind of spirit of camaraderie and shared purpose.

The first thing you need to make absolutely sure of in putting together an event that excludes women is location. Location is of primary importance. Pick somewhere like Hooters or a strip club or somewhere where your food and drinks will be served by someone in teeny-weeny little shorts and who doesn't mind if you make a bit of dirty innuendo. Be sure the place is in a nice and seedy place in the city. You are big and burly men, remember (yes, even you there in the NHS prescription glasses). Make sure the venue smells of piss and serves nothing but lager - lemonade and cocktails are for pansies and girly men, remember. It's got to be badly lit, have nowhere to sit down and serve as unhealthy a menu as you can find.

What with the smoking ban and the fact that Hooters (etc.) is not everywhere, you cannot rely on the venue alone to make sure women don't turn up. The best way is to let them turn up and then act like a complete arse to make sure they don't come back. This takes some work. The first way to do it is to make sure you make some offensive assumptions. They are there to accompany their boyfriend or husband. They are new and inexperienced. They do not have 'passion' the same way that the men do. They probably work in some soft field like marketing or PR, maybe design at a push - but certainly not someone who spends their time making software. Be sure to always talk about how you are a "hardcore coder", as if it's only a few rungs down from being a hardcore porn star. You are the man, remember, and there is no way a woman can know more about anything technical than you do.

If you work hard at it, you can make geek events completely unwelcoming to women. Way to go!

Seriously now. We have come a long way. I remember that in the mid-nineties, computer magazine publishers would include adverts for premium-rate commercial porn BBSes, and there would be lots of adverts like this for servers and software. Thankfully, that age is over. Or is it? Molly Holzschlag put up a post about how there is a geek event taking place in Tucson, Arizona, at Hooters, a restaurant which is based on female sex appeal and the work environment is one in which joking and sexual innuendo based on female sex appeal is commonplace (and that is company policy for chrissake).

It is possible to create a geek community where women feel welcome - or at least, a lot less unwelcome. I think the UK has been an example of how to do it right, with the London Girl Geeks getting everyone to raise their game, making events and communities more welcoming to women - the Girl Geeks aren't just planning events for women, but are making a valuable contribution by spreading the meme that geek communities need to listen to what women want. We need to constantly be working to stamp out the invisible sexism that is preventing women from being at home in our community, both online and offline.

Following accusations that I have 'stabbed the baby' on DataPortability.org, I took a listen to the audio from the DataPortability Steering Group telecon from today. It has only confirmed what I thought: DataPortability.org is about as agile as a quadriplegic in a coma. These guys are dead in the water. And before anyone goes crazy, I'm just the messenger.

The podcast describes a process of coming up with a formal governance structure. This involves a Plenary, Action Groups (aka. Standing Committees), Task Forces (aka. Special Committees), a super-complex set of voting rules, a 'sunset provision', a complex set of different statuses (Officers, Participants, Voting Participants, Representatives, organizational representatives and observers) and rules about who is allowed to participate and how. Hilariously, the notes for today's telecon contain the following gem of misplaced wisdom: Don't over-engineer solutions to problems that don't exist, yet. Go forward with simplicity, then correct course as necessary later. I guess the placement of the comma in the first sentence means that when a problem does exist, it can be completely over-engineered. Okay, cheap shot. If you omit that comma, it's a very good statement of principles. Now, it would be interesting to see what would happen if DataPortability.org adhered to it.

To pre-emptively rebut accusations that I've stabbed the baby: certainly not. The fact is that adding a few extra layers of bureaucracy to a supposed community effort is not going to improve things. The W3C is bureaucratic, but it actually produces some great work in spite of the administration, not because of it. Who wants to deal with that? It's insufferable enough if you are doing it for a living. But if you are volunteering for an online community project? Screw that. I'd look for some grand, conspiratorial vision - but I know it's not there. It's just a bunch of people who think that the best way to get people to collaborate is to create something more like an org chart than like an IRC channel. Mistaken, but not evil. We'll know the poor little baby is completely dead when we see some kind of WS-* or SOAP fluff coming out of DP.

There is a silver lining in this. A lot of people (wrongly) accuse the Semantic Web people of promising and not delivering (which is a bit silly - the SciAm article came out in 2001, and it's getting into industry seven years later - CSS took much longer than that, but nobody complains about that, because such an objection to the SemWeb is a convenient mask for other things). Well, compared to DataPortability.org, the SemWeb people are really freaking agile. (And we can do party tricks like following our own noses around.)

In other news: last night I started work on an initial design for the project I'm working on. It fits on to one Post-It note. I'm hoping that the final spec will not be longer than three A4 pages. If it is longer than that, that means I have not narrowed the problem space sufficiently.

First, some links to people who've been discussing DataPortability: Eran Hammer-Lahav (and an earlier mailing list post), Dave Winer and Scott Kveton. Read these, please, and read the comments.

