TTG Home » Robert Home » Daynotes Journal Home » Journal for Week of 16 July 2001
|
Week of 16 July 2001 Latest
Update: Friday, 05 July 2002 09:16 |
|
Monday, 16 July 2001 [Last
Week] [Monday] [Tuesday] [Wednesday]
[Thursday] [Friday] [Saturday]
[Sunday] [Next Week] |
|||||||||||||||||
A math error! A public math error! Arrghhh!
That's what happens when I'm running on minus sleep. Yesterday, I said,
"In theory, the 16" has nearly twice the light gathering
power (16^2 versus 10^2)..." I squared 10 properly to get 100,
but when I squared 16, I somehow came up with 192 rather than 256. Rip my
epaulettes from my shoulders. An error in binary, yet. A double bit flip.
I obviously need an ECC brain. Thanks to everyone who pointed out the
error to me. I expected to find public messages on the messageboard, but
no one posted one, perhaps from consideration for my embarrassment. Here's
a representative email I got, with my response:
Thanks. Of course, the 16" has 2.56 times the light gathering power of the 10". You don't need the other values; simply squaring the diameters works fine. 16^2 is 256 rather than 192 and 10^2 is 100. That's what happens when I'm running on little sleep. Light gathering ability has nothing to do with the exit pupil size, except that light is wasted (giving a lower effective objective size) if the exit pupil is greater than the maximum size of the observer's pupil. The diameter of the exit pupil is (roughly) the focal length of the eyepiece divided by the focal ratio of the scope. For example, a 30mm eyepiece used in my f/5 scope is 6mm. If one selects an eyepiece focal length that results in an exit pupil larger than "fits" into the eye of the observer, e.g. a 40mm eyepiece with an f/5 scope, then as you say the effective light gathering ability of the scope is smaller. There are other factors, as you mention, such as central obstruction and reflectivity of coatings, but I was assuming those to be equal. Relatively large central obstructions result in lower contrast, but have relatively small effect on light gathering. For example, a scope with a tiny central obstruction might have an 18% CO (linearly), whereas one with a huge CO might have 35%, again linearly. If you calculate the surface areas of those obstructions, they block relatively small amounts of light (about 3.2% in the case of the 18% CO and about 12.3% in the case of the 35% CO). Although they block little light, they do contribute to reduced contrast via diffraction. More important is the coating efficiency, which might be only 86% on some scopes to as high as 97% on others. Because the light bounces off the primary to the secondary to the eyepiece, an 86% reflectivity on both primary and secondary translates to a total reflectivity of only about 74%, and a 97% reflectivity on both to total reflectivity of about 94%. Worst case, such as on SCTs with low-efficiency coatings and large COs, you end up with perhaps 2/3 of theoretical transmission based on the mirror size.
I hope my five pounds of tobacco will show up today. With luck, I'll not need to open that emergency box of Prince Albert I bought Saturday night, although perhaps I've exaggerated in my mind just how bad it's likely to be. Oh, well, I'd best get this published and get back to work on the book. Short shrift around here until it's done.
Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
Tuesday, 17 July 2001 [Last
Week] [Monday] [Tuesday] [Wednesday]
[Thursday] [Friday] [Saturday]
[Sunday] [Next Week] |
|||||||||||||||||
| Oh, no! My tobacco didn't arrive yesterday, so
I called Cornell & Diehl. Craig Tarler had promised me that it would
ship last Wednesday and be here by Friday, but he didn't realize that he'd
run out of Latakia. So yesterday I smoked my last 968. With great
trepidation, I opened the Prince Albert box and gave it a sniff. It was
worse than I'd feared. Barbara sniffed it, and said it didn't smell too
bad to her. I found myself wondering what she was comparing it to. Not too
bad, perhaps, if compared to a compost heap, but pretty bad when compared
to any tobacco I'd choose to smoke.
