Are you worried yet?

Published at

Thursday 21st February, 2008

Tagged as

• Security
• Encryption
• Research

I haven’t had a chance to read the full paper yet, and I doubt it’s practical for me to verify the results, but a team including Ed Felten has figured out a way round on-disk encryption. The technique sounds dangerously simple to implement, and even if it isn’t totally reliable means you can’t trust on-disk encryption any more without considering the hardware platform around it. Which is perhaps not overly surprising; this is one reason I tend to avoid having nasty things like CD/DVD ROM drives in machines unless they actually need them. But now you need to go further: lock down the BIOS (ideally by flashing it with a new one) so you can’t boot off USB or similar; and somehow make it so you can’t easily remove the memory chips from the computer. I suppose if you’re prepared to spend enough money, you could simply make the entire unit unservicable, unexpandable, and pretty much impregnable.

It’ll still get broken, though. As with so many things, when it comes to security of data it really does seem like the only winning move is not to play.

I didn't tell you so

Published at

Tuesday 29th January, 2008

Tagged as

• Foolishness
• Arrogance
• Google

Okay, so I don’t think I’ve ever actually said this in public, but we were all thinking it, right?

It doesn’t say much for the quality of those 150 people Google hires every week.

From Joel Spolsky.

EuroIA and Mr Pepys

Published at

Thursday 20th September, 2007

Tagged as

• Information architecture
• Samuel Pepys
• History
• Inspiration

In a short while I’ll be heading off to Gatwick, and then to Barcelona for EuroIA, the third European information architecture summit. I’m not entirely sure what I’ll find there, but one thing other people will find is a poster (or lots of bits of paper taped to a wall, at least) trying to extract IA goodness from the life of Samuel Pepys. I freely admit that it doesn’t make much sense, and I’m not yet convinced that I’ve found anything out that’s useful to other people, although I’ve learned a lot in the process. I’ll be putting the images and drawings I made up on Flickr when I get back; unfortunately I’ve run out of time in the mad push to print onto tracing paper, wrestle our A3 colour printer into obediance, and other fun things.

If you’re there over the next few days, come and say hi, and try not to ask me anything really hard about the 17th Century - I never thought I’d say it, but I’m kind of sick of it. At least for the time being.

Portable social networking

Published at

Friday 3rd August, 2007

Tagged as

• Social networking
• Portable social networking
• Python
• Laziness

There's been a fair amount of discussion recently about the idea of portable social networking - that when you sign up to a new site, it should be really easy to pull your contacts and so on from whatever you've used before. Various people are attacking this problem, in various ways - in June, at Hack Day, I took the external route of writing a library you can use to pull down all your contact lists from different social networks, merge them, and list the results. It wasn't very pretty, and was really a demonstration of what could be done more than anything else.

Since then, a couple of people have expressed interest in it; either because they want to use it directly, or because they want to check how other people have tackled this problem before writing their own. You can download the python source code: psnlib.

I'm lazy at the best of times, so there:

• isn't an autoloader for the social network-specific plugins (it's about five lines of python, and I have it somewhere if anyone really wants)
• are unit tests (because it was quicker to write that way; however some of them have hard-coded specifics like who the first contact in my flickr list is)
• are my API keys, allocated for Hack Day, still in the source

There's probably some other stuff wrong as well; but you might find it useful nonetheless.

Hack day

Published at

Tuesday 12th June, 2007

Tagged as

• Hack Day London

So Hack Day is just around the corner - a few more days and we’ll all be rolling in APIs, pizzas, the fluff that gathers anywhere that several hundred people are, and so on.

I have no idea what I’ll be doing yet; I had a daft idea which turned out to be impractical without a lot of painful setup, and another idea which is just daft (and so probably just about worth doing). Things are floating around in my head, and with luck something will pop out just in time for me to pull it off. Twenty-four hours is a long time, and I do my best thinking at night anyway.

Done deal

Published at

Tuesday 17th April, 2007

Tagged as

• Egotism
• Google

I’ve never really abused job perks. Sure, I may have mailed out some personal items using the company franking system, and I might have done some photocopying for a launch party for my book, but I haven’t charged undue things on expenses, or stolen the phone system or anything.

