Google Image Labeller

I've recently discovered the Google Image Labeller (or Labeler as they call it) and am slightly hooked. However, I remain to be convinced that it's of any use at all.

It appears to be an attempt to improve image searching by "crowdsourcing" human cataloguing of images. It's presented as a game. You and a random partner are presented with a series of images which you each label. You score points when you and your partner assign the same labels. Certain labels are "off limits" - presumably these are ones Google already has enough votes for to believe they're valid. You are given two minutes to do as many images as you can. When the time is up you are shown what labels your partner chose for each image.

This all sounds quite clever, but I do wonder if the labels are going to be any good. Firstly the images presented are too small. In many cases I simply cannot see what the picture is of and so cannot label it meaningfully. Secondly the time element gets in the way. Thirdly your partner can "pass" on an image after which there is no point in labelling it, which can be annoying. Lastly, on seeing some of the labels my partners have chosen I find myself thinking "what is this guy on?".

Wisdom of crowds? Maybe, but if you do an image search and get crazy answers, perhaps it's because the crowd is full of village idiots. (Or mavericks, as I believe they're called in Alaska. Whoops, little bit of politics there.)

Happy birthday, NASA

According to today's edition of the always interesting Astronomy Picture of the Day, today marks the 50th anniversary of the creation of NASA.

I hope the next 50 years are as exciting as the first.

Slartibartfast syndrome

In a discussion with my manager today I described a tendency I have to approach any new area of work by finding something in it that's close to my previous work. Not only had my manager already noticed this problem, he'd given it a name - the Slartibartfast syndrome.

It's nice to work for someone who has the same cultural references.

Nothing to see here, move along please.

Those of you who read this blog avidly (you know who you are, even if I don't) will have noticed that I haven't written anything since early June. That's over two months in which I've not felt moved to put finger to keyboard here. I've been wondering why that is.

I could have ranted about the new look on Facebook or the BBC web pages, with their uncomfortably wide fixed width. I'm not changing my preferred browser window width for anybody, so it's their fault if I don't see the whole page now.

I could have written an essay entitled "what I did on my holiday". Sarah and I went to the Loire valley and stayed in a gite near Ste Catherine-de-Fierbois. We had a leisurely drive there and back, staying in hotels in Compiegne, Pouilly-sur-Loire, Angers and Beauvais. The weather wasn't too wonderful, but we enjoyed visiting all the sites and had some great meals out.

I could even bore you all to tears by going on about how much I like my new (to me) car, and how quickly I've got used to its high tech 6-speed computer controlled gearbox.

But I just haven't felt like writing anything.

Bringing colour back to Dr. Who

Over the years I've been at Kingswood Warren I've occasionally worked on projects that primarily benefit the BBC archives' restoration work, the most successful of these being the transform PAL decoder. A while ago I was introduced to another possible restoration job - recovering colour from black & white film recordings. This is all explained in some detail on the working group's wiki, but the general idea is that black & white film recordings sometimes captured the PAL colour sub-carrier patterning (when someone had forgotten to filter it out) and it might be possible to recover this colour.

My first, and second, thoughts were that this was a forlorn hope. You'd have to realign the film image with its original TV scan lines to stand a chance, and this just didn't seem possible. I'm glad to report that I've been proved quite wrong, as you can read in this thread on the Doctor Who restoration team technical forum.

The Moon and Mercury

Last night I took my first astro photos for a long time. There was a near conjunction of the Moon and the planet Mercury. I set up my Canon 350d and Sigma 70-300 zoom lens in my bedroom, pointing out through an unopenable (and dirty!) window towards a cluttered horizon.

By about half past eight it was beginning to get dark enough that I thought I might see something, but it was only at about ten to nine that the crescent Moon suddenly popped into view. I started taking pictures, one after another, playing with exposure time, focus and focal length in the hope of getting a good shot. I knew Mercury would be to the left of the Moon (thanks to the Planetarium program on my Palm PDA) and that it should fit in the field of view of a 300mm lens, but I couldn't see it in the viewfinder or with the naked eye.

Of all the pictures I took, this is the least worse. Mercury is clearly visible, and it's also the youngest new Moon I've photographed - just 1 day, 7¾ hours old.

Flash fashion

Why oh why oh why are so many websites falling victim to the current mania for Flash graphics? Obviously it's a good way of doing complicated things like the Scrabulous game or the BBC iPlayer, but just what does it add to a straightforward static page, apart from download time and accessibility problems?

The worst example I've been hit by recently is the Dilbert site. I used to visit this every morning, for a useful antidote to another day at work, but then they added so much futile Flash that it's become almost unusable.

However, all is not lost. I've recently started using Google Reader to subscribe to various web sites' feeds (such as the BBC Internet Blog or my internet connection daily usage figures) and have added the Dilbert site feed. This delivers just the cartoon strip, with minimal overhead. Perfect!

Fabulous Scrabulous

I think I might be addicted to the facebook application Scrabulous. I've only been playing since February this year, but have become completely hooked. I'd even go so far as to say it's the one thing that makes facebook worth bothering with.

Like many others, my first thought on seeing Scrabulous was "how are they getting away with it?" It's so obviously based on the SCRABBLE board game, but with no apparent connection to Spears / Mattel / Hasbro (the owners of SCRABBLE in various markets), that I was surprised they hadn't put a stop to it. Then I found out that they had already tried to.

Recently, Mattel has launched an official SCRABBLE facebook application. Feeling that I ought to respect the IPR of the game's owners I installed the app and challenged one of my regular Scrabulous opponents to a game. I am singularly unimpressed.

It starts by asking you if you live in north America or the rest of the world. Fair enough, as different companies own the rights in the two areas. But it doesn't remember this info - it asks every time. The game board is really slow to load, and is an ugly shade of turquoise. They've gone to some trouble to make it look like a travel scrabble board, with a 3D-look grid to stop the letters sliding off. I've got news for Mattel - electronic letters stay put. They've even gone to the trouble of replicating the little triangles around the double / triple letter / word squares that allow you to calculate your score without lifting the tiles. Hardly necessary when there's a computer to score for you.

Then I played my first word. Aaaaargh! It's got bloody annoying sound effects. Click, click, click as each letter goes down, whoosh as you take them back, sneeze as you sort or shuffle your letters and a ghastly muzac style chord as you play your word. How do I turn that off? A frantic search for user preferences, settings or whatever found nothing.

A bit later into the game I noticed that after playing a word it puts a sparkly star effect on the new letters. That must have taken ages to get right, and is completely and utterly pointless.

Last but not least it's a bloated memory and CPU hog. It's continually transferring data, even when neither player in the game is doing anything. It boosts Firefox's memory usage by about 65MByte, compared to Scrabulous's 5MByte. It consumes all available CPU time on my aging office PC, compared to Scrabulous's 44% or so.

To be fair, the SCRABBLE application is still a beta release, and some of these points may yet be sorted out. Other people say they like the appearance of the board, and I'm more than happy to concede that my ideas of aesthetic design are not representative of popular taste. But for plain simple playability, which is what matters, Scrabulous is way ahead at the moment. I can't help wondering why Mattel didn't just get on a flight to India and make Scrabulous's developers an offer they couldn't refuse.

Of course, Scrabulous is not without its own problems. For example, the stats calculator thinks I'm playing one more game than I actually am. Their online support is as useless as support lines often are: I asked if, when looking at an archived game, there was any way to print it, and got a reply telling me how to look at an archive game, which my question made clear I already knew how to do. It's also a victim of its own success, with games slow to load at peak times.

The shed - part 4

At last, the shed is finished. I wasn't sure if I should put it up on Friday or Saturday - I looked at several different weather forecasts, some said rain on Friday, some said rain on Saturday. Then Sarah got a short notice singing job on Saturday, so that decided it. We spent a very wet Friday screwing the panels together.

On Saturday morning I put on the roofing felt and all the finishing touches just before the rain started again. Then in a gap between showers I put in the window glass.

Next job, fill the shed by partially clearing the garage.

The shed - part 3

At long last, I'm ready to start putting up the shed. First I laid a sloping path along the side of the shed, using 450mm square slabs, and a raised path along the back, using 440mm x 220mm concrete blocks. Then I laid 12 "pads" to support the shed, using half slabs, bricks or blocks as required to get the right height. Getting these correctly positioned and level was not easy!

Finally I put in some blocks at the left to contain the soil under the compost bins and concreted the gap between the path and the fence. I don't want anything growing here. I filled the gap between the blocks and the back fence with the rest of the "granular sub base" material. I may put proper slabs over this later.

The photograph shows the finished work, just before I back filled with some of the previously excavated soil.

The shed - part 2

The weather over the last few days has been almost as bad as it could be for garden works. Cold, windy, rain and now snow.

After digging out the very soft soil mentioned in my last post, I've reached a heavier more clayey layer. (But not yet down to the solid yellow clay I found when putting in some fence posts last year.) Overnight rain turned one corner of the hole into a puddle, with very soft and squidgy ground below.

Yesterday I barrowed most of the 850kg bag of sub base aggregate from next to the house (where the lorry crane dropped the bag) to the hole at the far end of the garden. Having spread it out, roughly levelled it out and tamped it down a little, most of the area now has a firm surface I'll be happy to start laying the slabs on. However, the muddy corner is still very wobbly. The sub-base has yet to make much improvement. I hope that if it dries out a little over the next day or two I'll be able to compact it properly and make it firm.

The shed - part 1

I'm taking some time off work to put up a garden shed. It arrived on Monday, but as I'm nowhere near ready to assemble it the panels are stacked up on my back lawn, with a tarpaulin to protect them from the inevitable rain.

After a lot of deliberation, and several changes of mind, I've decided not to take on the work of a concrete base for it. Instead I'll lay paving slabs to support treated timber bearers, which then support the shed.

After stripping the top inch or two of soil and vegetation over the last three days I've realised that my waterlogged clay soil is just not up to having paving slabs laid on it (with a 50mm bed of sand & cement). Pressing down on the soil with the ball of my foot produces visible movement an inch or more away. I know that, come summer, this soil will be rock hard and cracked.

So I've started digging down another couple of inches to where the solid clay begins. This is only a bit firmer, but I've ordered a jumbo bag (~850kg) of "granular sub base" which should produce a firmer surface once I've spread and compacted it. I've also got to find somewhere to hide a huge pile of poor quality soil.

None of this would have been possible without the brilliant Paving Expert web site. It's taught me everything I know.

The joy of oxtail

Last weekend I cooked an oxtail stew that's keeping me going through the week. I've had 5 servings from it so far, but now there's no meat left, just a damn good stock which should make a good soup tomorrow if I throw in plenty of pasta.

I used a Daube recipe from Hugh F-W's meat book. Hugh specifies shin or skirt of beef, but the nearest thing I could find in Waitrose was oxtail. As well as the ~1kg of this, there is 150g of pork belly, to add even more unctuousness. I'm not sure it's needed. Left to get cold, the stew solidifies into a firm jelly, with only a few mm of fat on top. I'm tempted to buy some more oxtail just to make stock - the result has to be better than the overly runny pseudo stock (ground bones and mince) you can buy. It might even be cheaper.

Is my job safe?

I've been mulling over an article from the Sunday Times for more than a week now, and I still don't know what to make of it. It's a report of an interview with Zarin Patel, the BBC's director of finance, which says:

"Patel will also examine the future of BBC’s research and innovation centre, which could be commercialised. The division helped to develop industry standards such as high definition television and Nicam digital stereo. It has an annual budget of £30m."

The BBC's Research and Innovation department is where I work. It used to be the engineering R&D department, but has been renamed several times in the last couple of years. It's been continually reorganised as well, with a steady loss of staff. We're due to relocate from Kingswood Warren to west London this year, which will mean a further loss of staff. Now it looks as if the woman who controls the purse strings wants to sell us off.

Ashley Highfield, head of Future Media & Technology, sent out an email saying "nothing to worry about" soon after the Sunday Times article hit the streets, but neither he nor Patel has issued any public retraction of the story. Highfield's email makes it clear they will be looking for "partnerships" to get best value. A few years ago BBC Technology was sold, lock, stock & barrel, to Siemens. They are now described as "partners".

The last thing I want is to be TUPE'd to another employer in two or three years time. I'm prepared to put up with the relocation to west London, as I plan to take early retirement in 2014 and so will only have to tolerate the stress of commuting for five or six years. However, if my early retirement plans are to be blown to bits by being transferred out of the BBC to an employer not of my choosing, shouldn't I just go to another employer now?

Back on the bike

A few years ago I used to cycle to work almost every day. Then I moved house, which made the cycle journey longer but shortened the car journey, had some surgery, which stopped me doing anything physical for a while, and got lazy. The cycle commuting pretty well stopped.

Earlier this week I received my car's VED reminder, which prompted me to think "shouldn't I have had an MOT done by now?". I dug out the last MOT certificate and found it expired two weeks ago! Normally I receive a letter from my garage reminding me to arrange a service and MOT in January. I phoned the garage, but was put through to another garage entirely. It seems my usual garage has gone bust, which is a shame as it was conveniently close to work and staffed by good, honest people. Anyway, the car's due to go in to this other garage on Friday.

Until then, I'm back on the bike. It's a bit of a shock to the system, and my legs are not liking it. At least the weather's being cooperative. The frosty mornings might be a bit cold, but it's a hell of a lot better than wind & rain. Cycling through the park this morning, with layers of mist creeping over the trees and distant dog walkers silhouetted against the rising sun, really made me wish I had a camera with me.

Twickenham

Had an interesting day last Friday setting up machines in an outside broadcast van to record a dozen HD camera feeds of the 6 Nations England Wales rugby match on Saturday. This is for the iview project. This creates a moving 3D computer model from several camera views, which can then be rendered from (almost) any viewpoint to create a (moving) virtual camera view. Sport are very keen to have this, but there's still a lot of work to do.

Anyway, to capture the camera feeds we used 6 rack mount server PCs, each with 6 500GB drives and a dual channel HDTV capture card. It's important to grab uncompressed images, as any compression artefacts would interfere with subsequent processing. The data rates involved are pushing these PCs to the limit, and not everything went smoothly on the day. I'd spent the previous two weeks learning more than I ever wanted to know about Linux's disc caching strategy and memory management, but still didn't get to the bottom of why the process is not 100% reliable.

Test card on BBCi

For some time I've been aware that my TV's setup is drifting slightly, and I'm losing more picture off the bottom of the screen than I would like. Borrowing a test pattern generator is a bit of a hassle, so I've not got around to sorting it out. Now someone's revealed how to get a test card via BBCi on DTT (aka Freeview), so once I dig out the info on how to get into my TV's service menu (it's in my office, somewhere) I'll be able to tweak things a bit.

My joy is only slightly tempered by discovering that an old friend, who has worked on test card generators, had known about this "secret code" for months.

Does no one read what they write any more?

A colleague has drawn my attention to this news story from the BBC. The story is interesting in its own right, but I was struck by the phrase "A BBC journalist and an actor posing as fake academic were invited ...". I can't help feeling they'd have done better if they had posed as real academics.