🚀 go-pugleaf

In <4ksb9p$1eps@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>Today I was revising a program at work (OS2 REXX) and to my
>surprise I could not find a way to do date arithmetic. I found 
>the date("base") which returs a julian date, but I could not
>find anything that takes that format back into any other format.
>
>
>Did I mean something? I would think that date arithmetic should
>be something that one would want as part of the standard REXX.

Date processing involves algorithms and definitions of intention, somewhat like 
litigants in a courtroom.  It does not define a language.  

Go to a search engine and plug in the keyword variables you want.  Do a bit of 
research,  and quit whining.

Warren
http://www.borg.com/~warren

javier5@ibm.net wrote:
> Today I was revising a program at work (OS2 REXX) and to my
> surprise I could not find a way to do date arithmetic. I found 
> the date("base") which returs a julian date, but I could not
> find anything that takes that format back into any other format.


> Did I mean something? I would think that date arithmetic should
> be something that one would want as part of the standard REXX.

It's not standard part of Rexx, but take a look at:
    ftp://hobbes.nmsu.edu/os2/dev32/rexx/rxdate20.zip 

It might just be what you need.

--
Con Vissenberg                 | GCS  d-- H- s g- p?>1 a w+ v C+ US+$
cvissenb@emt.ie.philips.nl     |      !P L>+ 3 E N++ K W--- M? !V -po+
I don't speak for Philips.     |      Y+ t !5 !j R G? tv b+ !D B? e++
                               |      u+ h---- f r+++  n+ z+++
 
 

In <4ksb9p$1eps@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>Today I was revising a program at work (OS2 REXX) and to my
>surprise I could not find a way to do date arithmetic. I found 
>the date("base") which returs a julian date, but I could not
>find anything that takes that format back into any other format.

Javier5:

You might want to try

 ftp://hobbes.nmsu.edu/os2/dev32/rexx/datergf1.zip

 datergf1.zip Jan 31  1995    22043 REXX date arithmetics/conversions

The upcoming Object Rexx BTW has provisions to transfer dates from one
format to another.

Regards,

---rony

 
>  
> Today I was revising a program at work (OS2 REXX) and to my 
> surprise I could not find a way to do date arithmetic. I found  
> the date("base") which returs a julian date, but I could not 
> find anything that takes that format back into any other format. 
>  
>  
> Did I mean something? I would think that date arithmetic should 
> be something that one would want as part of the standard REXX. 
>  
 
I see that there have been some good, and, uh, not so good, replies  
to your note, but I'll go ahead and throw my in two cents worth.  The 
current EDM/2 (Electronic Developer's Magazine for OS/2) has a Rexx 
column on date manipulation.  There is code provided that will work 
on multiple platforms along with an OS/2 specific DLL. 
 
I tried to stay really close to the nature of Rexx when it was 
written so that it would be easy to use.  It can be found from 
http://www.iqpac.com 
 
joe r wyatt

In <4ksb9p$1eps@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>Today I was revising a program at work (OS2 REXX) and to my
>surprise I could not find a way to do date arithmetic. I found 
>the date("base") which returs a julian date, but I could not
>find anything that takes that format back into any other format.
>
>
>Did I mean something? I would think that date arithmetic should
>be something that one would want as part of the standard REXX.
It is if you write the code yourself....I found a program called fdater
that someone wrote and uploaded to compuserve.  It performs all date
arithmetic, and works fine (tested to 9999).  I will upload it here for you...
just add it in as a function (or procedure).

In <4l31s4$66j@globe.indirect.com>, dwright@indirect.com writes:
>In <4ksb9p$1eps@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>>Today I was revising a program at work (OS2 REXX) and to my
>>surprise I could not find a way to do date arithmetic. I found 
>>the date("base") which returs a julian date, but I could not
>>find anything that takes that format back into any other format.
>>
>>
>>Did I mean something? I would think that date arithmetic should
>>be something that one would want as part of the standard REXX.
>It is if you write the code yourself....I found a program called fdater
>that someone wrote and uploaded to compuserve.  It performs all date
>arithmetic, and works fine (tested to 9999).  I will upload it here for you...
>just add it in as a function (or procedure).