Well all that has changed: the top ten hits on Google for "James Aylett" are my homepage (twice, for complex reasons), two pages from The Uncertainty Division, one from Amazon, the two previously mentioned of random things I’ve done with my friends, two from Talk To Rex, and the ad:tech one. The Bath student has dropped off the front page!

Coincidence?

Back again

Published at

Monday 9th April, 2007

It's been a long, long time since I wrote on a personal blog - back in August 2001, in fact, when most people were still calling it a diary. (Okay, most people had no idea what anyone was talking about when it came to the Internet in general.)

Things have changed since then. I've blogged three theatre tours (Our Of Your Mind in 2003, An Extremely Memorable Emergency in 2004, and Impromime in 2005/2006). I've written a film/satire/random rubbish/Doctor Who blog, with James Lark, since late 2004, which now lives at Talk To Rex, also home to our adventures in film and print. Atom has arrived, which doesn't mean much to you, but makes my life an awful lot easier.

Back whenever it was (1999?) that I started putting my diary onto the web (before then it was on a telnet-based BBS called Monochrome), I wrote a series of Perl scripts to manage everything. There was a little header-and-payload file format, and Perl would run merrily over that, figuring out what was going on, then spit out all the calendar system, and the entry files themselves. (This was before I understood what the hell Dave Winer was talking about with RSS, and by the time I got it I'd stopped writing.) This lasted until the end of August 2001, after which I went to New York for a month to study film making, the world changed, I came back, got busy, got busier... and never really picked it back up again.

I can't guarantee I'll pick it up again properly now.

Of course, the Perl system just wouldn't do in this day and age. For a start, I haven't used the language seriously in years, so I have no idea how to do modern things in it, like, umm, opening a file, or reading in a load of XML, filtering out the bits I want, and rendering paged Atom archive files. So I'm now running a new set of code, in Python, which is heavily XML driven. In particular, this means that the source of each diary entry is actually Atom itself. No grungy hacking about for me, no sir. It's Atom with embedded XHTML right the way down.

This has presented some interesting challenges, most notably that I had to import about 1500 old entries into Atom, meaning I had to figure out how the original Perl scripts worked and replicate them all in Python. Right now I'm writing new entries by hand in Atom, but I might go back to the old source format, which is pretty neat for something I dreamed up on the spur of the moment while sitting in an office in Cirencester. Either way, it seems to all hang together.

So here we are again. I've taken the opportunity to redesign my site at the same time, and link to all the fun places you can find me on the web (look under Elsewhere, which is probably at the bottom right of your browser window). I've made a big effort to use semantic HTML properly, all that jazz, but no doubt not quite enough, so call out if something's wrong. I don't want to commit too much, and then disappoint anyone, so let's just say hello again now, and see what happens next.

Oh no!

Published at

Monday 9th April, 2007

Tagged as

• Egotism
• Information retrieval
• Google

For many, many years, all ten entries on the first page of Google for "James Aylett" have been about me. It’s interesting (if you’re me) to watch different sites move around in the Googleverse as the rankings change. My own website vies against The Uncertainty Division and Talk To Rex, and various mailing lists and so forth. It’s been reassuring that, no matter how much Google sucks, they’re getting the important things right.

But no longer. Interesting things are afoot at Google. Firstly, a pretty much irrelevant page has crept into the top ten. It’s entirely about me (look, shiny!), but it’s also pretty useless, compared to say random connections between me and my friends, which is on the second page on Google. How it got a higher rank is a mystery (well, actually it isn’t, but let’s not start deconstructing the latest generation of PageRank while I’m on a roll).

Worse is that another James Aylett has appeared. His homepage is heavily commented, and apparently auto-generated, with almost no information. (This tells us something else about PageRank April 2007, of course.) In any case, while it is of course possible that people may Google for “James Aylett” and want to find him, they of course will be delighted if they find me instead, so it’s my duty to wipe him back off the front page. He’s only at position seven, so it shouldn’t take too long.

Yes, it is petty for me to rel-nofollow the link to my competitor. But in a world run by Google, he is my nemesis, and must be dealt with accordingly.