Anonymous user
Rosetta Code:Village Pump/Dialects: Difference between revisions
added comments on classic REXX versus object-orientated REXXes. -- ~~~~
Walterpachl (talk | contribs) (20 more Rexx programs analyzed and tested. ONE real problem so far) |
(added comments on classic REXX versus object-orientated REXXes. -- ~~~~) |
||
Line 213:
Whether there are objects inside I don't care. I personally don't use those on the ooutside. Too old for that.
--[[User:Walterpachl|Walterpachl]] 02:34, 10 July 2012 (UTC)
-----▼
Walter, I wish you would refrain from stating what I don't love to do, or worse yet, assuming the reasons why I don't conform to your philosophy (based on your statements).
The major hangups (as you see them) are yours, not mine. My reasons of why I don't dance to your tune are my own.
I will ''never'' make as assertion that any of the classic REXX example programs (that I entered on Rosetta Code or elsewhere) will work on ooRexx (or whatever REXX, not just the object-oriented REXXes) if I haven't tested them first on those languages. Some, I'm sure, would work on AREXX or BREXX or OS/2 REXXes or CMS REXX or TSO REXX ..., but I don't have access to those (at least, not directly, and I certainly don't want to ask others to execute my versions on their computer, or worse yet, their employer's computer). I don't have an interest on what does or what doesn't run on object-oriented REXXes; I have no interest in those languages, and I don't have the time nor inclination to install them for whatever reason. -- [[User:Gerard Schildberger|Gerard Schildberger]] 20:37, 10 July 2012 (UTC)
Just because ooRexx and classic REXX were developed in the IBM family, doesn't make one a grandchild of the other. QBE was developed by IBM. So were a lot of other languages ... -- [[User:Gerard Schildberger|Gerard Schildberger]] 20:37, 10 July 2012 (UTC)
: The next 20 tasks
Line 218 ⟶ 228:
:: and one (Brownian tree) uses @ and a= and scrsize() which is not available in ooRexx.
--[[User:Walterpachl|Walterpachl]] 19:59, 10 July 2012 (UTC)
▲----
-----
Yes, the REXX code from the classic REXX example in the ''Brownian tree'' task is:
<lang rexx>if height==0 | width==0 then _=scrsize() /*not all REXXes have SCRSIZE.*/</lang>
which clearly states what you've stated. -- [[User:Gerard Schildberger|Gerard Schildberger]] 20:37, 10 July 2012 (UTC)
-----
An interesting case is under Rexx Roman numerals/Decode
: where Version 1 (good for all Rexxes)
Line 224 ⟶ 240:
but no longer good for ooRexx (due to the use of # and @
--[[User:Walterpachl|Walterpachl]] 08:30, 10 July 2012 (UTC)
-----
Er, no. The example version 2 for REXX was a straight copy of version 1's functionality (but with the superflous statements and dead code removed), and as such, mimicked the inability to handle the Roman numberal '''IIXX''' (and others of this type). This means that version 2 has the same functionality as version 1. I had added a comment to that effect (about the inability concerning version 1 and its cousin, version 2), but it was deleted by someone else.
This also implies that the NetRexx version has the same problem, as the REXX version (as stated) was taken/copied/transcribed from NetRexx), but I can't verify if the NetRexx version as I don't have access to the NetRexx language. Most likely, other examples have the same inability.
REXX version 1 ''may'' be good for all REXXes (which I'm not claiming, as I don't have access to all REXXes to test that claim, and I would be very surprised if anyone still does have access to all REXXes), but version 1 and version 2 still lack the ability to handle REXX numerals of the aforementioned type. I don't know if that makes it good or not. As I mentioned elsewhere, the Roman word for '''18''' is '''duodeviginti''' which translated literally means '''two-from-twenty''', and there exists a photograph(s) of '''IIXX''' chiseled on an archway. I also seen '''XIIX''', which I admit, looks very symetrical. That numeral may not be of the form that was approved of in the years after when the format/form of Roman numerals were codified more restrictively. Even the use of '''IIII''' wasn't approved of, but yet, wrist watches and clocks that use Roman numerals all have that form (I haven't looked at ''all'' watches or clocks, by the way, just a mere half-dozen or so).
Proof (tongue-in-cheek):
* 3 is prime,
* 5 is prime,
* 7 is prime,
* 9 isn't prime (a square),
* 11 is prime,
* 13 is prime,
* 15 isn't prime (ends in 5),
* 17 is prime,
* 19 is prime,
* ...
* all odd numbers are prime.
I'm not saying that those "unapproved" forms of Roman numerals should be used, but they ''have'' been used, and I think they should be converted correctely (to Arabic numerals).
Also, iterating again, I entered the classic REXX code to solve the task, using the language that I know best, and I never claimed it would work on an object-oriented version (such as ooRexx or NetRexx). Even if it would work elsewhere, I don't have the means or inclination (nor the time) to test if a classic REXX program will work on (any or all) object-oriented versions of REXX, or even other platforms, for that matter. -- [[User:Gerard Schildberger|Gerard Schildberger]] 20:37, 10 July 2012 (UTC)
-----
| |||