I packed my pipe with the Prince Albert, which could certainly be considered cruel and unusual punishment to the pipe. I doubt that many Dunhill pipes have ever suffered the indignity of being packed with Prince Albert. I thought I heard a tiny muffled scream from the pipe as I filled it, but perhaps that was my imagination. I lit the pipe and found that it was as bad as I expected, perhaps worse. Somewhere on the scale between abominable and hideous. I think they use ethylene glycol (antifreeze) or something similar as a moistening agent. Seriously. I'm not kidding about that. As bad as the taste is, the aftertaste is worse. Price Albert is to real pipe tobacco as ersatz coffee is to real coffee. I think I'm going to have to boil this Dunhill pipe out with alcohol to get rid of whatever junk they put in Prince Albert tobacco.
I have several chapters in progress, and hope to send a total of three or four off to my editor by the end of this week. I'll post them on the Subscriber page when they're complete. Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
Wednesday, 18 July 2001 [Last
Week] [Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday] [Sunday] [Next
Week] |
|||||||||||||||||
| The king is dead. Long live the king. Intel
announced that the Pentium III will be no more as of the end of this year.
Intel is transitioning from the 815 chipset and Pentium III to the 845
chipset and the Pentium 4. And not a moment too soon. Clock speed sells,
and the Coppermine core upon which the Pentium III is based has a top end
of 1 GHz, more or less. AMD has been embarrassing Intel by shipping
Athlons that run at speeds much higher than those achievable by the
Pentium III. The Pentium 4 brings an end to that, because Pentium 4 clock
speeds are faster than anything AMD can achieve with the Athlon.
Of course, clock speed doesn't equate to performance, particularly when comparing processors as different as the Athlon and Pentium 4. The Athlon and Pentium III were reasonably well matched, clock for clock, but the Pentium 4 must run at significantly higher clock speeds to match the overall performance of an Athlon. That doesn't really matter, though, because it isn't performance that sells, it's clock speed. System buyers perceive a 1.5 GHz Pentium 4 as being "faster" than a 1.3 GHz Athlon, even though the converse may actually be true. Volume shipments of the Pentium 4 are going to hurt AMD badly. The Pentium III isn't really dead, of course. It lives on in mobile form and, more importantly, as the Celeron. The new 0.13µ Pentium III core, code-named Tualatin, will be the basis of the next generation of Celeron, which I will for convenience call the Celeron 4. I've heard various estimates, some from people who should know, about how fast the Tualatin core will eventually be able to run. The initial cores are shipping at not much more than 1 GHz, but the die shrink should allow core speeds eventually to reach 2 GHz or more. Of course, Intel will manage release schedules to avoid channel conflict between the Pentium 4 and faster Celeron 4 models, but the point is that Intel now has a great deal of headroom in both product lines when it comes to clock speeds, neatly reversing the previous situation when Intel was playing catch-up with AMD on clock speeds. AMD has nothing to match a 2 GHz Celeron 4, either on absolute clock speed or for that matter on actual performance. Unless AMD does something fast, they're going to be in the same old situation of playing catch-up with Intel, being forced to sell their processors as inexpensive, slower alternatives to Intel processors. Although Intel now holds most of the trumps, they still have work to do. They need a trouble-free roll-out of the 845 chipset, for starters. They also need to provide a DDR solution for the Pentium 4. The 845 is actually DDR-capable, but it will ship as an SDR-only platform, supporting standard PC133 SDR-SDRAM. That's fine as far as it goes, but the Pentium 4 really wants more bandwidth than SDR-SDRAM can provide. Intel's entanglements with Rambus have influenced their product lines too long already. Intel badly needs a Pentium 4 platform that uses cheap, fast DDR-SDRAM. There's no technical reason why they couldn't provide that now, and I expect to see an Intel DDR Pentium 4 chipset by year end or slightly thereafter.
