The Outsider's Inside View post#009 - Shoot and Score

After nearly a month of dealing with computer downtime and upgrades, a new SETHI datacube finally was created this morning. The last new one was written to disk just under 4 weeks ago, so it's good to have that part of my work back on track. That one was #96, so I'm finally 2/3 of the way done. A few more should finish this week, though the number of cubes running at one time has been reduced quite a bit since sidious started doing its designated job. I've got 9 cubes crunching on lando, plus 3 each on thumper and sidious, so I guess that makes 15. With 44 left to go, we're looking at about 3 duty cycles, or about 6 weeks until they're all done. We recently passed 14.7 million spectra in the SETHI database, so Matt keeps splitting tapes and I keep trying to catch up to him.

GALFA data-crunching also continues to go steadily. By the end of the week I hope to have finished the first-stage processing for all the data we have taken for the TOGS project. I've been making good use of ewen's new RAM, though it's still behaving somewhat peculiarly, like rebooting itself overnight. Why?

Looks like the San Jose Sharks will make the playoffs, but I'm worried that the Colorado Avalanche might catch the Calgary Flames. Although it would be really exciting if the Flames/Avs game on April 8 was to see who made the playoffs, that might be too much excitement to handle. I'm hoping for a good Stanley Cup playoffs, in any event. The NHL has done a good job of making its game more fun to watch again.

I was mostly right about what would happen on Jeopardy! on Friday. Should be a good rematch tonight.

> as for reboots / ewen (why?) - iT's an *Anomaly* - even though there may be a 'reasonable' explanation . . . as in 'nature of the beast' ;)

Thanks for the Update Kevin - and Congratulations re: the SETHI database and that's to Matt also . . . Keep up the Great work Berkeley . . . and take a *brEathEr* occassionally :0 . . .

> as for reboots / ewen (why?) - iT's an *Anomaly* - even though there may be a 'reasonable' explanation . . . as in 'nature of the beast' ;)

That may be so in the world of Microsoft systems. Reboots, "unexplained" or otherwise are a fault condition and should not happen. *nix systems are usually left running for months and years between deliberately initiated reboots/shutdowns.

There is a quote somewhere that one of the greatest crimes of Microsoft is in how that one company has so lowered user expectations.

They are still in court (a full decade and counting :-( ) waiting to lose for their other crimes.

Computers should just simply work. That they don't is just a very bad sloppy bodge fix.

Happy crunchin',
Martin

Well, I wouldn't put the blame solely on Microsoft. The entire PC industry and their entire quest for cheaper parts have made an entire market out of poorly tested hardware/software that have helped push the PC experience down the wrong path.

And I wouldn't say that all *nix systems are not prone to crashes. They do happen less, but the do happen.

Agreed. However, such problems in the *nix world tend to get fixed.

Another gorgeous day in the East Bay. I'm listening to the Flames vs. Predators on NHL radio, and I just finished watching San Jose beat Atlanta. It's been a good week for catching a lot of hockey.

Today was a good day. My 100th SETHI datacube got made, so that's a good milestone. I ran some successful datacube tests for GALFA data too, which is very encouraging. Best of all, a modification that I suggested for the GALFA processing software appears to have done a good job of improving gain corrections for our survey data. It's a little technical to explain, but basically the previous method worked fine for GALFA projects that observe with Arecibo in a certain scanning mode, basically nodding the telescope back and forth on the meridian. We call that mode basketweaving because when you combine the scans from all the days you basically "weave" together a bunch of scans that look like zigzag "ribbons." Targeted HI studies use that mode because it's really efficient for covering a lot of sky rather quickly, plus the observed positions that overlap between one day and another allow us to get really good calibration corrections. However, our all-sky survey (TOGS) doesn't use basketweave scans, but drift scans, where you just choose one declination each night and don't move the telescope at all. So this fundamental difference in observing mode means we needed a fundamental change in the way we calibrate the data for day-to-day beam variations, and now we have it.

We had a science meeting today; Dan's back from South Africa. The spurious pulses in multibeam SETI@home data is still an unsolved issue, but our contacts at Arecibo are looking into the problem quite closely now. I'm sure we'll find the origin soon enough. Apparently Astropulse doesn't suffer from the same problem, so that's been working in beta for a while now. I was told pretty bluntly not to expect to use bruno for SETHI crunching, but I'm fine with that. I'm good with the rate of SETHI progress at present.

At lunch I went to a seminar entitled Softwood Lumber and the American Assault on Canada's Domestic Policy Sovereignty. It was an excellent summary of the Canada-US dispute over lumber import/export duties, and how the US was found to be in violation of trade agreements but because of Canada's reliance on the US market, they didn't get a fair shake in the settlement. It was an interesting example of a more deferential Canadian mindset versus the adversarial US point of view. And the speaker was American, but he was clearly sympathetic to the Canadian side.