I would keep in mind that calendars have a political aspect, and some political bodies will sometimes create arbitrary calendars (I remember a 41st of September used by the U.S. Congress a few decades ago - they had set themselves a deadline where they needed to resolve an issue before the end of September and then decided it was more motivating for them to modify what September meant than to modify the deadline).
Calendars might also be used in less official works of fantasy. Anyways, if you do not like this behavior, it's easy enough to create a validDate verb which is an identity verb which generates errors for invalid dates. And, once created, that verb could be used with a variety of date manipulation verbs, in addition to todayno. Thanks, -- Raul On Wed, Sep 24, 2014 at 7:52 PM, robert therriault <[email protected]> wrote: > I certainly see your point Linda, > > The tests that I run my program through are an example of creating triggers > to will reveal when my programming has created errors. If I was designing a > language that could be used in as wide a variety of ways as J, I think that > knowing when to generate meaningful error messages would be a very > challenging task. I do think that in the case of 2002 3 1 that the value > returned is meaningful, since that day exists. The one that I would question > would be 2002 2 29, which is not found on the calendar but still returns a > value. > > Cheers, bob > > On Sep 24, 2014, at 3:42 PM, Linda Alvord <[email protected]> wrote: > >> Using J often causes error message and they are very helpful. They cause a >> reassessment of the statement in question. Most often it causes a quick fix >> as you notice the problem. Since there is reasonalble response to todayno >> 2002 2 29 >> I believe an error message would cause me to find the problem. >> >> It seems incorrect to provide the value for 2002 3 1 is not helpful. >> >> This may just be a minority opinion. >> >> Linda >> >> -----Original Message----- >> From: [email protected] >> [mailto:[email protected]] On Behalf Of robert >> therriault >> Sent: Wednesday, September 24, 2014 10:49 AM >> To: [email protected] >> Subject: Re: [Jprogramming] Weekend Puzzle - Age of Groundhog born 2002 2 2 >> >> Hi LInda, >> >> Seems okay to me, since >> >> (2001 3 1) diff 2001 2 28 NB. non leap year >> 1 >> (2000 3 1) diff 2000 2 28 NB. leap year >> 2 >> >> diff=: 13 :'(todayno x)-todayno y' >> (2005 3 1) diff 2004 2 28 >> 367 >> (2004 3 1) diff 2003 2 28 >> 367 >> NB. both these intervals contain 2004 2 29 because of a leap year >> >> (2003 3 1) diff 2002 2 28 >> 366 >> (2002 3 1) diff 2001 2 28 >> 366 >> >> NB. these intervals don't contain a leap day >> >> Cheers, bob >> >> On Sep 24, 2014, at 2:06 AM, Linda Alvord <[email protected]> wrote: >> >>> Does this seem ok to you? >> >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm >> >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