I got a royalty statement from O'Reilly yesterday, covering Q1/2001. For the first time ever, there was a cover letter with it. The letter basically said that everyone's royalties were much lower than usual because of the blood-bath in computer books. Distributors aren't buying many computer books because bookstores aren't buying many computer books because people aren't buying many computer books. Like all computer books, PC Hardware in a Nutshell did poorly, selling fewer than 1,500 copies a month during the first quarter. The second quarter isn't likely to be any better, although O'Reilly did say that sales started to pick up some late in the second quarter. Of course, everything is relative, and a lot of computer book authors would love to see their titles selling at an annualized rate of 15,000 copies or more even during good times. Ah, well. Perhaps the third and fourth quarters will be good.
My tobacco finally arrived yesterday via UPS. I spent the afternoon surrounded by a haze of good tobacco smoke. The Dunhill pipe that I'd used with the cheap tobacco is currently undergoing decontamination. I filled the bowl with salt to absorb the moisture, and will later soak it in rubbing alcohol to remove any remaining contamination. I also made the mistake of filling my pouch with the junk tobacco, and now it smells of it. I'm temporarily using a plastic baggie as a pouch, and airing out my real pouch on the kitchen table. I'm afraid that it may be unsalvageable, though. Oh, well. I guess I'll have to put in an order for someone to club yet another baby seal so that I can get a new tobacco pouch.
I finally got around to ordering another Telrad finder for the 90mm refractor. It'll probably look like an elephant sitting on a chipmunk, but it'll let us find stuff, which the bundled finder does not. We decided to use both the standard 8X50 finder and the Telrad on our 10" scope, but I'll probably just remove the 6X30 finder from the refractor. Optically, the 6X30 is fine, but both the mounting position and its inability to hold alignment make it pretty useless. As with the 10" scope, the Telrad will get us into the general vicinity of what we're looking for, after which a low-power/wide-field eyepiece will allow the telescope to serve as its own finder. Barbara is off with her parents and sister today to visit a pottery. She'll be back in time for us to have dinner and head off to the Forsyth Astronomical Society meeting at SciWorks tonight. Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
Thursday, 19 July 2001 [Last
Week] [Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday] [Sunday] [Next
Week] |
|||||||||||||||||
| I mentioned this a while ago, but if you're
considering buying a CD writer, hold off until next week. There should be
some interesting news Monday. I suspect the price of the current
generation 16X CD writers will begin falling next week. That's very good
news, because it will put the price of 12X and 16X writers within the
budgets of a lot more people.
The truth is that faster CD writers are a matter of diminishing returns. There's always a minute or so of overhead that's required to write the TOC and so on, so increased X numbers have limited benefits once you get above 12X or so. A 12X writer records a full 74-minute CD in a bit over six minutes, plus a minute or so for overhead. Call it seven minutes total. A 16X writer reduces that to a bit under six minutes total. Unless you're sitting there timing the write, you're not likely to notice much difference between 12X and 16X writes, and the same will be true when faster writers ship. That's especially true with BURN-Proof writers, which allow you to continue doing other things while the CD is being written. We now regard a 12X or 16X writer as standard equipment in any PC we build other than absolute entry-level systems, and recommend that you do the same. The things just have value all out of proportion to their minimal cost. Well, I'd best get to work on the book... Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
Friday, 20 July 2001 [Last
Week] [Monday] [Tuesday]
[Wednesday] [Thursday]
[Friday] [Saturday] [Sunday]
[Next Week] |
|||||||||||||||||
|
On this day in history, Diana Rigg and Natalie Wood were born (1938), and Oberst Claus Schenk Graf von Stauffenberg plants a bomb under Hitler's conference table at the Führerhauptquartier of Wolfschanze near Rastenburg, nearly succeeding in killing the son of a bitch (1944). And, of course, on this day in 1969, Neil Armstrong stepped from Apollo 11's Lunar Lander onto Luna's surface and uttered the most famous blown line in history, "That's one small step for man, one giant leap for mankind." And that same day was personally significant for me for an entirely different reason.
I've gotten some unbelievable spam, but this one tops them all. I reproduce it here unedited, via a copy/paste from the original HTML message.
