Below are the 20 most recent journal entries recorded in David Malone's LiveJournal:

[ << Previous 20 ]

Monday, November 2nd, 2009
8:05 pm	1792 A while ago, I noticed a number of sources that said in 1792, there was a move from the use of apparent solar time to mean solar time. For example this page mentions it. There were no details about who made this decision or why. After a bit of digging, I managed to find that one page was using a pamphlet by David Fisher called "Time Changes" as a source. It turns out David Fisher is an astrologer in the UK, and astrologers are quite interested in calculating historical dates. The TCD library had a copy of the astrological periodical in question, so I ordered a copy from stacks. After some confusion, I got the right issue, but the pamphlet was missing. I managed to contact David Fisher, who kindly went looking for the article, but had no luck. Later, I managed to get a copy of Derek Howse's Greenwich Time and Longitude (Derek Howse was a UK astronomer). I noticed the same date (on p 88 of my copy): Communities began to keep mean time in preference to apparent - Geneva from 1780, England 1792, Berlin 1810, Paris 1816. Unfortunately, he gives no citation or further information. He gives a much better discussion of some of the other events later in the 1800s. Derek Howse is no longer with us, so I couldn't really have followed it up. Recently, I noticed that the Oxford Companion to the Year has the same information (p.664) As clocks and watches improved in accuracy and declined in price, cities and nations began to adopt mean solar time for purposes of legal definition, Great Briton doing so in 1792. The authors are still alive, but can't remember the exact sources. Patrick Wayman's book on Dunsink Observatory mentions some things in passing. The footnotes on page 152 mention that the equation of time is included in Watson's almanac from 1792 and continues to be shown by Thome's Directory. It comments that mean time is the "time shown shown by well regulated clocks and watches" in the 1954 Thome's. Later, in the footnotes on page 154, there is a comment that in the 17th and 18th C, whenever accurate time was required, people would use a mean time clock, but there was some doubt if "true time" would be better. I spent some time looking at old editions of the Nautical Almanacs. The first one, for the year 1767, has an explanatory section beginning on page 145. On the first page it comments This Time [Apparent Time] is different from that shewn by Clocks and Watches well regulated at Land, which is called equated or mean Time. Later in the explanatory section, it actually mentions But if Watches made upon Mr. John Harrison's or other equivalent Principles should be brought into Use at Sea, ... which is a nice touch. This section basically remains unchanged - the text in the 1794 Almanac (which was printed in 1793, and so would have been produced immediately after the mysterious 1792 change) is basically the same as that in the original 1767 edition. Indeed, apart from some modernisation of the English and an attribution to (the late) Dr Maskelyne, this text is intact until 1833. In 1834, there is a big change, and the almanac moves to mean time. An impressive committee of the Astronomical Society of London (including Airy, Babbage, Beaufort, Bishop of Cloyne (Brinkley), De Morgan, Hamilton and Herschel) decided that it was silly to list information in the Almanac in apparent time, because everyone has to convert it to mean time before reading it. They add a note to the pages saying "MEAN TIME" to make sure no one missed it, and the explanatory section on p. 491 now says All the articles of the Ephemeris have been computed for Greenwich MEAN Solar time; ... So, if there was any change in 1792, it passed the astronomers and navigators by without making much of an impact. My most recent lead came from Google Books, which turned up a reference in "Murphy's bed: a study of real sources and sur-real associations in Samuel Beckett", which in passing talks about the 1792 change in London and says that page 770 of the 1935 Nautical Almanac describes what is going on. So, back to the library. Indeed, on page 770 it says: Eventually states imposed local mean time, so that everyone in a community would use the same time. In 1780 Geneva was the first to do this, then England in 1792, Berlin in 1810 and Paris in 1816. However, still no mention of what or how, but at least this looks like the source that Howse and the Oxford Companion to the Year are quoting, and probably David Fisher too. The Irish Merlin (an Almanac) from 1793 doesn't seem to say anything either: It's enough to make me wish for the simplicity of leap seconds. (5 Comments |Comment on this)
Thursday, October 29th, 2009
9:20 pm	Twitter We were doing some simple software radio as part of a summer project, and discovered some shocking long wave interference. It made it impossible to decode BBC 4 on longwave (which wanted to do), because of a chiriping sound, with about 2 chirps a second, plus some gaps that seemed to show some structure (there's an MP3 of what it sounds like). The noise seems to be really wide band, based on a spectrogram we generated with the software radio. Someone had heard the same chirping elsewhere on campus, and after wandering around with a hand held radio, we discovered that the noise seemed to be coming off the mains cables. Some nice people from Comreg have been out to help us find the problem, and it looks like it might be the fire alarm. I checked some of the fire stuff in our building on the way home and it looks some of it is emitting so much RF, that it is confusing my hand held radio's squelch. It looks like an "Analogue Addressable Fire System". I'll have to find out if I can see how the signalling is supposed to work. (1 Comment |Comment on this)
Saturday, October 24th, 2009
9:48 pm	MeTube I'm in a video on youtube of when I did a talk at the Ignite Dublin event. Mad - I don't think I've been on youtube before. I need to learn to speak into the mic. (2 Comments |Comment on this)
9:41 pm	Shiny Cable TV? When we moved in here, NTL had disconnected the previous resident by the time we moved in, even though the previous resident had had to pay for the service for that month. You could see the cable dangling where it would previously have been shoved into one of those junction boxes. This is a few years ago. About a week ago, we noticed it had been reconnected. Yet the next day it had been disconnected again, and the next day again it was connected and then disconnected again. We were wondering what was going on. In jest, I suggested it might be magpies, who often clean our gutters for us. This morning, I was peering out the window and saw the local magpies pulling and playing with various cables on the front of the houses on the other side of the street, including phone, TV and power cables. I guess they could well actually shove a F connector in hard enough to get it to stay in... (Comment on this)
Monday, October 5th, 2009
9:21 pm	Pay gaps There's been talk here about the public/private pay gap here, and people are suggesting that the public sector should take a pay cut, to put them in line with the private sector. There are some CSO statistics on it here. The statistics suggest that to get public and private sectors in line: Women in the public sector should take a bigger relative cut than men. The lower paid in the public sector should take bigger relative cuts. It also looks like private sector workers in big companies should take a cut to put them in line with smaller companies. There's also a hint that everyone should join a union. (Though, remember correlation is not the same as causation...) (Comment on this)
Monday, September 28th, 2009
7:15 pm	Nice to meet you I love the radio ad about the man with the one eyebrow. (Comment on this)
7:26 am	Loop I built myself a loop antenna, which is supposed to be directional. Interestingly, if you but your volt meter between the earth and one end of the loop, it doesn't seem to be. However, if you put it between the two ends of the loop it is. I produced an animation (34MB) showing the spectrum as I turn the loop through 360 degrees in 60s. If you look, you can see the stations at 180kHz and 60kHz fading in and out. The stations at 198 and 200 don't seem to show it as clearly, which may just be an issue with the colours for the spectrogram. I wonder if the ticks around 100kHz are actually the LORAN signal? It looks like they fade in and out too. (2 Comments |Comment on this)
Tuesday, September 22nd, 2009
7:31 pm	The phone network sucks... Today I saw someone get two calls from +33-00-353-their-Irish-phone-number. It turned out to be a recruitment company cold calling. I can only presume that they have paid someone to be able to set their source phone number. At least the Internet doesn't pretend that address spoofing isn't possible. (Comment on this)
Wednesday, September 9th, 2009
8:25 pm	All address space... ... is not created equal. (Or, at least, is no longer equal.) (Comment on this)
8:23 pm	All the nines % date +%y 09 7% date +%y/%m 09/09 8% date +%y/%m/%d 09/09/09 9% date "+%y/%m/%d %h Unmatched ". 10% date "+%y/%m/%d %h" 09/09/09 Sep 11% date "+%y/%m/%d %H" 09/09/09 09 12% date "+%y/%m/%d %H:%M" 09/09/09 09:08 13% date "+%y/%m/%d %H:%M:%S" 09/09/09 09:08:13 14% date "+%y/%m/%d %H:%M:%S" 09/09/09 09:09:04 14% date "+%y/%m/%d %H:%M:%S" 09/09/09 09:09:06 14% date "+%y/%m/%d %H:%M:%S" 09/09/09 09:09:07 14% date "+%y/%m/%d %H:%M:%S" 09/09/09 09:09:09 (Comment on this)
Thursday, September 3rd, 2009
9:11 pm	Don't throw out old junk A paper about how some changes to junk DNA looks likely to have produced new genes in humans. (Comment on this)
8:13 pm	Apple Server Tips Apple have a Server Tips video at the moment, where they show how stuff like Terminal, ssh and iostat work. Interestingly, the first ssh connection they make is actually to an IPv6 link-local address. There's also some good tips for command line versions of some of the Apple graphic utilities too. (Comment on this)
Saturday, August 22nd, 2009
8:27 am	Cache speed I recently watched an MIT lecture on cache and algorithms. I was surprised by one comment, that the bandwidth of different parts of the memory hierarchy was similar, but the latency was different. Maybe I misunderstood him, but that isn't really how I remembered it. When we used to assess CPUs, one of the tests we used to do was to see how fast Matrix dot products would run on the CPU, because some of our users did a lot dot products. We used to plot graphs of how fast the dot product went in a warm cache. I just reran those tests for different vintage CPUs - the x-axis shows how much space the data took up and the y-axis shows how quickly we got through the data (one flop is a multiply plus add). The code is just straight C compiled with gcc -O3 with no unrolled loops or other fancy stuff. Each point is about a billion multiplies. On the left the data set is very small, the various overheads are big enough that they aren't being amortised over enough real operations, so performance seems to increase. On the right, your data is too big for the cache, so you you're seeing main memory performance. In between, you're seeing the performance of the L1 or L2 cache (for some of the CPUs the steps in performance are quite clear). For all the CPUs, there's a pretty big difference between cache performance and main memory performance (though both CPUs and ram have obviously got faster over the last few years). If you look at the log scale graph, it looks like this gap has been narrowing over time - from roughly 10x to 5x. However, that still seems like a pretty substantial difference. I guess the latency gap is probably even bigger. I might be able to see that by looking at random accesses rather than sequential accesses. (Comment on this)
8:22 am	LJ Interesting - there was a paper about Livejournal in the Proceedings of the National Academy of Science in 2005. (Comment on this)
Thursday, August 13th, 2009
8:14 pm	Twitter Bots Interesting - It seems twitter is being used as a command channel for bots. (Comment on this)
Tuesday, August 11th, 2009
8:55 pm	Only Connect One of the only connect sequences that you had to spot the connection was 70!, tan 90, √-1, 1/0. I'm surprised that neither team got it, as one team was called "the mathematicians". We all knew about 69! in school... (Comment on this)
Monday, August 3rd, 2009
6:48 pm	*-cert Maths I've been working my way through the leaving cert and junior cert maths exams (available via the examinations web site, choose "Examination Material Archive"). It's a while since I've looked at these exams, as I haven't been giving many grinds since finishing being an undergrad. People do the leaving cert when they are about 18 and the junior cert when they are 15-16. The exam is the same for everyone in the country, though you can choose to do a harder or easier course. Some people have claimed that the standard associated with these exams has fallen over time. My impression (which would largely be based on the LCH papers) is actually that the questions are of a similar standard to those from 15-20 years ago. The "part c" questions are often quite tricky, and questions often draw on material from several parts of the course. Its possible that the marking is easier, but anyone who can do most of the questions should have a pretty solid base in maths. (The optional parts of the course have changed since my day. They now know how to integrate to get the volume of a cone, do some coordinate geometry of the ellipse and know a little about how to solve recurrence equations). (Comment on this)
Sunday, August 2nd, 2009
9:26 pm	Interesting I located a mortgage statement from the 80s. Check out the interest rates in the third column. (2 Comments |Comment on this)
Friday, July 31st, 2009
10:03 pm	Central Bank I was trying to figure out how many people the Irish Central Bank employs. I checked their web site, and their org chart suggests they have at least 33 managers. I thought I'd check the McCarthy report, because it lists numbers of people in various parts of the public sector. The only mentions of the central bank that I could find was one quoted statistic and mentioning that one of the committee was an ex-head of the central bank. (Comment on this)
9:52 pm	Predictive Order I don't understand how my phone orders predictive text homographs. Lust, Just, Kurt. Shot, Riot, Pint. Good, Home, Gone, Hood, Hoof, Hone, Goof. Done, Food, Fond, Dome. Deep, Fees, Deer, Feds, (Comment on this)