Robin Green | 6 May 2012
No comments

Pragmatism in computer science research

I can’t help but think that if computer science grants had more industry input into the writing process and more importantly the process of selecting grants to fund, we’d see a great wailing and gnashing of teeth and then some far more useful and relevant research being done.

For example, there has been research into “semantic wikis” and “semantic knowledge systems” for years. While fun, the sums invested seem to have been out of all proportion to what is basically quite a simple idea, with quite low uptake so far. (The main beneficiary of this particular line of research, other than the researchers themselves, may have been IBM, who I think received some EU funding to help it build its proprietary software. But I don’t know if that software was successful, so perhaps even IBM got hoodwinked.)

Also, the open source software created in those projects is frequently abandoned after production – in one case quite deliberately so, immediately after funding ceased, in order to focus on the next “latest, greatest” semantic knowledge system, which happened to involve a total rewrite. It’s great that there’s some open source code, but no-one seems to care about it or maintain it, which indicates something fundamentally wrong with the whole process.

It’s also possible that semantic wikis are actually brilliant, but not enough effort has been put into commercialising and disseminating the existing research. Again, industry advice could help here.

Another benefit of this idea might be that if grant assessors started to say, “No, this distributing computing calculus of yours has absolutely no relevance to bridging the digital divide in India”, funding bodies might realise that if they want to keep funding the amount of “basic research” that is being done, they ought to be honest about it and stop attaching spurious criteria to all funding tranches that grant writers then have to write nonsense to “fulfill”. By all means have practical criteria for more practical funding tranches, but don’t try and pretend that all research will have practical applications in the next 5 years, because some of it won’t.

Category: Tech | Tags:
Robin Green | 6 May 2012
No comments

The unthinkable – state aid for Open Source software

Originally published here with 7 comments

Why is it OK for the UK government (a Conservative-led government no less) to propose competing with private landlords in order to subsidise UK tech startups (something that already goes on to a limited extent, it’s just outsourced to universities), and OK for Ken Livingstone to propose competing with estate agents to help poor people (something that already goes on, it’s called social housing), and it’s widely applauded when governments introduce enforced open access or open textbook policies to compete with Elsevier and other big publishers to help everyone… but somehow it’s unthinkable Marxist heresy to propose that the government fund open source software to compete with Microsoft to help startups (and indeed everyone else)?

I know politicians don’t know their Java from their Ruby, but still…

Perhaps that’s it, though. Maybe virtually the entire British political class (including the political staffers, pundits, wonks and columnists) haven’t got a clue about programming, since they all studied things like English or PPE at university, and the civil service is chronically short of expertise due to all this outsourcing and privatisation that’s been going on for the last couple of decades. Perhaps, though they recognise the centrality of tech to this new economy, they think that spending a bit of money on computer science research is the only legitimate role for government here, due to the universal received wisdom that “government should keep its nose out of everything” (except when it shouldn’t).

Of course Microsoft has been arguing against any positive mention of open source in government policies for years. Their argument has always been that competition or even just procurement preference towards open source from government is “unfair”.

But surely the policies I mentioned at the beginning would be unfair to landlords and/or estate agents. You might say that they’ve benefited unfairly from increases in the value of land in recent years… whereupon I’d just point out that Microsoft is a convicted monopolist.

Category: Politics | Tags: , ,
Robin Green | 3 May 2012
No comments

3 things that are posted on Pinterest that are just wrong

  1. Photos of morbidly obese kids with “funny” captions. I am all for making fun of fat adults like me, that’s fine – but they’re just kids! To those who do this: You’re an adult! You don’t do that – pick on someone your own size!
  2. “Jokes” that involve murder or hyperbolic rage. If I don’t know you, and I see that, I’m going to think you are a complete psycho. Or maybe that’s just me.
  3. Cartoons of little boys talking to little girls with “cute” sexual captions. (It’s always women who post this.)
Category: Entertainment | Tags: ,
Robin Green | 1 May 2012
No comments

Open letter to desktop (and CLI) developers who hate web dev

Originally posted here with 2 comments

You have a point. The basic underlying technology is (still!) primitive, and things are being shoehorned into where they arguably shouldn’t go, partly because of totally inevitable corporate (and airport wifi) firewall policies. (Blocking everything except port 80 and port 443 is becoming increasingly meaningless – but that doesn’t mean it isn’t still rational.)

Nevertheless, pretty much all the issues have been or are being worked around, or obsoleted. (No thanks to Microsoft, which has been dragging its feet every step of the way – ever wondered why so many web developers hate IE?)

Notably, there are now a plethora of client-side libraries (JQuery being one of the most famous) which claim to be cross-browser – the early 21st-century equivalent of cross-platform. Some of them even are. Some of them even support ancient browsers like Safari 2 (which you shouldn’t – it has an awful bug in its Javascript implementation).

And yes, some of them even support adapting to mobile browsers.