You are receiving this email because you requested to receive info and updates via email. To unsubscribe, reply to this email with "unsubscribe" in the subject or simply click on the following link: Unsubscribe 
Would you trust someone who marketed his services via spam to do eye surgery on you? I certainly wouldn't. And why would I travel to Ohio to have laser eye surgery done when I could instead travel across the border into Canada and have it done for something like a tenth the price?
I've finally given up on the weather-liars for predicting which evenings will or will not be good for observing. I've settled on a new method. I now use the Magic 8-Ball that Barbara gave me for Christmas. Yesterday evening, I asked the Magic 8-Ball whether tonight would be a good observing night. Its response was "Don't count on it". So I asked it if Saturday night would be a good night for observing, to which it replied with a simple "Yes." As I remember, the Magic 8-Ball responses are two-thirds positive and one-third negative. Because I probably won't solicit its advice unless I think there's a pretty good chance the weather will be suitable, that means its response set probably corresponds pretty closely to the actual likelihood of good weather. In that respect, it will probably give more accurate forecasts than do the weather-liars. I remember a class in Operations Research in which we discussed the accuracy of weather forecasts. The upshot was that the most accurate means of forecasting tomorrow's weather is to predict that it will be the same as today's. If you do that, you're right much more often than the weather-liars are. That was nearly twenty years ago, but I suspect things haven't changed much. Oh, well. The Magic 8-Ball didn't say that tonight would be bad weather, just that we shouldn't count on good weather. So we may be up at Bullington, depending on what the weather looks like this evening. The Telrad for the small scope arrived yesterday, so we may install it tonight. I have lots to write and machines to build, so I'd best get to it. Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
Saturday, 21 July 2001 [Last
Week] [Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday] [Sunday] [Next
Week] |
|||||||||||||||||
| The weather-liars strike again. Barbara watches
the NBC affiliate, WXII, for news and weather. Miss Perky (AKA Michelle
Kennedy, but I didn't know that when I christened her Miss Perky) forecast
clear skies for last night. She even showed a cloud map, allegedly from a
satellite, that showed the clouds all clearing off. Even the Magic 8-Ball
claimed we'd have clear skies. They both lied.
We left here about 8:30, even though there was about 7/10 cloud cover, on the assumption that the clouds on the satellite picture appeared to be moving south and so it should be clearer at Bullington and clearer still as the evening progressed. The nearer we got to Bullington, the cloudier it got. We arrived at 9:00 and sat in the truck for 45 minutes or so, watching the clouds. Barbara did spot Mars, although it was usually behind enough haze that it required averted vision to see. Pretty bad for an object whose visual magnitude is nearly -2. I'd estimate that its apparent magnitude last night was about 3.5 at best, sometimes dropping to less than 5. That means the haze was blocking literally more than 99% of the light. At one point, it did seem as though the clouds were clearing in the north, so we got out one scope and set it up. I wasn't optimistic, so we set up only the scope rather than all the usual equipment. My pessimism turned out to be justified, as the clouds soon moved in. We finally packed up at 10:30 and headed home. Oh, well. It's supposed to be clear tonight, if one has any faith in the weather-liars. I do like Miss Perky, though, despite the fact that she lied to us. She's definitely Weather Channel quality. She's young, very pretty, blonde, enthusiastic, and likeable. I don't expect her to be at WXII for long. I expect her to show up on the Weather Channel in the near future. What's interesting is that she's a new hire at WXII. They already have two "meteorologists", Austin Caviness and Jeff Hardin. Those guys have completed their two-week course (or whatever) that allows them to call themselves meteorologists. That's a joke, of course. A real meteorologist is someone who has an MS or PhD in meteorology, but then we live in a world where garbage men call themselves sanitation engineers, so I'm not surprised that the title of meteorologist is now being used by people with no legitimate claim to it. If it weren't for the AMA, we'd have people who'd taken a two-week course in first aid calling themselves physicians. But at least Mr. Caviness and Mr. Hardin have their AMS certificates, which Miss Perky does not. They must both be very aggravated that Miss Perky is now the lead weather anchor for WXII. She does the primetime news on weekdays, leaving them with only the scraps (early morning, daytime, weekends and so forth).
