Change Displayed Text SizeGrow Displayed Text SizeShrink Displayed Text Size


Saturday, February 08, 2003

Yup, she's pretty awesome. Pretty much every day I wonder how I got so lucky. She doesn't even have to think about how to make me happy, which says a lot. Same goes for me I guess - when I randomly was killing time in Fresno and picked up a Pocket Homie for her, I almost forgot I got it for her. Man, I sure wasn't prepared for how much she liked it!

We're destined for great things :)

2/08/2003 09:48:18 PM ] [  ]


Friday, February 07, 2003

By Dawn's Early Light

So I have 20, 30 pages maybe of notes, and a fairly complete page with graphics, etc. on the Columbia accident, explaining it from the point of view of an aeronautical engineer. Surprisingly enough, the news media, and NASA, have done a reasonably good job of keeping the investigation open, honest, and informative for the general public. The entire NASA team is doing a great job. So my mother of all posts is not going to happen.
CNN, in the first day or two of the investigation, decided that the ET's insulation breakaway just had to be the cause, since they had video of it. The likelyhood of 2.67 pounds of foam insulation damaging the orbiter enough to cause a breakup and reentry is a tad remote. Anyone who has taken the Kennedy Space Center tour and seen a demo of the Shuttle's tiles should know that. CNN continues to astound me with their lack of responsibility, as in the past few days NASA has said publically that the chances of the insulation issue being the root cause of the problem has been tossed out, yet CNN keeps bringing it back.
Give it up, Ted.

Now, why would little old me have a lot to say about what happened to the shuttle?
I may have been in this silly computer business for a while, but it's not the path in life I originally chose. When I was in college at the wonderful University of Maryland, I wasn't studying Computer Science, but aeronautical engineering. Specifically, hypersonic aerothermodynamics.
Yeah, try saying that 5 times fast. I dare ya!
Hypersonic aerothermodynamics focuses on the flight environment at speeds above about Mach 5 (5 times the speed of sound). That's literally faster than a speeding bullet. The fastest operational airplane in the world (retired now, unfortunately), the SR-71 family, flew at a max speed of Mach 3.4. The experimental X-15 flew up to Mach 6.7 during it's test program in the 1960s (though it only did so once, and none of it's flights spent more than a few minutes or seconds at hypersonic speeds). We don't know much at all about flying faster than Mach 5. The total flight time of all aircraft above Mach 5 would amount to less than an hour. The shuttle is the only aircraft that flies at those speeds right now - there are no X-planes flying in the hypersonic regime, no military flights, nothing.
During the 1980s and early 1990s, there were the beginnings of a renaissance in hypersonic flight. President Reagan announced the X-30 program in 1986, designed to create a family of aircraft that would not only provide cheap access to space, but fast "Orient Express" airliners that would cross the globe in minutes rather than hours. New York to Tokyo in 2 hours, etc. Over the next few years billions of dollars were spent pursuing this vision with not a whole lot to show for it. The X-30 was cancelled not long after the overall configuration was finally approved in 1994. With the death of the X-30 came a decline in interest in hypersonic flight, and here we are today still with no hypersonic flight test data to speak of. And this cycle of spending billions before we get an airframe has been happening since the 1950s.

Why is this important? When the shuttle broke up, it was travelling at 207,000 feet and Mach 18.3. At that speed, more than 12,000 mile an hour, you could circle the entire planet in about 2 hours. Los Angeles to New York would only take minutes. That's how fast they were going. Fast enough that even that high up in the atmosphere, on the edge of space where the air is super thin, the friction of the air moving over the shuttle's wing created a 2,000 degree plasma around the shuttle. The shuttle is moving so fast that the force of the air around it might as well be harder than diamond. In that environment, most metals melt instantly, if they can survive the strange chemical effects of oxygen and ozone molecules bumping into them at extreme speeds. In that environment, the smallest defect will become a catastrophic structural failure. If, say, a landing gear door is not sealed properly, the forces acting on the orbiter will tear it off and start a chain reaction, taking the rest of the wing structure with it before crushing the rest of the orbiter.

