The side of St Paul’s

•October 24, 2010 • Leave a Comment

The side of St Paul’s

Originally uploaded by zeristor

Whilst strolling past St Paul’s Cathedral I thought this odd angle picture shot using HDR was quite a unique angle.

Cooking Tech: Bread machines

•May 12, 2010 • Leave a Comment

Being of a techy disposition I never really got the hang of cooking. I love food, and I love to eat, and slowly I have been impressing myself with my kitchen arts.

Bread making had always been a bit a step too far. All this rolling sleeves up, and looking at clocks, and working out how long things take because of the temperature has been a bit off putting to get started. As with most things though the Japanese have an answer.

Gadgetitis is how I regard the Japanese electronics market, make something to do anything. Very inventive, and looks like good fun. They have sushi machines you know. The South Koreans now have specific fermented cabbage fridges, it is amazing to see how technology can be used to augment a culture. I can’t wait to find out what the Chinese will come up with to automate their culture.

Japan, or is it just Panasonic, have a bread machine. Bung the ingredients in, work out how to programme it. Wait a bit, taking in the wafting smells of the baking bread and wait for the beep, when you get your fingers burnt trying to shake the loaf out. I do love fresh bread, you can get it in the shops it isn’t the same. They sell what they think you want, if they have it, and if they don’t then what? True it is not the same as fully making it oneself, but it is a step closer. It takes the awe out of it.

Mechanically a bread machine doesn’t amount to much more than a rotor spinning a paddle with a heating element. The software though has to work out how to get from lid shut to crusty loaf. No simple task. Bread making is quite a specialised art. Temperature and quantity all play parts as the living yeast is soaked, fed, and then cremated. It might be colder at the start, and so take longer to rise.

The front panel is quite basic, and no doubt daunting to those legions who left their video clock flashing. I am only borrowing my bread machine from my Mum as she stopped using it. In return I bake her the bread she wants.

The British cultural gadget would have to be the Teas Maid. A cup of hot tea to wake you up  in the  morning. A heating element,  and a timer. I have an idea for a quiet toaster, when the toast is ready instead of a load clang the toaster makes a soft clearing the throat noise. Particularly understated, and very quaintly English; of course it may never get heard…

Finding REST web services with DNS

•May 12, 2010 • Leave a Comment

InfoQ has an interesting article on finding REST services with DNS. Quite lovely, like REST, going with the grain of how the web works.

SOAP, had WSDL, which seemed rather clunky. How REST works is fairly standard across the board, however the format of the data in it will change. But annotating in RDF (the Semantic Web again) would make things more explicit.

Having read up and learnt about Erlang, I keep thinking that writing DNS in Erlang would be obvious. Apart from the fact there is a working fully mature system at the moment. Mmmmmm….

Isn’t there more to the Semantic web than Syllogism?

•May 5, 2010 • Leave a Comment


Syllogism. There are some words that I can’t get my head around. This one I did, finally. I have read about Aristotle and logic no end of times. Knowledge that happily flowed through me. I was caught up short as I perused Clay Shirky’s deep vault of delights. Articles that he had written several years ago.

Now I have been wide eyed at the Semantic web for a while now. I was wide eyed when I found out about email back in 1981, and distraught to see people agape at Fax machines. I toyed with the Internet, read about it in books anonymous ftp was amazing, and gopher? Blimey! People were aghast as I shrieked when I finally saw a web connection back in 1994 in the library at Heriot-Watt. An anecdote propping up no end of braying middle aged dinner parties.

So Semantic Web. The next big thing?

Medieval monks went to great lengths to illustrate their books, wonderous capitals spilling out in riotous ornate colours. Web pages too look quite amazing, but to a glum server choking on bits from around the world it knows nothing of natty pictures. But in a way the semantic web (via RDFa) illustrates the text of a web page.

More than words. It takes a lot of nous to get the meaning of a text. Data though can be larded in on the web, behind the scenes in the html unbeknownst to the reader can lurk labels and tags explaining what something is, or where. The reader can’t see it but search engines can, and so use it to deliver better results to advertisees.

OK. So only a tiny proportion of the Semantic web is embedded in web pages, virtually all of it resides in things called RDF triple stores. Triple because each piece of knowledge consists of three parts a subject, with a predicate giving rise to an object. “The sky  has the colour  blue” for example. Put enough of these statements together and you get a graph. Each part of these triple items is unique, it has its own URI to represent it. This is fine for  small area of knowledge but to meld these fields together one needs some way of translating them, and this is where OWL comes in, it allows one to work out how different vocabularies can fit together. So now you can search for similar patterns in large databases of Semantic data, using SPARQL (almost for the glory of Rome).