But so what? Isn’t this all a big hack?

And what? The x86 platform wasn’t? UNIX wasn’t? MS-DOS and Windows 3 weren’t?

UNIX/Linux took a long, long time to get rid of some of their laughable hacks (like /usr). As did DOS/Windows/Windows NT. This kind of thing has a long pedigree.

More to the point, some things are actually being done better this time around. Did CORBA have caching, or support Twitter-scale scalability? (Maybe it did, but I never heard of it.) You may think Java Swing or Qt is the bee’s knees for GUIs (really?) – but isn’t Swing actually pretty imperative and old hat these days, and isn’t Qt adopting Javascript and becoming in that sense more web-like?

Maybe worse is better.

Aren’t Swing and Qt huge monolithic libraries built by out-of-touch corporations and inadequately maintained and improved, and aren’t say JQuery and Firefox much better, healthier ecosystems, more responsive to developers’ concerns, more frequently released, more easily hackable (or at least perceived as such, in terms of the ease of writing extensions)?

And RIA development in the browser, and other web trends, are sneaking distributed computing – and knowledge of such – back into pedestrian, everyday application programming. This is wonderful! This is one of the best things that ever happened to computing!

Think of web development as like a massive franchise reboot. Being done extremely slowly and painstakingly, and still not finished.

The end result is going to be awesome.

Category: Tech | Tags:
Robin Green | 29 April 2012
No comments

Another atheist giving atheists a bad name

Originally posted here, with 2 comments

Sam Harris thinks we should profile Muslims.

I will just repeat the argument of security expert Bruce Schneier: as soon as you profile Muslims, terrorists will recruit beardless white (or light-skinned) men. Or maybe even women. Whichever groups from which they can recruit, that are least likely to be suspected. As a devout Islamophobe, Sam Harris should know that the Koran explicitly permits deception in wartime (taqiyya).

P.S. Some people from the Middle East are lighter-skinned than some ethnically southern Europeans. Not that I’m suggesting that Europeans can’t be terrorists – they can be and have been.

 

Category: Politics | Tags:
Robin Green | 28 April 2012
No comments

Computer games help children to pay attention

Today I heard an old woman moan that “kids today don’t read, they don’t speak [sic], and they don’t know how to write a [handwritten] letter”. Well it’s not all bad…
This is another one of those studies where the conclusion of the study would have been thought “obvious” whichever way it had turned out.
Category: Politics, Tech | Tags: ,
Robin Green | 22 April 2012
No comments

Oil spills, GDP and pushing on a string

Originally posted here with 10 comments

I think arguments of the form “we shouldn’t pursue GDP because, ZOMG, oil spills increase GDP” are misguided.

As a practical matter, elected politicians don’t pursue GDP, they pursue votes. And voters don’t pursue GDP either.

I’d like to draw a leftfield analogy with morality. Christian theologians identified morality with what God said we should do, which had the obvious problem that it could not be extended to matters not mentioned in the Bible. Then some secular philosophers like Bentham tried to “discover” an objective, absolute and timeless standard for morality, which reproduced the religious delusion that morality is something out there in the world rather than a set of judgements made by groups of people based on reason, evidence, and shared values – values which are not necessarily universally shared. Now many philosophers recognise this point, more or less – but until then they were missing the point about the nature of morality, and as a result there was much misdirected effort.

While elected politicians do listen to economists, the people they ultimately listen to are voters. And voters already don’t consider GDP to be fundamentally very important, if they consider it at all. They may value closely-correlated concepts like reduced unemployment or growth in useful, productive activity, of course. And they may consider GDP (or real GDP) to be a useful proxy for those things, in normal times at least. But to some extent, this argument is already accepted by voters.

So once again, it arguably comes back to: do voters have sufficient influence over politicians? Do we need more democracy?

In a two-party system dominated by big money, like the US, voters have limited influence over politicians, because they can either switch between “left” or “right” or throw away their votes. In a multi-party system with proportional representation, voters can express a more nuanced opinion, and punish politicians effectively without having to betray their principles completely.

Category: Politics | Tags: , ,
Robin Green | 15 April 2012
No comments

If you can’t say it to people who don’t want to hear, what’s the point?

Originally published here, with 6 comments

I’m an atheist, but I really dislike the view that religious people should not put forward political or indeed any other opinions based on their religious beliefs.

If someone thinks suicide and/or abortion are immoral, they should be able to say so. Even protest those who assist such things. Even say things which other people might find offensive. Even on Facebook. Come on, this should not be controversial.

Likewise, I’m a liberal, but I really dislike the view that non-liberals should not put forward views based on their political beliefs. OK, racism and other offensive stuff like that is kind of a special case because it’s a powder keg, but apart from that…

Can’t the people who have these views see that this is where white male conservative Christians get the idea that they are an oppressed group from?

Category: Politics | Tags:
  Recent Entries »