Let's put a wrap on this thread.
Date processing is the business of application code, just as you said.
Date processing is not the business of an operating system interface.
Date processing involves nationalities, religions, languages, boundaries, and 
treaties.
Date processing is going to cost 600 billion dollars to the U.S. of A., in programming 
effort alone, over the next six years.
Date processing, by my estimates, will cost business an additional TRILLION dollars 
before the final costs are tallied.

So get experienced in the nuances of the calendar, and look in the proper nooks 
and crannies for the proper solutions. 

Warren
http://www.borg.com/~warren

warren@borg.com wrote:
: In <4ksb9p$1eps@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
: >Today I was revising a program at work (OS2 REXX) and to my
: >surprise I could not find a way to do date arithmetic. I found 
: >the date("base") which returs a julian date, but I could not
: >find anything that takes that format back into any other format.

: Date processing involves algorithms and definitions of intention, somewhat like 
: litigants in a courtroom.  It does not define a language.  

: Go to a search engine and plug in the keyword variables you want.  Do a bit of 
: research,  and quit whining.

Geez Warren, why don't you just tell him to FOAD while you're at it? 
You've posted two messages on this subject. an neither one has anything
helpful to say. Maybe you should just keep whatever anger you have about
the subject of date processing to yourself, and work on lightening up a
bit. 

Anyhow, I'm the person who wrote the program mentioned by C. Vissengen, 
available at:

    ftp://hobbes.nmsu.edu/os2/dev32/rexx/rxdate20.zip

and would be happy to answer any questions about it.  As Warren mentioned,
date processing is complicated by many factors including religion,
language, country, etc.  But if you happen to be interested in working
with English US-style dates, then this package might do the trick.

    Barry

(If you're thinking of flaming me Warren, please go back and re-read your 
previous posts first...they _are_ pretty harsh)

Barry B Pederson (bapeders@plains.nodak.edu) wrote:

: Anyhow, I'm the person who wrote the program mentioned by C. Vissengen, 
: available at:

:     ftp://hobbes.nmsu.edu/os2/dev32/rexx/rxdate20.zip

Oops, that should have been: C. Vissenberg, sorry about that :)

    Barry

In <4l9677$bfq@daily-planet.nodak.edu>, bapeders@plains.nodak.edu (Barry B Pederson) writes:
>warren@borg.com wrote:>
>(If you're thinking of flaming me Warren, please go back and re-read your 
>previous posts first...they _are_ pretty harsh)
snips
>Barry

Oops.  Too many people dumping on REXX, doing too little research or study on 
their own, and too little time between now and 1999 December 31, Gregorian, 
Roman, Western, which is a year BEFORE the millenium.  Your procedures are 
easily found by reading commonly available documents, faq's, tips, etc., as well 
as by executing any of the readily avaiable search engines.  

The next such abusive request that comes through may get a FOAD!   Well, at 
least in my thoughts. My concern is that the evolution of REXX to Object-REXX 
will be contaminated by inferior demands on its structure.  

Warren
http://www.borg.com/~warren

In <4ksniq$9fn@ns2.borg.com>, warren@borg.com writes:
>> javier5@ibm.net writes:
>>Did I miss something? I would think that date arithmetic should
>>be something that one would want as part of the standard REXX.
>
>Date processing involves algorithms and definitions of intention, somewhat like 
>litigants in a courtroom.  It does not define a language.

I truly believe that the only people who define a language are those that
create it. Look at Xbase for instance. It has  date processing in it.

My reasoning behind having date arithmetic in REXX is the following. The
domain of the tasks best handled with REXX are very different than what
would be done with languages such as Pascal and C/C++. One task 
commonly performed in REXX are simple scheduling of tasks, and as a
means of easily calling other programs. Because of its ease of use it
is also used in simple programs (or even big ones) in which speed of
development is more important than raw performance. In any of the
mentioned cases date arithmetic would be something that would
substantially add to the language.