Heh. Here's a picture you may or may not see in the next revision of PC Hardware in a Nutshell. It's the "before" picture of the keyboard on my main system, after nearly a year of not being cleaned. I normally clean keyboards much more often than that, but I wanted a good illustration. Now all I have to do is run it through the dishwasher and take an "after" picture.
Barbara is off to the Border Collie trials today, so I'm babysitting Mom and the kids. She said on her way out the door that she'd be back in time to head up to Bullington tonight if the weather cooperates. I'm debating what to do today: read, write, or build a system. I'm down to only five systems around my desk now. Under or next to my desk are meepmeep, the Roadrunner box, and thoth, my main workstation. On and under my credenza are theodore, our main file server; orion, the Win98SE box I built for the scanner; and hathor, a Windows 2000 Pro workstation. I think I'll take some time off today and just relax by reading a book or two. Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
|
[Last
Week] [Monday] [Tuesday]
[Wednesday] [Thursday] [Friday]
[Saturday] [Sunday] [Next
Week] |
|||||||||||||||||
| We finally had a decent night for observing
last night. Barbara got home from her Border Collie trial around 18:45 and
picked up KFC chicken on her way in. I was going to say "Kentucky
Fried Chicken", but apparently the state of Kentucky trademarked the
name Kentucky and insisted on KFC paying a royalty, whence the change in
name. It seems bizarre to me that anyone could trademark
"Kentucky" in a global sense, or indeed that a government should
be permitted to trademark, copyright, or otherwise protect anything. And
even if they could, it seems that the scope of that trademark should apply
only to naming geo-political subdivisions. I mean, if Rhode Island decided
to change its name to Kentucky, well then the original Kentucky might have
a valid complaint. But certainly not if a fast-food franchise decides to
use the name, particularly because it was using that name well before
Kentucky trademarked it.
At any rate, we got up to Bullington around sunset, and the weather appeared generally good. The cloud cover was less than 1/10, although the humidity was noticeably higher than forecast. Bonnie Richardson had arrived just before us and was getting set up. Priscilla Ivester arrived shortly after we did. We all got all our scopes set up and then sat around and waited for them to equilibrate. I mounted the Telrad on the small refractor while we waited. By 22:00 it was dark enough that we could begin to make out a few bright objects and by 22:30 it was fully dark, or as dark as it ever gets at Bullington. Unfortunately, that's not very dark. At the darkest, it's still possible to read newspaper headlines. We spent the next couple hours bagging various objects, including M51, several Messier objects in Saggitarius, M57 (the Ring Nebula), Epsilon Lyrae (the Double Double), and so on. By midnight, Saggitarius had begun to climb out of the muck and the light dome on the southern horizon. Unfortunately, just about then clouds started to move in and the high humidity started causing some serious dewing problems. The Telrads fogged up, the eyepieces fogged up, our charts and books started getting soggy, and the telescope tubes had beaded moisture on them. It was pretty obvious that observing was at an end, so all of us packed up and headed out around 00:30. That was a shame. I was hoping we'd be up there until 02:00 or 03:00. Still, not bad for a summer night in the Carolinas. I really can't wait until the fall. Not only are more interesting things up (Jupiter, Saturn, Orion, etc.) but the observing conditions will be much better. Also, the earlier sunset times mean we'll be able to start observing in the early evening rather than having to wait until 22:00 or later for darkness. Barbara and I have talked about it, and this fall we'll probably start heading up to the club observing sites on the Blue Ridge Parkway. That involves a round-trip commute of roughly three hours, but I think it'll be worth it. The Parkway gains us dark skies (or at least as dark as they get around here) and enough altitude to get us above most of the haze and muck. We'll have some organization to do to make sure Mom and the dogs are taken care of while we're gone, but that should be doable.
