Sunday, September 16
So, the penultimate episode is up.
Looks like my predictions are at least partly right.
Spoiler for Potemayo:
They also drop some hints about the ending, but I suspect there's some misdirection going on there. If I'm right, next week's episode could be rolling-on-the-floor funny.
I was copying a few* files on the new mu.nu server and it was running a little slower than expected, so I stopped the copy and ran a quick performance test using hdparm.
And I was all, damn, there's something wrong here, it's only doing about 6MB a second. (I'd expect to see upwards of 40MB a second on any modern 3.5" drive.)
So I checked one of the mee.nu servers, and I got the same disappointing results.
And then I looked again, and I realised... Those numbers aren't in MB. They're the next larger unit.
I was looking at the cached (not raw) performance, and yes, I'm getting 6GB per second from the cache.
Unfortunately, that still doesn't explain why the copy is so slow.
Answer: Ah. If I'm copying archives from both the old servers to the new servers and uncompressing them on the fly, copying files from one disk to the other on the server at the same time is unlikely to deliver 100% of the system's potential.
* A few = slightly less than a million.
Friday, September 14
I am going to turn off my style sheet now. If the theme builder is installed correctly, this will make no difference at all.
If my site suddenly disappears, well, there's probably still a bug or two hanging around.
Update: Oops. Let's try that again.
Not just my air conditioner, but also my dishwasher and my oven pull the voltage down enough to reset my external drive.
Western Digital has to be getting complaints about this thing.
Thursday, September 13
Step 1: Announce that you'll be releasing detailed specs for your graphics chips to open source developers, albeit under an NDA.
Step 2: Announce that you've reconsidered. Never mind the NDA stuff, it will be open to all.
Step 3: Actually deliver the documentation.
Step 4: Watch x.org's server melt.
Step 5: Profit!
It's a star map. The problem with star maps in general is that stars are distributed in three dimensions. The solution this map uses is that it is also in three dimensions. It's engraved inside a solid transparent block using a carefully focused laser.
Still not plushies, but ideal for sitting atop your monitor.
The inimitable James Randi on homeopathy:
Wednesday, September 12
From what I've read so far, the anime is very faithful to the manga, but the jokes are still funny even after you've seen them in the other form.
Or, How the Continuing Merger of Reporting and Editorials Came to Produce Something Good
Pennsylvania Man Claims to Burn Salt Water
Oh, great. There goes the Second Law of Thermodynamics.
ERIE, Pa. — An Erie, Pa., cancer researcher says he has found a way to burn salt water, a novel invention that is being touted by a retired chemistry professor as the "most remarkable" water science discovery in a century.Yes, it would indeed be the most remarkable water science discovery in a century if it weren't total nonsense.
John Kanzius says he happened upon the discovery accidentally when he tried to desalinate seawater with a radio-frequency generator he says he developed to treat cancer. He discovered that as long as the salt water was exposed to the radio frequencies, it would burn.
The discovery has scientists excited by the prospect of using salt water, the most abundant resource on earth, as a fuel.Water isn't a fuel (unless you have a working fusion reactor). Water is what you get when you're finished burning a fuel.
Rustum Roy, a Penn State University emeritus professor of chemistry, has held demonstrations at his State College, Pa., lab to confirm his own observations.This might not mean anything to you, but it does to me. Rustum Roy is a distinguished materials scientist, with published papers going back six decades. Which makes him an octogenarian, yes.
He is also something of a crank - no, I take that back; he's all of a crank. His written work also includes some... odd... opinions on various social issues, vitriolic attacks on the pure sciences (physics and chemistry) and bold predictions that these computer things are just a passing fad and that we will shortly all move back into villages.On top of that, he's a supporter of homeopathy, and has recently attached his name to some astoundingly shoddy research purporting to support that particular brand of quackery. (He may have been directly involved in the research; I don't know for certain. But his name is on the paper.)
Which is how it goes. Some crank claims some cranky thing, and some crank scientist (scientists are hardly immune from crankery) supports the claim.
What floored me is the very next line:
[Roy is also a specialist in "whole person healing" and Christian sexuality.]Which is entirely true.
But... Did a major news organisation just look up the references of source and make an editorial aside casting doubt on his competence? Well, it looks that way to me.
But maybe FoxNews thinks that specialists in "whole person healing" are qualified to comment on matters thermodynamic - even though they themselves put the term "whole person healing" in scare quotes.
Or... maybe not:
[Internet commentary upon Kanzius and Roy's assertion points out that creating fire from salt water is possible by first separating it into hydrogen, oxygen, sodium and chloride, then burning the sodium. However, such a process would consume much more energy than it produces.]
Tuesday, September 11
And I am too.
Running live on 1.1 now. The blue links are a side-effect: my blog was set up on a slightly dodgy 1.1 alpha, so a few things glitched when I merged it back into the mainstream. It should fix itself, though, if I do this...
Yes! Okay, those colours aren't quite right either, but that's what you get from a quick clicky-clicky.
Update: Colours are back again. The colour picker does let you put in hex values if you really want to. (Also colour names for the 140 standard named colours.)
The one remaining bug is something that didn't break that should have; namely, my banner. I changed the banner sizing from exclusive to inclusive at the last minute; either way is confusing, but inclusive is probably less so. Having the same number but assigning it a different meaning should have screwed up my banner image, but it's fine, so there's something wrong with the code there.
58 queries taking 0.4098 seconds, 327 records returned.
Powered by Minx 1.1.6c-pink.