bellyphant → page five

Perhaps I should start setting the white balance properly on my Pentax *ist DS — it would appear that the camera actually pre-processes the raw data coming off the sensor based on what you’ve chosen as the white balance setting. This sort of goes against what I’ve been taught about raw files before — that the white balance setting (and indeed most of the picture processing settings) doesn’t matter. It would appear it does — the camera apparently does per-channel gain control, and then the processing software scales the channels appropriately afterwards. Sort of handy, but it would have been nice to know earlier: I always leave mine on “Daylight” white balance, and just fix stuff up afterwards — perhaps I’ll have to change this behaviour.

Also it would appear that ISO 1600 and 3200 are just software multipliers. If you look at the raw file data, there are holes in the graphs — indicating the use of a software multiplier. Which means that you’re better off using ISO 800 and underexposing by 1 or 2 stops (for ISO 1600 and 3200 respectively). It also seems that ISO 200 may be a bit weird, and that the sensor is actually best at 400.

(Note that all this only really applies to the Pentax *ist DS — it would appear that Canon cameras don’t do the same thing, although the Nikon D200 does: it’s an advertised feature.)

RAW vs White balance on the DPReview Pentax SLR Forum — you need to read all the posts, people were initially skeptical but seemed to be convinced in the end.

Needless to say, at some point I’ll have to test this for myself. Fetch me a grey card!

I can see why people might find the way that Google (the search engine, not the company) works a little strange. I’ve now basically dropped out of the listings for “tube snot”. (I generally do a check every so often just to see where I am in the rankings… It’s not weird at all… really! I have to know about this stuff for work.)

I guess it would be understandable — the page is over a month old now, and I haven’t really mentioned it again. What is a little weird is that it doesn’t show up as indexed as all — Google has actually completely lost track of the page. I have it in my sitemap that it gets updated monthly, perhaps Google just ignores all pages it thinks aren’t going to be updated?

Reading various SEO sites — urgh, it’s definitely a strange industry — they have all these mystical beliefs about how Google works. I can see how those might develop; it doesn’t seem to work in a particularly measured way. It’s more like a core algorithm with a bunch of hacks chucked on top of it. Which might not be too far from the truth?

A few weeks ago, while watching TV, I came across a miniseries called “Tsunami the Aftermath”. In it were some known actors (e.g. Toni Collette) and some lesser known ones, such as my brother-in-law’s brother, Leon Ford. Needless to say I was slightly surprised: he’s been in a bunch of movies and TV shows in Australia; but this was the first time I’d come across him randomly.

Anyway, after watching the show I started to wonder how he (and more importantly, I) would rate in a “Six Degrees of Kevin Bacon” game. We didn’t have the internet at the time, and the best we could come up with was:

1. Leon Ford → Tsunami the Aftermath → Chiwetel Ejiofor
2. Chiwetel Ejiofor → Love, Actually → Denise Richards
3. Denise Richards → Wild Things → Kevin Bacon

(Incidentally, I just got some spam from “Denise Richards”. It wasn’t that exciting.)

There was also a link involving Toni Collette, Cameron Diaz and Tom Cruise. But that was obviously longer. So I left it at that, with Leon on three degrees of separation — that meant I was on five degrees (or four if you leave out my brother-in-law; I was willing to do this to fit into the six degrees…)

Tonight I discovered the Oracle of Bacon. Using IMDb data, it allows you to find the Bacon number for any actor. Apparently Leon has a bacon number of two; i.e. he’s two degrees away from Kevin Bacon. So I’m closer than I thought: I’m actually four degrees away. You can also use the site to connect any two actors, and it also has a page analysing the best centres of the Hollywood universe — it would appear that Kevin Bacon is in fact not the most connected actor.

Hooray for Graph Theory!

I’ve been a fan of Shaun Inman’s blog for a while — it was actually coming across the original site of his a couple of years ago that caused a bit of redesign of the UI for ICE: suddenly the UI we’d done looked amateurish. So we sat down and came up with something we were happier with (maybe with a fair amount of “inspiration” — not too much — from sites we liked such as Shaun’s).

Anyway, somehow-or-other I ended up back at it about a month and a half ago. At some point he had redesigned the site to use “The Heap” — basically stories are coloured according to month of the year, category and age. (Obviously the “age” colouring changes as the posts get older.) It’s a marvellous piece of ambient UI engineering. Who knows if it’s actually useful; at least it makes all the pages of the site look slightly different :). I tend to think it works a little better than that, though.

The heap concept was interesting enough that it’s actually one of the things that made me setup this blog again, and do it by writing my own software — I was hoping to put some interesting things in (perhaps I still will, someday…) Back to the point of the post: it’s a cool idea, and it’s one that would actually translate well into a whole bunch of UI situations. If you haven’t already seen it, you should check it out.

We’re off to Brussels (wikitravel) this morning! (Urgh, it’s early.) Hopefully we’ll get to see Bruges as well, but since it’ll involve random traversal of the Belgian rail network, who knows.

We haven’t exactly done much research, but Eurostar had a package deal with a fancy hotel, so it sounded like a good idea at the time. This will be the first time we’ve left the UK since June, so it’s got to be good just for that! Will report when we get back.

In the spirit of adding as much random crap to my blog as I can, I’ve just added the del.icio.us tagometer. I imagine it will soon reveal that none pages of my blog are tagged in del.icio.us. Interestingly, “http://localhost/” is tagged by about 200 people… (“Now what was that site again? You know the one — the one that’s right in front of me all the time; that host that’s local?”)

In other crap-on-a-site news, I’m still trying to tell whether I like the Snap popup pictures thing. I think I’ll leave it on here for at least a little bit longer.

Finally we have internet at home! We’re using Virgin.net — mainly because it was cheap(ish), and because they didn’t have a 12-month contract. Service seems okay so far; not getting the full 8mbit (nothing like it, more like 1 - 1.5mbit), but it is peak time and the speed is supposed to adjust.

But it sure is handy having the internet at home; so much stuff to read! (And people to talk to; the Australia-UK timezone difference is fairly inconvenient if you don’t have internet at home.)

Do the “analyse” and “defragment” buttons not do anything in your Windows XP Disk Defragmenter? You probably need to make sure the “DCOM Server Process Launcher” service is running. I had mine turned off. Apparently this is a bad thing to do — it would appear that it is no longer enough for defragmentation to run as a normal program; it needs to be a DCOM process connected somehow to some service. Way to Microsoft it! Also, you don’t seem to be able to start that service normally, the only way I could make it work was to set it to “Automatic” (it was on “Manual”) and reboot.