February 2009 Archives

I believe you can learn a lot by trying to understand how your own mind works. It can help you to make sense of why you make mistakes, or why you forget things, and to have confidence that your mistakes are a natural consequence of being human.

I can't claim any kind of psychological knowledge, but I understand my own mind as a powerful pattern matching engine. It matches patterns in what I see or hear, then matches patterns in my own memory, then patterns in random noise leading to things like dreams and new ideas. And it'll do that matching of patterns with an enormous amount of subjectivity.

For me, this leads to two main failures. The first is the making of false assumptions.based on what I was expecting rather than what I objectively perceive. Today, I answered the door and perceived the person I was expecting to be arriving at around that time, even though they were someone completely different - they look fairly similar, but not hugely. But then, I've always had a bit of trouble recognising faces. On many other occasions, I've read text and seen the words as something different, to the point where I could visualise the shapes of the incorrect words on the page.

The other main failure is to do with memory. My mind certainly doesn't work like a computer. I can't program it to bring something back to my attention at a pre-programmed moment. Remembering things, for me, relies entirely on setting up something to remind me. For example, I keep supplies of general bathroom stuff in my room, bringing them down to the bathroom as necessary. When it's time to bring the new tube of toothpaste down, I'll always forget to do so unless I physically put the new tube in my dressing gown pocket. This kind of thing happens with almost 100% consistency.

I think it's important to keep this kind of consideration in mind.  So to speak.

Perhaps I should expand on my earlier, rather short, post about how the job interview went...

In my two days in Hamburg, I spent one and a bit in DESY and the rest exploring the city. My talk went well, with lots of interesting questions being asked. They have some really exciting work going on in the next few months, the group is REALLY friendly (and that particular bar is set quite high by my current group in Cambridge), and the head of the group said he'd be happy to have me back to work there starting as soon as I'm ready!

The catch, as ever, is that I have to have completed the PhD qualification before I can start. That'll mean, at the very minimum, having had my viva and done my corrections to the satisfaction of my examiners. That'll delay things, probably until around the end of April. I don't want it to be much later than absolutely necessary, but it's going to depend on the availability of my examiners.

The city is great. Very clean, and it's a nice size - lots to explore without being overwhelmed. And the accent in the North of Germany seems to be a bit easier for me to understand, but maybe that's just practice on my part.

So, now I just have to finish "that thesis thing"...

Got the job :D

A Chapman Stick is a bit like a cross between a guitar and a piano. It's a stringed instrument, but you play it by pressing the strings in different positions along the fretboard, with very sensitive pickups used to capture the sound. Players use both hands to create more complicated melodies. In this video the musician, Guillermo Cides, goes further and uses a digital looping machine to build up a progressively thicker sound over the length of the music. It's a nice moment of calm in an otherwise slightly frantic time.



It'd be a good instrument for me, I think, if I was going to take up music again. It's a bit of a shame that they cost thousands of pounds, and are all made to order with a waiting list that's about a year long...

I'm off to DESY on Monday for my sort-of-interview-chat-visit thing.  I've been given a map which has lots of circles and lines on it, where the particle accelerators are.  If you take a wrong turning on the way to the one I'm visiting, you end up literally climbing over one of them, which is above ground and covered by a long mound with grass on top.

I'm alternating between more excited than I've ever been before, and utterly terrified.

I'm going to London for the weekend to see friends.  It's going to be great.  And it's going to take my mind off the whole thing until right before I have to leave.

I practiced the talk I'm going to be giving in Hamburg next week to my group today. It's great to have an atmosphere where people will happily come to your talk even when they know that it's just a practice run for something else. I got some great feedback, which is going to take a little while to implement and will make things go a lot more smoothly next week.

One of the difficulties for me is to translate work which I performed as a microscopist for people who are fundamentally particle accelerator scientists. I'm going to feel really silly standing in front of such high-energy people and saying: "... the electrons were accelerated to high energy (300 keV)...".

It makes an odd noise a bit like a steam train when it's on the ground:

Quite a bit of pain caused this weekend due to not really understanding OpenEmbedded, but I seem to have got through it.  Otherwise, it's been a reasonably productive weekend.

Mesa now selects the correct driver.  I was slightly misinterpreting the causes of my problems here, concentrating on XF86DRIGetClientDriverName when in fact it's much simpler: AIGLX gets upset if swrast_dri.so isn't available, even if you know that later on it's going to choose myshinyhardware_dri.so.  So, adding that file to the package pulls it back into line.

There was also a segfault due to a slightly strange way that drm.ko has of doing things: a DRM device's unique identifier gets zeroed out by drm_lastclose() even though the unique identifier remains the same right until the DRM device gets deallocated at some point in the distant future.

All of the DRI initialisation pipeline seems to be happy now.  Right now there's a problem with Mesa mapping the framebuffer, which causes it to fall back to software, but otherwise it's looking quite promising.  Hopefully, this indicates the end of the "groundwork" to create a working 3D driver...

[root@rafiki:~]# DISPLAY=:0 LIBGL_DEBUG=verbose glxgears
libGL: XF86DRIGetClientDriverName: 0.1.0 glamo (screen 0)
libGL: OpenDriver: trying /usr/lib/dri/glamo_dri.so
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 4, (OK)
drmOpenByBusid: Searching for BusID platform:glamo-drm
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 4, (OK)
drmOpenByBusid: drmOpenMinor returns 4
drmOpenByBusid: drmGetBusid reports platform:glamo-drm
libGL error: drmMap of framebuffer failed (Invalid argument)
libGL error: reverting to software direct rendering
libGL: OpenDriver: trying /usr/lib/dri/swrast_dri.so
27 frames in 5.1 seconds = 5.281 FPS

Incredibly neat: