Does anyone have a basic program for calculating elapsed days between two given dates? I would like to run this program on a 15C, 33S or any Pioneer series model, and a 48-series.
Data arithmetic / elapsed days calculation program
|
|
« Next Oldest | Next Newest »
|
▼
09-28-2005, 03:32 PM
▼
09-28-2005, 03:47 PM
That seems like a "data-intensive" program. Can you make a good alogrythm? Try july31 1947 to april 6, 2004. I guess I would write a supbroutine that 1st figures out the "day number". So you do 31 <enter> 4 <enter> 1947 GSB(x) and calculates days through march---how to store that----a series of if/then could build essentially a "lookup table" so that with "4" it returns 90 days--what about the leap years--a nested if/then for that? Then you add the days to the month "lookup table reuslt" and that is stored in a register, as is the start year. Then you run the same thing for the second date. Then the years subtract--oh, we need a leap year figurer for that part, too. I think this is too much memory for hte 15c, but it would fit in the 33s and the 48g. I am sure there is a program for (on the 48 and 49) this available at hpcalc.org. Regards ▼
09-29-2005, 04:18 AM
Hi, Bill: Bill wrote:
"I think this is too much memory for hte 15c, but it would fit in the 33s and the 48g." You're joking, right ? The HP-15C can do this sort of calculation using 1/10th of its programming resources, at most. As for the 33s and its 32 Kb of RAM, try and invert an 8x8 matrix using it. The 0.4 Kb HP-15C can do that directly from the keyboard. Best regards from V. ▼
09-29-2005, 08:52 AM
Hi Valentin, Glad I got your attention; I have not conversed with you in some time :-) Well, seeing that some here are more clever than me, yes indeed it looks like you can easily fit the problem into a 15c. I am afraid that my stream-of-consciousness idea of a solution would never fit into the 15c. Yet another example of how much more effective brains (15c, excellent built-in resources, efficient memory allocation) wins over brawn (33s, gobs of memory, difficulty using most of it) any day. On this day, I fit into the brawn category :-( Regards, Bill ▼
09-29-2005, 09:45 PM
Hi Bill, folks, Valentin is right: a DAYS BETWEEN DATES program does fit in the 15C memory. It even fits in the 11C memory. The following program has 140 lines and includes DAY OF WEEK as well (Of course Valentin would need only half of these, or even less if he manages to concoct a more amazing formula :-) . This is a direct implementation of the formulas used by the Master Library Module of the TI-59 calculator, one of my earliest 15C program, so do not expect anything close to perfection, quite the contrary! (Apparently, at the time I did not know what LSTx was for. I would also sometimes use 2 10^x instead of 100 just to save a step, even though this just makes the program slower, sometimes in the same program I would use simply 100).
f A works just like the DeltaDYS on the 12C, when D.MY format is set. Unlike the DATE function on the 12C, f B requires only one argument (the date) and returns 0=Sat, 1=Sun, ... 6=Fri. This is quite handy in Portuguese as the days of the week are the corresponding ordinal numbers from Monday (segunda = 2nd) to Friday (sexta = 6th) :-) Example:
01.012005 ENTER
29.092005 f B Regards, Gerson.
----------------------------- 1) The listing is supposed to be ok as I have just keyed it in the 11C and it worked. The program seems to be ok too. Of course g TEST 7 should be replaced by f x>y on the 11C. For speed, occurrences of 2 10^x and 4 10^x should be replaced by 100 and 10000 (or 100 g x^2), respectively. 2) The Brazilian edition of the TI-59/59 Master Library Module Manual has some typos (missing parentheses and brackets). I think the pencil annotation I made is correct:
if m < 3 then A funny mistranslation in this manual: The HI-LO GAME (ML-21) was translated as JOGO: Alô? Veja! (GAME: Hello? Look!) which made no sense at all! Apparently, the translator thought LO was short for 'LOOK' and HI was just the interjection 'hi'. (Edited to correct a couple of typos - I never get rid of them!) (Edited again to include a missing g INT in third line) ----------
To the list of beginner's mistakes in this program, I have to mention at least ten unnecessary ENTERs :-)
Edited: 1 Oct 2005, 1:19 a.m.
09-29-2005, 06:24 AM
Hi all, Hi Bill. Good question, Stephen. I'm into astronomy, so also into this stuff... (JD etc...) One thing about the forum, you sure get some REALLY GOOD ANSWERS!
Even if you were on a desert island and had no calculator, it's easy You just work out what the leap years are (easy since if they divide into four they are, except for years divisible by 400).
Make a small list of them (a single small look up table). DW
09-28-2005, 03:54 PM
For the 48 (and 49) series, you can use the DDAYS command, as long as the dates are from October 15th, 1582 through December 31st, 9999. With system flag -42 clear, the arguments should be formatted as MM.DDYYYY, and with flag -42 set, the format is DD.MMYYYY.