Now Clay Shirky’s article “The Semantic Web, Syllogism, and Worldview” predates embedded web knowledge. It is interesting to read his words, a lot has unfolded since then. The upshot is an extension of a criticism of AI, a branch of science devoted in making machines that can think. Logic is brittle, so the labels can quickly end up contradicting themselves making the whole thing pointless. As though the whole point is to deduce a deeper understanding of the world and not the rather more prosaic bait for agents scouring the web for the betterment of searches.

True now I think about it distilling an item into labels has its issues, and with that in mind you can’t construe too much. But there is plenty to be getting on with for the time being. Locations of restaurants, etc. One of the biggest users of the semantic web has been the biosciences, huge amount of data that needs to be cross referenced. The semantic web can feed their giant graph databases looking to turn up leads for new medicines. Now combing through data profitably looking for similarities, this isn’t AI to me but turning the cogs and wheels and crunching data. For AI to replicate a human I can only think of an imaginative machine. Perhaps this too is a mirage.

But murky knowledge muddies the pool. One piece of supporting evidence isn’t as much use as twenty more. This is how Cyc works, looking for a bit of evidence, but wary that what but me true might only be true locally, or just plain wrong. And now it has started to read, learning unaided by puny humans. The singularity started a year ago…

Ubuntu 10.04 LTS under VirtualBox

•May 4, 2010 • Leave a Comment

Hurrah; to an extent.

Mac OS X is a form of Unix, moreso than Linux if you look into the details. Ubuntu though is in wide demand and I have been quite curious about it, installing the software from apt-get always seemed so elegant.

Now I have run Linux in dual boot on my old Windows machine, Suse, and Mandrake ages ago, BeOS too! On my MacBook  I thought OS X would be similar enough to Linux for me to learn from it, it isn’t quite. Parallels, a virtual machine would do the job for me I thought, well it lets me run Windows to keep my hand in which is what I got it for but the Ubuntu installs always spluttered out before properly booting.

In a heightened level of effort I stumbled upon VirtualBox, founded in Germany, but slurped up by Sun and made available to one and all. A Virtual Machine that is free and works. And it does. A new Ubuntu and a new attempt to get the thing working. VirtualBox seems more polished than then Parallels 3 which I paid for, although that has been twice superseded.

MacPorts on OS X always seemed to get tangled up, I suppose I  just need to make more mistakes, and more clean installs then everything would settle down. But the apt-get seems to be the way to install software. True a new software can languish a while before the package is released, but convenience counts for a lot.

My only gripe, is I am stuck with a 800 x 600 window, like squinting through a letter box at a desktop. I’ve found and read up about VBoxMachine commands to set the resolution, but they appear to be impotent for some reason. I can leave this for a while, it always interesting to look at a new OS, challenging the usability I might have become too used to.

Now where is my guide to VI short cuts…

NOTE: A friend came up with this web site for installing the Virtual Box guest additions in Ubuntu. Works quite well, although everything seems to be more glacial. Perhpas I have far too many browser tabs open?

Which cloud am I?

•April 25, 2010 • Leave a Comment

I have been following cloud computing for ages, well at least since Amazon started their web services system some while ago.

Much as everyone is behind the Amazon system I want to do some work in Python, and Google’s AppEngine seems the best place to start, mainly because it is free (well the first 5 million hits/month are).

There seems to be very little flexibility, but you get DataStore and a few other things thrown in. I had wanted to use RabbitMQ for something, and for that I’ll need my own server.

I’ll get something working, then work out how to tweak it. This is all a big leap for me out of the Microsoft stack, but it should be great fun!

East London Line

•April 23, 2010 • Leave a Comment

When will the East London Line ever start?

First one week, then another, slipping delays even though it is a month earlier than anticipated. I spent an hour reading a book by Spitalfield’s farm the other day. 16 trains went past in an hour, it is good to iron out all the teething problems, and there have been enough premature openings when everything collapsed a day or two in.

Alarmingly Shoreditch High Street is a Zone 1 station. This means getting to anywhere is going to cost a lot more for passing through pricey central London, it does make sense being so close to the City, but people won’t be able to skirt Central London.

I also wonder about the line’s capacity, they say there is untapped demand, and with other services so cramped surely four carriages will be stuffed full of commuters. Traffic is good, but  how could they increase capacity.