Now, I've pretty much given up on DataPortability. I haven't written about my opinions on DP basically out of politeness and hope that they may turn themselves around. I didn't want to say bad stuff about DataPortability, because I agree with the broad aims of DP. But I have to. DP is not going anywhere. The scope is unclear and the whole thing seems to exist more as a talking shop than a doing shop. While people are busy implementing OpenID and OAuth, marking up their pages with microformats, publishing RDF, writing great tools and working on new specifications and vocabularies - both on their own and through things like the W3C and microformats.org, DataPortability just chats. Various people I know have told me in private that they have no time for the DP group. Don't worry, I won't be outing them.

The group seems to have attracted a lot of endorsement and coverage - from companies like Facebook, Digg, MySpace and so on. Why is that interesting? Well, it's really not. At the same time, these companies have started publishing more open data and implementing open standards. That's actually interesting - because developers and users can start actually using them. The whole thing very much reminds me of the parable at the beginning of Concluding Unscientific Postscript - Oh, dear author, you have published your big book; but when is the next one coming out? You can get as many bloggers writing about a project, and as many Web 2.0 startups endorsing something, but show us the fucking code already! Not to pimp or anything, but this is exactly what Microformats and Linked Data and plenty of other projects do. Developers can't implement press releases - they need substance. Meanwhile, those wanting to avoid actually implementing the already existing standards and good practices can fob everyone off by joining a talking shop, safe in the knowledge that they will just go on talking forever and never actually force them to do anything the talking shop supposedly stands for. Is this not obvious?

Almost every suggestion made seems to end up with someone getting hot under the collar about privacy, but making no practical suggestion as to solving the privacy problems, and also seemingly blind to all the ways people have tried to do this so far. This kind of thing is depressing, if only because it shows that people aren't very creative in how they think about data. That's why I care about the Semantic Web - it's not a choice. The data is already out there - either it's public, and being indexed by Google, or it's private and within a walled garden. If it's public, it's usually not very well structured, meaning that you need the engineering resources of a Google to do anything useful with it. If you want a big profound vision of what I think the Semantic Web is, here it is: everyone having their own version of Google sitting on their desktop, chatting with every other shard of Google. Complete parity between consuming and publishing data. Being able to move my friend relationships from Facebook to Bebo is so not the end of the story. This is why I don't like standards which presume the existence of some Google or Microsoft-like power. You need to get out of the 1980s and realise that the Web exists. The killer app of the Web is the Web. The killer app of the Semantic Web is the Semantic Web! (And porn, but let's not think too hard about that)

Another big problem I have with DataPortability is that it feels profoundly un-hackery. They have chosen Confluence, a horrible, bloated Java 'enterprise' application instead of something like MediaWiki, which is very much un-enterprisey, but just works damnit. Wiki syntaxes are another discussion, but forcing me to learn yet another crappy wiki syntax is a good way of discouraging involvement. Similarly, using a proprietary chat system like Skype rather than something like IRC is a good way to help people like me not participate. I have IRC open a lot more than I have Skype open, and I can run an IRC client on almost anything I can imagine. Some may see these things as piddly. I don't. They are important. If you want people to participate in a meaningful way, you need to let them participate using open tools. There's a reason open source software isn't developed with Microsoft Visual Source Safe.

DataPortability doesn't seem to have breathed in the true spirit of the Internet. You know, all that warm, RESTy, URI-driven, user-driven, distributed extensibility stuff. I don't know how many more times I'm going to have to explain to people that URIs exist and are much more useful than numbers and strings because you can dereference them. If you haven't read TimBL's Linked Data design issues document, you need to. It's profoundly depressing to watch the DataPortability get sucked into the void that is all that XRI/XRDS insanity (do read this post on www-tag). An aside: I'm tempted into suggesting we have a new alternative to OpenID and call it ReallyOpenID, and it would be OpenID without all the XRI/XRDS/i-names stuff. Nobody has explained to me why "=tommorris" is so much easier for my poor brain to understand than "tommorris.org", except that the former I have to buy from an i-names provider, while the latter I can get from any old domain registrar. Anyway.

I'm planning to start working soon on solving a well-defined problem that we've been seeing with the way that some people have started using APIs, microformats and Semantic Web data (specifically, FOAF and similar profile description formats and protocols). I've identified a clear problem, going to sit down and write out exactly what the problem is, how users feel about it, and how we can solve it - then try and put together a simple, well-functioning way of solving the problem across a variety of approaches. I won't be inviting mass participation, setting up bloated enterprise wiki systems, having waffly mailing lists or sending press releases to TechCrunch. Committees gave us XML Schema. We had to wait for James Clark to give us RELAX NG.

I hope that the DataPortability people actually do something soon. Go on, prove the critics wrong. Until then, I'll be hanging out with the cool kids over at microformats and the Semantic Web Interest Group - on our nice, open IRC channels, discussing real stuff. DataPortability needs to figure out what developers want. Hint: it's something broadly along the lines of rough consensus and running code.