Regards, ▼
09-30-2005, 05:09 PM
Hi; the day-of-week (DOW in the HP41CX or HP41 Time Module) in the HP48G may be obtained with TSTR (Time STRing). A brief description of TSTR can also be found here: HP48GII Users´s Guide (chapter 25). TSTR uses the contents in both Level 2 (time stamp) and Level 1 (date reference) to compose a string (ALGEBRAIC mode demands arguments in the same order),. Given that both contents are valid, the resulting string (Level 1) is added a three-character reference head for the day-of-week (MON, TUE, WED...). Cheers. Luiz (Brazil)
Edited: 30 Sept 2005, 5:14 p.m. ▼
09-30-2005, 06:42 PM
Quote: Hi Luiz, TSTR is very handy, but the output is in English. What about days of the week in languages other than English, like good ol' Portuguese, for instance? << "SábDomSegTerQuaQuiSex" 1.012 ROT DDAYS 7 MOD 3 * 1 + DUP 2 + SUB >> Flag -42 setting is irrelevant! Just enter the date according to the date format on your calculator: 30.092005 => "Sex" (sexta-feira) Replace "SabDomSegTerQuaQuiSex" with your favorite string:
"SamDimLunMarMerJeuVen" etc... Cheers,
Gerson.
09-28-2005, 05:14 PM
I don't have a calculator program, but it is fairly easy: 1) Move new year to the first of march, so that march is month 0 and february is month 11 of the previous year. 2) Compute the days since some day in january 1 BC as dn = year DIV 400 - year DIV 100 + year * 1461 DIV 4 + month * 153 DIV 5 + dayofmonth DIV is integer division, a very useful operation on the HP 33S (but easy to substitute with divide and INTG). 3) Do this for both dates and subtract the results. If you want the day of week, this can be reduced to dow = (year DIV 400 - year DIV 100 + year * 5 DIV 4 + month * 13 DIV 5 + dayofmonth) MOD 7 This is called Zeller's congruence. You can ignore the year DIV 400 - year DIV 100 part if you are only interested in the 20th and 21st centuries. ▼
09-28-2005, 05:48 PM
Convert each to Julian Day Numbers and subtract. http://scienceworld.wolfram.com/astronomy/JulianDate.html Or if you can limit your range, use the Modified Julian Day Numbers (note that these change at midnight, not noon!) http://scienceworld.wolfram.com/astronomy/ModifiedJulianDate.html ▼
09-28-2005, 08:21 PM
Thanks (everyone) for a terrific response. I knew it had something to do with Julian date, but I was wondering if there were other ways since some of these calcs keep time / date. ▼
09-29-2005, 05:38 AM
Quote: You do know that the HP-12C even has a button to calculate days between dates, don't you? ▼
09-29-2005, 12:23 PM
Yeah, I have the 12C and love this feature about it. I have several other HPs that calculate elapsed days. For the ones that don't calculate this, I want to write a short program to calculate it. Reason: I'm a medical physicist and frequently I need to calculate elapsed days for a cancer patient's treatment (from start to finish). This is most helpful for the physicians, really, but I have to "QA" every patient's chart weekly. For years I either do it in my head or manually (but using a calc). Now I'm learning to think smarter (or just being lazy!). I think I will first try the Modified Julian Date method since it seems easier. I have over a dozen new HPs added to my collection and trying to rotate them in/out of use (at home and at work) and need them to do what I want them to do. (My first HP was a 28S in '89 and cut my teeth on it.) ▼
09-30-2005, 12:27 PM
Meanwhile, you can use the 15C program above. It is no masterpiece but it works! In testing it, I can see that today I am exactly 16141 days old, a prime number. And that I got married on a Wednesday (11/11/1987, all prime numbers too... but that was not on purpose), which matches what I can remember. To use the MM.DDYYYY format you're used to, just replace the first two lines with these:
f LBL 0 x<>y STO 0 x<>y ENTER g INT STO 2 Rv RCL 2 - 100 * ENTER Best regards, Gerson.
09-28-2005, 07:27 PM
A simple program appears in the HP25 Applications Handbook; a more sophisticated version is published in the HP41 Applications Pac. The HP25 version is quite simple, due to the limitations of that model (one of my favorites, indeed!). If you would like to understand how it works, and then adapt it to other machines, it may be a very appropriate starting point. ▼
09-29-2005, 02:22 AM
I wrote a version of this prg for the 42S and I like very much how it works. Feel free for asking if interested. Raul ▼
09-29-2005, 12:25 PM
Yeah, I'm interested. I have a 42S, too, and would like to rotate it in/out of professional / home use and be able to calculate elapsed days. If it makes it easier, my email address is.... s_easterling@earthlink.net Thanks!
Stephen
09-29-2005, 02:50 PM
Also check out the PPC ROM Manual (for the HP41) for the functions CJ and JC, these have an excellent backgrounder. These work with both Julain and GRegorian calendars. Meindert
09-30-2005, 12:30 AM
Have a look at the Calendar Functions program for the HP-67/97 http://www.hpmuseum.org/software/67pacs/67calend.htm **vp ▼
09-30-2005, 02:21 AM
I was wrong above: the program I adapted for the 42S, is this one for the 67 of the software library. Raul ▼
09-30-2005, 09:43 PM
It might be appropriate to be careful about using days beween dates programs from the 1900's since many of those programs did not recognize that the year 2000 would be a leap year (the Y2K problem). If you translate one of those old programs you should check that the number of days between 01/01/2000 and 01/01/2001 is 366. The routine from the TI-59 Master Library mentioned in Gerson Barbosa's submission gets 366. If you are a dedicated purist, and what RPNer isn't, you might want to check the number of days the routine finds between 01/01/4000 and 01/01/4001. It should be 365. The routine from the TI-59 Master Library gets 366, but everyone knows that AOSers aren't purists. ▼
09-30-2005, 11:01 PM
4000 will be a leap year since it is divisible by 400, won't it? Anyway, the program fails for 1700, 1800, 2100, 2200, etc. However I am not sure whether this is due to something wrong in the TI-59 Master Library routine or in the 15C program since I don't have a TI-59. ----- I have just checked 01/01/2100-01/01/2101 and 01/01/4000-01/01/4001 on Miroslav Nemecek's TI-59 Emulator and the results match those of the 12C. I have to see what is causing the 15C program to return 366 for 01/01/2100-01/01/2101.
Edited: 30 Sept 2005, 11:36 p.m. ▼
10-01-2005, 02:25 AM
There was just a missing g INT in the third line. Here is a 15C lighter version after removing a lot of unneeded ENTER's:
Date format is MM.DDYYYY f A: works just like the DeltaDYS on the 12C. f B: similar to DATE on the 12C, but requires only one argument, the date, and returns 0=Sat, 1=Sun, ... 6=Fri. On the 11C, g TEST 7 should be replaced with f x>y Everything appears to be all right now. Sorry for the inconvenience. ▼
10-01-2005, 03:41 AM
I notice that you have Saturday=0 through Friday=6, but ISO 8601
Here in the U.S.A., our calendars have stuck with the tradition of
Just out of curiosity, of the cultures represented in this forum,
Does anyone know when European calendars switched to starting the
Regards, ▼
10-01-2005, 12:01 PM
Quote: I thought it was quite the contrary, because as least in two European languages, Portuguese and German, the week seems to start on Sunday. The word for Wednesday in German is Mittwoch, which means, I think, "middle of the week". If this is correct, then Wednesday is the fourth day of the week and therefore Sunday is the first day. In Portuguese, as I have already mentioned, the business days are ordinal numbers: segunda-feira, terça-feira, quarta-feira, quinta-feira and sexta-feira, litteraly "second fair", "third fair",... "sixth-fair" (fair = market place). I think this is because the medieval fairs that were held throughout the week, but I may be wrong (I have heard of another reason, but I don't remember what should it be). If European calendars now start the week on Monday, you may be right about the reasons. Soon after the French Revolution, for a while they even abandoned the Gregorian calendar. Regards, Gerson.
▼
10-02-2005, 01:50 AM
Note that language doesn't always stay in synch with the calendar. For example, our English September, October, November, and December have roots in the Latin for seven, eight, nine, and ten, even though March hasn't been the first month for quite some time.
Regards, Edited: 2 Oct 2005, 2:15 a.m.
10-05-2005, 03:26 PM
Quote: Before 1976 in Germany Sunday was the first day of the week, since 1976 Monday is the first day according to DIN 1355 (for those who understand German http://de.wikipedia.org/wiki/Woche). About a later question if calendar weeks are used, I can definitely say "Yes" for Germany. Mostly every working dead line is given as calendar week. In practise this means, when the dead line is for example KW39 (German for Kalenderwoche 39, 09/26/05-10/02/05) you should have it on Monday KW40. ;-) Finally I want to point to my date conversation routines for the HP-42S which can be found at http://www.hp42s.com/programs/date/date.html working with every date of the Gregorian calendar (October 15/1582 to December 31/9999). Technical notes about implementation can be found at http://www.hp42s.com/programs/date/olddate.html. Regards, Christoph, Germany ▼
10-07-2005, 07:44 AM
Thank you! I was very curious as to when the week changed from Does anyone know of a country that changed earlier?
But really now, a DIN was published and everyone changed their
Okay, I realize that it's an ISO standard and ANSI (American
"Weekly planning" pages in business "planners" start with Monday
But the planners do show the ISO 8601 week number on the weekly
"KW40" is just fine for use within German-speaking countries, but
Of course an American would understand 10/02/05 or 10-02-05 as
I doubt that ISO week dates will catch on in the U.S., except
First there's the oddity of the week starting with the "wrong"
Worse, the first week of the year may lack up to the first 3 days
Regards, ▼
10-11-2005, 05:33 PM
Quote: Hi James That's funny, we all here are aficionados of artefacts that really does catch with week days: our hp calculators. Have a look at the serial number coding that hp uses for the calculators. Here they are - in a big U.S. company. The calendar weeks you call obsolete... :-D Valentino ▼
10-12-2005, 03:26 AM
First off, the use of a calendar week number in HP serials would be an example of "very limited use". And yes, I was well aware of it. I wonder, does HP's week numbering system follow ISO 8601? There are quite a few other ways one could number the weeks in a year. I never called it "obsolete", just of limited usefulness.
Regards, Edited: 12 Oct 2005, 6:06 a.m.
10-07-2005, 01:11 PM
Quote: Very interesting routines! There's even one for calculating the day of the Easter! About twelve years ago a wrote a 42S program for doing this. No, not for religious purposes, just for planning my vacations. I had to choose my vacations at least six months in advance, and I wouldn't want them to coincide either with Carnival or Easter, since here they are a five-day holiday anyway :-) By what I can remember, the program is based on a formula by Euler (I simply ported a BASIC program that was published in a local scientific magazine to the HP-42S). It works for any year in the range from 1900 to 2099 (enough for what I had in mind). For example, given 2005 the output will be:
Carnival: Feb-06-2005 The program (361 bytes long) is far from being of general interest. Anyway, I can provide a .RAW file for Emu42 (Thanks, Christoph!) if anyone is interested. Regards, Gerson.
Edited: 7 Oct 2005, 1:18 p.m.
10-02-2005, 04:42 AM
Quote: I did some googling. It turns out the reason is theological: 1) The third commandment says: Remember thou keep holy the Sabbath day 2) The Christian Sabbath is Sunday 3) The Sabbath is the seventh day Therefore, Monday is the first day of the week. QED. Here are some links:
The Days of the Week. Lots of information about weekdays names in different languages, and this: Quote: Quote:
Sunday is NOT the Sabbath!. This page explains why real protestants keep Saturday holy: Quote:
▼
10-02-2005, 08:35 AM
Granted, the Jewish may use "Remember the Sabbath Day to keep it
Since the word "remember" is used in the commandment, I surmise
Note that the seven-day week wasn't restricted to Jews in
If I recall correctly, my catechism said "Remember to keep holy
As far as I can determine, the traditional seven-day cycle of the
Anyway, in the U.S.A., calendars still show the week as starting So where does the week start on your calendar?
For that matter, do you ever use the ISO 8601 week of year and day
By the way, I'm trying to avoid starting any religious-based flame
Regards, Edited: 2 Oct 2005, 8:44 a.m. ▼
10-02-2005, 10:23 AM
Quote: I think that if the reason really was to make Sunday the seventh day of the week to make it fit the Bible, then it sounds more like a 18th or 19th century thing, at the latest. Few 20th century Europeans would bother.
Quote:
The week always starts on Monday here. Software that does not allow configuration of the start of the week is broken. Quote:
My Nokia supports weeks starting on Saturday, Sunday or Monday, and shows ISO week numbers if it starts on Sunday or Monday.
Quote: No, that was my worry after I posted :-)
Quote: Yes, it is very interesting to discover the logic, to the extent there is any, behind the things we take for granted. It is very often based in religion.
10-03-2005, 02:01 PM
In Chinese, Monday is "Week 1" Tuesday "Week 2" etc until sunday which is "Week Day" Arnaud
09-30-2005, 11:02 PM
Must be a part of the algorithm I'm unfamiliar with. Why wouldn't 4000 be a leap year, since it's divisible by 400? ▼
10-01-2005, 12:58 AM
In the Gregorian calendar, a year divisible by 4 is a leap year, unless it's also divisable by 100, in which case it's not a leap year, unless it's also divisible by 400, in which case it's a leap year after all.
Regards, ▼
10-01-2005, 01:16 AM
Yes. ▼
10-01-2005, 02:41 AM
Sorry Howard, I guess that I got mixed up as to whom I meant to reply to. By the way, I strongly suspect that what Palmer had in mind was a proposed additional rule, suggested by the famous astronomer John Herschel, that years divisible by 4000 not be leap years. This would change the average calendar year length from 365 97/400 days to 365 969/4000 days, and would be closer to the observed length of a tropical year, although still a trifle long. However, as far as I know, no church, government, or standards organization has ever officially adopted this rule; it's certainly not part of the Gregorian calendar. Of course, with the leap seconds, years are (on average) slightly longer, and I suppose that leap seconds will be needed more frequently in the future, which would move the calendar even farther out of synchronization with the seasons, so maybe Herschel's proposed rule wouldn't be enough by the time the year 4000 gets here. But I've read that the U.S. has (unfortunately, IMO) proposed eliminating these adjustments to keep UTC in synch with astronomical time. I think dealing with the leap seconds is enough for now; changes to the calendar rules can be left for future generations to deal with.
Regards, ▼
10-01-2005, 04:29 AM
Quote:According to HP, there are 365.242198781 days per year. ▼
10-01-2005, 05:47 AM
Actually, at least in the 48/49 series, the units base for 1 year is 31556925.9747 seconds, but I don't know where they got that value from. Of course they equate 1 day to 86400 seconds, so they don't attempt to take leap seconds into account. So yes, 1_yr converts to 365.242198781_d in these calculators. Anyway, a Gregorian calendar year, averaged over the 400 year cycle, is 365.2425 days. With Herschel's proposed change, averaged over the resulting 4000 year cycle, it would be 365.24225 days; still a little long. Of course the Greek Orthodox calendar uses a different set of rules for exceptions to having a leap year every 4th year, which brings them a bit closer to the observed tropical year.
Regards, Edited: 1 Oct 2005, 5:50 a.m.
10-01-2005, 09:47 PM
I was relying on the following quotation from page 619 of Volume 4 of the 1969 version of Encyclopedia Britannica: "... Later, a slight change was made in the Gregorian calendar to bring it still more closely into line with the tropical year. The Gregorian calendar was still in error by one day in 3,323 years and, in consequence, a further rule of intercalation has been adopted that makes the year 4000, 8000, etc., common years, i.e., years without an intercalated day. The calendar is now, therefore, correct to within one day in 20,000 years. ..." ▼
10-02-2005, 12:50 AM
I'm surprised. At least ISO 8601:2000 section 4.3.2.1 includes:
"The Gregorian calendar distinguishes common years with a duration of 365 calendar days and leap years with a Could it be that the Encyclopedia Britannica is mistaken? That said, I haven't seen a copy of ISO 8601:2004, and don't feel like paying $101 for a copy of it.
Regards, ▼
10-02-2005, 01:34 AM
PS: With an error of one day in 3,323 years, the proposal to drop leap years every 3200 years seems to make more sense. See: http://mindprod.com/jgloss/leapyear.html. Of course, I don't see calendar reform as being particularly urgent.
Regards,
▼
10-02-2005, 02:14 AM
This is all very interesting and enlightening. I've learned a whole lot more about calendars than I knew before this thread. One thing I did know was that leap-seconds aren't added for the purpose of correcting for the base error in the calendar the way additional leap-year days are. They are there to compensate for the slowing of Earth's rotation, and consequent lengthening of the day. But what makes us look for an even multiple of 100 to correct an error whose period is 3,323 years? In other words, why not make the year 3323 a leap year, instead of 4000 or even 3200?
Edited: 2 Oct 2005, 2:15 a.m. ▼
10-02-2005, 05:02 AM
Yes, seconds are added (or dropped, though that hasn't occurred)
However, a leap second does have the additional effect of
How many years will it take for 86400 leap seconds to accumulate?
As for looking for an integer multiple of 100 instead of simply
Of course, by the year 83100 (or even 3323, for that matter), who
Maybe we'll just wait until the discrepancy between the calendar
By the way, the Gregorian leap year rules are intended to keep the
Of course, even though the calendar reform was for ecclesiastical
Regards, ▼
10-02-2005, 09:05 AM
Ok, folks, let's strive to reach the year 3323 at all! Remember oil will be exhausted within the next 50 years. And there is the greenhouse effect leading to some more severe hurricanes and similar stuff. Oh sorry, I forgot, according to the present administration there is no greenhouse effect! So there is a good (?) chance a well known superpower will start a war on this with the result there is no need for long range calendar corrections anymore. Good luck! ((;-) ▼
10-02-2005, 09:40 AM
Hey, I'm pretty certain that I, personally, won't reach the year 3323. Yes, there are certainly plenty of problems far more urgent than additional leap year reform to be solved. I think we can leave that issue for (hoped-for) generations far in the future to address. Of course, I do rather wish that the silly ISO 8601 week of year numbers would be dropped; since a Gregorian calendar year never has an integer multiple of 7 days in it, there doesn't seem to be any good way to number the weeks of the year. And since there seems to be very little chance that those cultures that start the week with Sunday will ever use the ISO day of week numbers starting with Monday as 1, they seem much more likely to cause confusion than to standardize usage
Regrds,
10-02-2005, 02:50 PM
Hi all. Interesting thread. NOTE: The following IS kinda o.t.: (but provided as a community service announcement of sorts)
Actually the permanent oil crisis is here now.
The (fatal?) impacts will probably arrive in an irreversible form Maybe someone might like to do a quick simulation (on an hp calc of course) to predict economic collapse from incremental rise in oil prices (with resulting flow on effects into EVERY COMMODITY REQUIRING TRANSPORT). The modern economy is in fact very fragile and Bush and Cheney know it. So does John Howard and Tony Blair.
Of course, one (tried and tested) way to avoid collapse is to steal All the best for the future everyone. Don W
Edited: 2 Oct 2005, 2:51 p.m.
10-02-2005, 10:55 PM
Quote: In order to achieve this, shouldn't we also compensate for the effect of the precession of the equinoxes on the dates seasons begin? Any ideas? Regards, Gerson. ▼
10-04-2005, 06:17 PM
Quote:
I think that if we were using a sidereal (star-based) calendar
But in fact the Gregorian calendar is designed to keep the
Regards,
10-02-2005, 01:00 PM
Perhaps neither TI nor HP expected their products to last until the year 4000 A.C. :-) Though the calendar is now that accurate, you'll be celebrating Christmas in the Summer, like we do down here, by the year 15,000 A.C. ... Will this be a problem? Besides your post being enlighting, it helped me spot a mistake in my listing that would prevent the program to work correctly for some years ending in 00. (My hand-written listing on the Master Library was ok though). Thanks.
Edited: 2 Oct 2005, 1:03 p.m. ▼
10-03-2005, 05:51 PM
Quote: But, apart from battery problems and capacitors drying: Is the "plastic" body of a classic, woodstock, voyager, pioneer or 41/48/49 still all right in a few 100 years? Or even thousands? And what about the metal parts of the older calculators? And the platin and chips on it? How long could they last? Anyone here with that material knowledge? ▼
10-03-2005, 10:28 PM
Assuming they are used, the ICs will eventually fail from electromigration of the metal layer. The time period depends on feature size and current density.
10-03-2005, 10:58 PM
My guess is that plastic cases will become a problem. Some plastics become brittle with age. One of the prized items in my collection was a Speedee Add-a-matic (see the section of the museum on old calculators). Last spring I inadvertently knocked it off a table onto a terrazo floor. The plastic case shattered into about twenty pieces. The metal mechanism still works. By comparison back in the early 1980's I inadvertently knocked my almost new Radio Shack Model 100 off the same table onto the same terrazo floor. There was a small crack in one corner of the case. Some of the keys came off and the batteries came out of the compartment. I put the batteries back in, pressed the keys back in place and the machine was able to complete the calculations that were in process. That machine still works today. A "shock test" which illustrates how sturdy some of the hand-helds are when they are relatively new is the test of the oscillator described in the service manual for the TI-59. I'm not going to repeat that test today. Sorry about that! Who's going to do a drop test on his HP-35? I do have some old Sharps, Casios, and a number of spare TI-30's at my winter home. When I'm back there I just may have to do some drop tests. I also have three of the Addometers (again, see the old calculator section) which were all metal and made by a typewrter manufacturer back in the 1920's. Two of the three work well. The third is somewhat rusty and the dials only turn with some difficulty. A bad choice of metal can be a problem. The early Pickett slide rules used a magnesium alloy (I think) base. They often didn't move smoothly even when new but that problem can be solved these days with a little WD-40. Many of those now have the slide and frame fused together by corrosion. The late aluminum base devices have fewer problems.
10-03-2005, 07:53 PM
"Though the calendar is now that accurate, you'll be celebrating Christmas in the Summer, like we do down here, by the year 15,000 A.C. ... Will this be a problem?" Nope! It is, in fact, precisely the addition of leap years which KEEPS the seasons where they belong, due to the precession of the equinox. Us northerners will still have spring starting (about) March 21 and Christmas will still be in the winter. What will happen in 13000 years (half of a precession cycle) is that the stars which you now think of as "winter" stars (again, for us northerners; perhaps better stated as "December" stars) will be instead the summer/June stars. i.e. the constellation Orion will be prominent on June/July evenings in the year 15000, rather then the December/January evenings when we see him now. Another minor effect will be due to the ellipticity of the Earth's orbit. At this time, the Earth is closest to the Sun in early January. The effect on incoming solar energy is only a few percent, but this means that northern hemisphere winters are now somewhat warmer than they will be in 13000 years, when winter will be occuring when the Earth is farthest from the sun. Leap seconds (or the alternative now being touted by some folks at the US Naval Observatory - officially charged with keeping track of time for the United States - of changing the rate of atomic time by about 20 parts per billion) keep the rotational position of the Earth aligned with the stars. I have to think a bit about whether the accumulated leap seconds affect leap years. The Earth is not only slowing down overall, but at the level of 10's of microseconds per day speeds up and slows down - somewhat at random but also with seasonal variations that basically relate to the fact that there is more land mass in the northern hemisphere than in the southern hemisphere. This stuff is all very critical for very accurate (sub-centimeter) global geodesy. For more details on precise earth orientation and timekeeping, go to http://maia.usno.navy.mil . ▼
10-03-2005, 10:50 PM
Quote: I always thought the addition of leap years keeps the seasons in place because this makes the year closer to its actual length. But the seasons shifts due to the precession of the equinoxes would not be corrected by this method. Have I misunterstood anything? Here is an excerpt of Isaac Asimos's "A choice of catastrophes": "... In 12890 years, the [Earth's] axis shall turn to the opposite direction... ...and the Summers's Solstice [in the northern hemisphere] will be on December 21, and the Winter's Solstice on June 21..." (This is a bad translation back to English. The actual text should be somewhat different, but hopefully I may have kept the right meaning).
10-04-2005, 12:11 AM
This Message was deleted. This empty message preserves the threading when a post with followup(s) is deleted. If all followups have been removed, the original poster may delete this post again to make this placeholder disappear. ▼
10-04-2005, 12:49 AM
Don, Precession can be viewed as the "sliding" of the celestial reference frame (right ascension and declination - "RA/Dec" in astronomer shorthand) which is tied to the Earth (basically projection of latitude and longitude onto the sky) along the ecliptic (the more-or-less plane of the planetary orbits). If you think of the position of stars in an ecliptic coordinate system, they will not change with time. In the RA/Dec system, the star positions change continuously. So yes, the RA/Dec of Orion will change, but Orion will still be in the ecliptic, as will ALL the ecliptic constellations (so beloved of astrologers (gack, I HATE that word!). The path of the Sun, as viewed from the Earth, will continue to circle around the ecliptic once per year. Also, as I said earlier, leap years REALLY DO take care of the sliding of the seasons. The reason calendar reform was necessary was because the date March 21 was clearly no longer coinciding with the "beginning" of spring (defined as the day of the year on which the declination of the Sun is ZERO (i.e. the Sun was exactly on the celestial equator = Declination of the Sun is 0 degrees)). What was the problem: the length of the year was too short (as noted elsewhere in this discussion) if only 365 days were alloted to a year. So, leap years merely adjust the length of the year to match the actual time that the Earth takes to go around the Sun - with year here defined as the length of time from the beginning of one spring to the next spring. That's what our current calendar is based on - the spring-to-spring year. If the date of the seasons was going to switch, we would have already seen the effect. This problem has been known, more or less, for several thousand years. That would be enough to switch the beginning of spring by a month (i.e. around a twelfth of the 26000 year precession period) - which has not happened. So, (as long as we stick to the current calendar or some close approximation) Spring will continue to occur on March 21 for the forseeable future. In particular, for the next 13000 years. For more on this, try http://www-istp.gsfc.nasa.gov/stargaze/Sprecess.htm ▼
10-04-2005, 11:26 AM
Quote: Now you have convinced me! Thanks! Ne sutor ultra crepidam. From now on this shoemaker will stick just to heels, soles, shoelaces and the like :-)
10-06-2005, 05:14 PM
My attempt to explain the "precession of the equinoxes",
But I'm not anastronomer; I wouldn't even consider myself an
First off, we have at least three major definitions of a year
A "tropical" year, the period of the seasons. For example, from
A "sidereal" year, the period of the earth's revolution around the
An "anomalistic" year, the period from the earth's perihelion
The ecliptic is the plane of earth's orbit, so called because an
The celestial equator is the earth's equator projected into the
The celestial poles are the earth's axis of rotation projected
Currently, the planes of the ecliptic and celestial equator are at
An equinox occurs when the earth passes through the line of
Because the earth spins, it has an equatorial bulge. The sun's and
As noted above, the point of perihelion is precessing, and the
Because the earth's orbital speed (like a comet's) is fastest at
We all know that the sun appears to revolve around the earth due
After all, if the earth didn't rotate at all in relation to the
A sidereal day, the period of Earth's rotation relative to the
Because the earth's orbital speed is fastest at perihelion, and
No doubt Earth's actual orbit is much more complicated than I've
The relationships of the above periodic motions are the bases of To simplify timekeeping, we use a "mean solar day".
Of course now we no longer define a second as 1/86400 mean solar
As noted in another post, earth's rotation varies, and it's
Regarding leap years, to be pedantic, they're used simply because
Before Julius Caesar's reform, Roman years varied in length,
Of course now we change the dates of the "fiscal year" instead,
Anyway, a fixed-length year would be an obvious solution the
Julius Caesar's calendar reform came fairly close. The idea is to
Unfortunately, with Julius Caesar already dead, his reform was
Of course it would've been more "elegant" to distribute the days
Month normal year leap year
and for that matter, start both month #1 and the year on (okay,
Why the vernal equinox? Well, it's the beginning of spring, a
Of course the year has actually started on various dates in
But the Julian calendar was certainly a huge improvement from the
Notably, under the Julian calendar, the vernal equinox was
Of course they also dropped 10 calendar days, so 1582 had only 355
In my opinion, it might've been simpler to just acknowledge that
Of course other leap year rules, some arguably better, can be (and
Regards, ▼
10-07-2005, 08:17 AM
Quote:
But after thinking a little more, I realize that this would've
Regards
10-04-2005, 04:12 AM
Quote: "A.C."?
Indeed, the valid "system time" on the 48/49 series covers only a
I find it interesting that they didn't move the range ahead a bit
Apparently it's an "artificial" restriction; with the 52-bit clock
Maybe they simply chose the range so that TSTR would only have to
Similarly, the range of 1582-10-15 through 9999-12-31 for date
But I doubt that they expected the calculators to still be in use
But with the most recent models, I sometimes can't quite help
Quote:
Well, barring some major medical breakthroughs and several other
As others have pointed out, maybe it could be considered
Anyway, the way I figure, ignoring the effects of leap seconds,
Quote:
I rather doubt that the few days drift would be a problem, unless
Certainly it would be noticeable to astronomers, but I trust that
As for celebrating Christmas in the summer, at least they wouldn't
Do Australians include artificial snowflakes, snowmen, and such in
Quote:
Well, I'm glad that it helped. I confess that I didn't even
Regards, Edited: 4 Oct 2005, 4:19 a.m. ▼
10-04-2005, 06:05 AM
This Message was deleted. This empty message preserves the threading when a post with followup(s) is deleted. If all followups have been removed, the original poster may delete this post again to make this placeholder disappear. ▼
10-04-2005, 08:47 AM
Maybe your snowflakes and snowmen make it feel a little cooler? I
Regarding the pagan origins for Christmas, St. Valentine's day,
Around here, we really need something like Christmas
The kids colour hard-boiled eggs before Easter, so they know where
Christmas is a different story though. I'm always amazed that most
Regards,
10-04-2005, 09:50 PM
Quote: I should have said AD (Anno Domini). As I know you use the abbreviation BC (Before Christ), I wrongly guessed you'd use also use AC (After Christ) as we do here.
Regards, Gerson.
▼
10-06-2005, 01:57 AM
A.C. (After Christ) is new to me, though I'm familiar with C.E. (Christian Era) and B.C.E. (Before Christian Era) for the proleptic Gregorian calendar. Incidentally, this is usually taken to have a year 0000, which is a leap year, so the same leap year rules apply for negative years. For the Julian calendar, which doesn't have a year 0, A.D. and B.C seem to be the rule, but of course these have also been used for the Gregorian calendar. I suppose that C.E. and B.C.E. might by applied to the julian calendar too. I guess for dates before all western cultures adopted the Gregorian calendar (around 1750?), it's best to be specific about which calendar is meant. For my niece's genealogical research, I think that "New Style" and "Old Style" dates have caused some apparent anomolies. Which date something occurred on depends on which calendar is used.
Regards, ▼
10-06-2005, 08:19 PM
More generally referred to in non-religious context as Common Era (CE) and Before Common Era (BCE). See the Wikipedia article on the Common Era. ▼
10-07-2005, 03:47 AM
Okay, now that you mention it, I think I've also seen "Common
But if they wanted to get references to Christ out of the date,
And of course Gregory was a pope, so shouldn't we stop using the
Like it or not, Christianity has had a real influence on history,
Not that I feel that a calendar for secular purposes particularly
Regards, ▼
10-08-2005, 11:55 PM
The problem is that A.D. stands for Anno Domini, "Year of our Lord". For a huge number of people, Jesus isn't "our Lord". ▼
10-09-2005, 12:37 AM
Quote:Yes, obviously, but few would seriously doubt that Jesus Christ was a very real historical person. Even those who might think that the New Testament were completely fictitious wouldn't dispute the historical reality of Christianity, hence my mistake that "C.E." stood for "Christian Era". After all, whoever came up with "C.E." and "B.C.E." didn't choose to base a calendar on a new epoch that had nothing to do with Christianity.
Regards, Edited: 9 Oct 2005, 1:14 a.m.
10-09-2005, 03:02 PM
Quote: So that they wouldn't have to change the dates of everything. The epoch of the Gregorian calendar (and the Julian calendar) isn't really the birth of Christ anyhow; scholars seem to think that happened some time around 4 or 5 BCE. ▼
10-12-2005, 04:42 AM
Quote:So it seems to me that they weren't really very serious about removing any religious references from the calendar. Obviously, "A.D." (Anno Domini, Year of the Lord), would seem very inappropriate to non-christians, but I don't see why anyone would object to, for example, "Christian Era". Quote:Yes, there's very good reason to believe that the birth of Jesus of Nazareth actually occurred a few years before our calendar would seem to indicate. However, the particular scholar (Dionysius Exiguus) whose work the epoch is based on, apparently believed that Jesus turned 1 year old in 754 A.U.C., so he established that as the year 1 A.D. How certain he was, I don't know.
Regards,
10-05-2005, 12:26 AM
Quote: The current Gregorian calendar is slow by 25.92 seconds/year (much better than the 11 1/2 minutes of the Julian calendar!); the Y4K proposal makes it slow by 4.32 seconds/year. However, that means that there will be a Y20K problem. Note too that the Eastern Orthodox church in 1923 proposed an alternative rule to correct the leap year problem in the Julian calendar. In their system, century years modulo 900 must have a value of 200 or 600 to be considered a leap year. This first year of divergence is in 2800, where the Gregorian calendar has a leap year but the Orthodox calendar does not until 2900. The Orthodox calendar is slow by 1.91 seconds/year, so their day of reckoning doesn't happen until Y45K. An additional problem is that the end of the current interglacial period (not "the next ice age" since we're in the middle of an ice age now!) is due around Y25K and may cause further perturbations on the Earth's rotation and revolution. ▼
10-05-2005, 11:10 AM
re: "An additional problem is that the end of the current interglacial period (not "the next ice age" since we're in the middle of an ice age now!) is due around Y25K and As I try to make the difference when I teach Astronomy 101, "rotation" is the turning of the Earth on its axis, whereas "revolution" is the orbital motion of the Earth around the Sun. So, while a new ice age may well affect the rotation rate of the Earth (due to changes in the moment of inertia of an ice-clad Earth) it WILL NOT affect the revolution (i.e. the orbital period). Sorry to be so technical, but when among geeks and nerds, we should make sure we have it right!! ▼
10-06-2005, 12:35 AM
Quote: Note that I said "the end of the current interglacial period" and not "the return of the glaciers." The latter is one aspect of the former. Consider the possibility that ice ages and interglacial periods are caused by something that also impacts the revolution of the Earth. That's what I meant when I said "rotation and revolution". Rotation can be affected by the ice, whereas revolution *and* ice could be affected by some common external cause. This is all speculation, of course.
10-05-2005, 10:43 PM
My impression is that the additional rule was actually adopted by
I wouldn't expect lunar calendars for religious and traditional
Regarding ice ages, land in many northern (and southern?) regions
What happens if (when?) the ice caps on Greenland and Antarctica
Of course the distribution of inland water varies, largely due to
I've read that Earth's core rotates at a different rate than its
Quite a lot of things affect the distribution of Earth's angular
But all of the above affects Earth's rotation (its day length),
To the extent that the average length of a day varies, the number
To affect the revolution, we need something outside of the
Regards, ▼
10-06-2005, 12:25 AM
Quote: Hence my reference to the Earth's revolution; as we are not certain what causes ice ages, we need to consider the possibility that it is due to something external to the Earth-Moon system.
10-01-2005, 02:59 AM
The 67 program, and the 41 and 42 versions I rewrote, give 366 :-( |