I got another one of those stupid chain-letter/Ponzi scheme spams yesterday. Several copies, actually, including one that had header information forged to make it appear it had come from me to me. Geez. At any rate, I thought the mathematics was amusing. You can send this moron $20 for his "reports". He suggests that you start by sending 5,000 spams, and says you can expect a 0.2% response rate, or 10 responses, each of which sends you $5, for a total of $50. According to him, it then pyramids, with the 10 responders each sending 5,000 spams, from which you receive 10*10 or 100 responses, each of whom send you $5, for a subtotal of $500, and a grand total of $550. Each of those 100 responders then sends 5,000 spams, with a 0.2% response rate, for 10*10*10 or 1,000 responses, each of whom sends you $5, for a subtotal of $5,000, and a grand total of $5,550. Each of those 1,000 responders then sends 5,000 spams, from which receive 10*10*10*10 or 10,000 responses, for a subtotal of $50,000 and a grand total of $55,550. Let's see how it really works. As I established last week, the actual response rate for a spam message is more likely to be on the close order of 0.005% than 0.2%. So, when you send out your first 5,000 spams, you can expect 0.25 responses. That partial responder sends you 0.25 * $5, or $1.25. He then sends out his 5,000 spams, but since he's only a quarter of a responder, he actually sends only 1,250 spams. He gets a response rate of 0.005%, or 0.0625 responses. That 1/16th of a responder sends you $5, for a subtotal of $0.3125 and a grand total of $1.5625. The 1/16th of a responder sends out 5,000 spams, which actually totals only 312.5 spams. He gets a response rate of 0.005%, or 0.015625 responses. That 1/64th of a responder sends you $5, for a subtotal of $0.078125 and a grand total of $1.640625. The 1/64th of a responder sends out 5,000 spams, which actually totals only 78.125 spams. He gets a response rate of 0.005%, or 0.00390625 responses. That 1/256th of a responder sends you $5, for a subtotal of $0.01953125 and a grand total of $1.66015625. So there you have it. For only $20 and quite a bit of work, you can earn $1.66, for an actual loss of only a little more than $18. But at least the morons who go for this plan are getting a cheap education. Of course, the original spam mentions in passing that one must get the required 10 responses, even if that requires mailing out more than the planned 5,000 spams. Let's see how that might actually work. With a 0.005% response rate, getting 10 responses would require sending 200,000 spams. In order to get their required 10 responses, your first ten responders would also have to send 200,000 spams each, or 2,000,000 spams. Their 10*10 responders would also have to send out 200,000 spams each, or 20,000,000 spams. And their 10*10*10 responders would also have to send out 200,000 spams each, or 200,000,000 spams. And their 10*10*10*10 responders would also have to send 200,000 spams each, or 2,000,000,000 spams. So, cumulatively sending 2,222,200,000 spams might theoretically result in the originator getting his $55,550. Of course, in practice, there are a couple of problems with this whole scheme. First, it assumes unique email addresses, and the chances of all 2,222,200,000 spams being addressed to unique addresses approaches zero so closely as to make no difference. Second, as with any Ponzi Scheme, the originator and early adopters are the only ones to benefit. Anyone who joins the scheme later (in this case, any time after the originator) has no chance of getting out even as much as he paid in. Third, of course, the likely response rate to such a scheme is nowhere near 0.005%. That was the response rate I estimated for someone who was actually selling something that had some value, albeit likely a small value. For the chain-letter spam, the response rate might be two to three orders of magnitude smaller, which means it has no chance of succeeding until we have a few more email users out there. We should have enough for it to work, but only once we've colonized this entire galaxy and have begun to establish colonies in other galaxies. Of course, there'll then be a problem with speed-of-light and response time. So, assuming that the response rate is actually 0.00005% and this moron sent 10,000,000 spams, that means he's likely to get five responses and a total of $25 for his efforts. That in turn means that this particular idiot will never try it again. Unfortunately for all of us, there's a continuing supply of idiots. Did I get the numbers right? I don't know. I didn't bother checking, because once one begins dealing with such large/small numbers, moving the decimal one place to the right or left doesn't make a lot of difference.
Click here to read or post responses to this week's journal entries Click here to read or post responses to the Linux Chronicles Forum [Top] |
|||||||||||||||||