The last seconds of the Columbia's flight were a battle. The flight control computers were trying to correct for an increase in drag on the left side of the orbiter (which would pull the nose to the right). The computers tried to correct using the flight control surfaces on the wings and tail of the orbiter, but could not generate the force necessary to combat the increasing drag (most likely increasing as more of the left side orbiter structure disintegrated). As a last resort, the computers even tried to use the Reaction Control System thrusters in the right side nose of the orbiter to compensate (the RCS thrusters are rockets, normally used in space). At that point the computers lost the battle for control of the orbiter and seconds later the vehicle broke up, most likely from aerodynamic stress.

Photos of the orbiter over the southwestern United States taken by the [Starfire Optical Range] telescope were released today by NASA, and at first glance they seem to show something on the left side leading edge of the orbiter coming apart.[Piece of Columbia Wing Found, Photos of Re-Entry Examined][Air Force imagery confirms Columbia wing damaged][NASA Examines Air Force Photos of Shuttle]

Starfire is actually one of the most advanced optical facilities in the world, being part of missile defence research since at least the 1980s. Starfire's telescope use [adaptive optics]- flexible, computer controlled mirrors- to correct for atmospheric distortion as it looks up into the sky (ie it removes the twinkling from the stars). Since the photo of Columbia was taken not of a star in the night sky but an aircraft crossing the morning sky at 18 times the speed of sound, it will take the experts to determine what, exactly, those photos show. If it is damage to the leading edge of the orbiter, all the other pieces of telemetry would make sense, from the loss of temperature and pressure sensors in the landing gear bay to the increase in drag on that side of the orbiter.

Last Saturday, the day of the accident, Chelle asked a very good question.
"Was it preventable?"
While we still don't know all the details of the accident, there are some things that are certain already. The orbiter was lost because of structural failure due to extreme aerodynamic loads. We also know that the flight control system not only knew there was a problem, but fought to correct it for several minutes before it completely lost the struggle and the orbiter broke up. Wether or not this is a software "defect" might be open to debate. Modern fly-by-wire flight control systems such as that on the F-16 Falcon would have been able to maintain control in a subsonic or supersonic flight environment. There are flight control systems that are even designed to adapt to large portions of the aircraft coming off or being inoperable. If we had more data from a dedicated hypersonic flight test program, such as any one of the dozen or so cancelled in the last decade alone, the shuttle's software probably could have handled the changing flight conditions (ie increased drag on the left side) better. The software could easily have been updated to handle a loss of control authority at high speed. It may even have been smart enough to not try to conpensate for the increase in drag, as the induced rolling moment and yawing that the computer's actions generated are probably what caused the sudden structural failure of the orbiter.

It's unfortunate that the [Shuttle Infrared Leeside Temperature Sensing] cameras mounted in Columbia's tail were removed during it's last overhaul. During past flights these cameras have imaged the reentry of Columbia and the effects of heating on the thermal protection system - and specifically the left wing of the orbiter. If this camera had been in place, we would have full imagery of whatever happened to the left wing of Columbia.

2/07/2003 04:41:10 PM ] [  ]


Shit. Browsing through [coderlog] after randomly finding it using my fixed Mozilla/IE [What's Related?] bookmarklet (drag that link to your toolbar to use it), I happened across [CodeCon 2003], which is looking a HELL of a lot more interesting than DefCon was this year. It's in 2 weeks in SF, so I somehow doubt I'll be able to go :(

2/07/2003 04:02:36 PM ] [  ]


Wednesday, February 05, 2003

The author of [Zoe] left me a comment here the other day and I was about to send him a reply (he turned me onto Zoe's XML-RPC interface), but it appears our good friends at Haloscan are acting up, and [I'm not the only one] who can't even dig up records of the post on haloscan's site. Bleh.
Overall, haloscan is good, when it works. At some point when the AP is all set, I'm just going to host the blog from the AP machine (chelle's too, probably) using something that has trackback, etc. etc. Until then, I'll keep on listening to Pantera's [Cemetary Gates] followed by T.A.T.U's [All The Things She Said].
Kissing russian schoolgirls that can sing are pretty cool. But yes, playing that after Pantera makes me seem a tad unbalanced.
Got a problem with that? ;)

2/05/2003 01:19:31 PM ] [  ]


What OS are you?
Which OS are You?
Which OS are You?