As it stands REXX already has a date function that returns different
formats including a julian format. Why it is so terrible to consider
having a function that takes a julian date and returns a string
in date format? You could use the same options as date, excluding
 the B option which is Julian, to make it easier on users.

In <4l9m6u$etd@ns2.borg.com>, warren@borg.com writes:
>The next such abusive request that comes through may get a FOAD!   Well, at 
>least in my thoughts. My concern is that the evolution of REXX to Object-REXX 
>will be contaminated by inferior demands on its structure.  

I started this thread on "Date arithmetic missing in REXX" and I don't
recall having asked for anything. My initial post says:

>> "I would think that date arithmetic should
>> be something that one would want as part of the standard REXX."

Did I request anything??

Maybe the "should" got there by mistake instead of a "would", but
other than that I didn't think that message was as bad as you have
made it seem.

Perhaps you have used REXX much longer than I have, but I have
every right to have an "opinion" on REXX. I didn't ask for a solution
to a problem. I am not trying to do a homework or just get someone to
send me a function over the newsgroup. As a matter of fact I got done
what I needed with date arithmetic before I posted the message
on the group. All I did was to post an opinion.

The commitee working in the standard REXX should be the ones
to decide what goes into the language. I don't think they would
take every little single request that people post and incorporate
it to the language, but if an overwhelming majority of the REXX
base asked for date arithmetic to be included should they say
no because YOU BELIEVE that it should not be included.

REXX together with Xbase are some of the most productive
languages I have used. I don't want it "contaminated by inferior
demands on its structure" as yourself. I just think that at least
having a function that would convert from Julian back to
Gregorian would be usefull in the language.

In <4lcnc7$29u6@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>In <4ksniq$9fn@ns2.borg.com>, warren@borg.com writes:
>>> javier5@ibm.net writes:
>>>Did I miss something? I would think that date arithmetic should
>>>be something that one would want as part of the standard REXX.
>>
>>Date processing involves algorithms and definitions of intention, somewhat like 
>>litigants in a courtroom.  It does not define a language.
>
>I truly believe that the only people who define a language are those that
>create it. Look at Xbase for instance. It has  date processing in it.
>
Hmm.  Should it treat 1900 as a leap year as did Lotus and dBase?  Even though 
that was a bug.  Should it limit the dates returned as does the B option?  Would 
the date values have concordance, i.e., be able to be converted from one form,
to another, and back again, yielding the same value?  Should it yield dates that 
did not exist?

The B date format is "ordinal" (counting from one), not Julian (year, then day of 
the year).


Warren
http://www.borg.com/~warren

In <4lcoei$29u6@news-s01.ny.us.ibm.net>, javier5@ibm.net writes:
>In <4l9m6u$etd@ns2.borg.com>, warren@borg.com writes:
>>The next such abusive request that comes through may get a FOAD!   Well, at 
>>least in my thoughts. My concern is that the evolution of REXX to Object-REXX 
>>will be contaminated by inferior demands on its structure.  
>
>I started this thread on "Date arithmetic missing in REXX" and I don't
>recall having asked for anything. My initial post says:
>
>>> "I would think that date arithmetic should
>>> be something that one would want as part of the standard REXX."
>snips
>
>REXX together with Xbase are some of the most productive
>languages I have used. I don't want it "contaminated by inferior
>demands on its structure" as yourself. I just think that at least
>having a function that would convert from Julian back to
>Gregorian would be usefull in the language.

What you are asking for, is a language extension.  They abound: 
math, editing, data base, sockets, and, yes "date conversion".  The 
flakiest of all, the least predictable, most problematic are the date 
conversion routines.

The simple problem of relating to legacy systems typically mandates that 
1900 be treated as a leap year because it was divisible by 4, for example.
They are flakey by the nature of the problem.  Hey, not even Cobol had 
date conversion.

Warren
http://www.borg.com/~warren