That was actually a reasonable quiz, though hey, I really do think I'm more BeOS than OS X. At least I didn't end up Windows ME (yeeeech).
BeOS: Cause I'm a little country, and a little rock n roll. OK, no country.
And speaking of the terror from Redmond, yet again the PC at home would not mount a Windows XP Pro CD last night. Terribly odd, but it's an eMachines, and even under Be it sees a network card, but can't actually provision an IP on it. Again, this is why I use Macs.

2/05/2003 01:06:53 PM ] [  ]


Tuesday, February 04, 2003

Well, after some fiddling around I got the CocoaEmbed project to work in an "offline" mode, ie no network connections. So when viewing email, it won't load an image or whatever from a spammer's server. The key was to call CHBrowserView's stop method with the flag NSStopLoadNetwork immediately after I called the loadURI method. That should prevent anything from loading over the network, but I'd much prefer to have a flag set when the loadURI method is called. Hopefully nothing will slip in or screw up because i'm stopping it right after I tell it to load. Here's the code:
- (IBAction)load:(id)sender
NSString* str = [urlbar stringValue];
[browserView loadURI:str referrer:nil flags:NSLoadFlagsNone];
[browserView stop:NSStopLoadNetwork];

Just drop that in as the load method and you should have it working. So far I haven't been able to break it any more than the debug CHBrowserView is already broken-ish.

2/04/2003 10:11:50 PM ] [  ]


[Pig], [haloblack], and Chemlab are all going to have new releases this year, with haloblack's in a month or so. In the words of one of the board posters,
wtf is this, 1996?

Not that any of us are complaining, mind you. And there is a rumor that 16volt and/or h3llb3nt are now on a major label.
Me, I'm making do with finding outargeously good [Schwein] tracks out there in the noosphere, with the occasional bit of [crackwhore] pop thrown in.

Hey, it beats listening to Kelly Osbourne. Or seeing her.

Speaking of keepin it [real], if you know someone like [this], be a real friend and set up an intervention (picked up from zFilter).
You can make a difference, yo.

2/04/2003 04:38:42 PM ] [  ]


Monday, February 03, 2003

Well, I was waiting to post something about Columbia, but it's going to be a doozy, so that'll wait till tomorrow.
[Zoe] is pretty cool, but it's hard to get it working, and even when it's going it does eat a lot of memory. It does a lot of cool stuff, but it does it through a web browser interface, running a web server on your machine, which ends up being a little clunky.
[Emila] is an attempt to do much the same thing, with a less clunky Cocoa interface. Both of these apps do a lot of the things, and express the ideas that Evil Toaster, my mail app does. Of course, Evil Toaster is still a bunch of bits and pieces without a coherent whole still.

Email has always been the killer app for the internet, but it's becoming harder and harder to use. We're bombarded with information. At home I still have only the most basic spam filtering turned on with the account I've had for quite a while, last night I managed to get 30-something spams in the space of an hour. Of the 100 or so legitimate emails I get a day, maybe 90 of them are mailing list related - skunk-works, cocoa-dev, mac-games-dev, bsd-wireless, etc. Filters put those into the right places so that everything is tidy. At work, on the otherhand, I get cc'd on everything, people send Word, etc. attachments, various automated systems send me emails about things I don't care about, and all of it is impossible to filter. For work I've been using Mozilla's Mail and News component, which does a pretty good job at what I need it to do- fast searching.

But all email apps are still pretty much the same. Zoe and Emila are cool if only because they don't look and feel like Outlook. The things that I find important or useful in an email app are very different from what other people look for, but there's no reason that you can't build a back end of mail protocol and storage abstraction, indexing for searches, filtering, and all that other good stuff, and choose between the front end for the masses or the Batmobile of email.

Which, of course, is what I'm doing. Evil Toaster is now Evil Toaster (the GUI) and Mailroom (the backend daemon).

2/03/2003 11:21:44 PM ] [  ]