Talk:Sudoku/REXX: Difference between revisions

:: I named it (and a few others) by what I named it on my system, and I have an update system in place for updating/maintenance/verification/backups/etc (we're talking about more than just a few files);   also "uploaded" were similarly named REXX files (subroutines) which are called (invoked) by various REXX programs which are also "'''$'''"-named routines (programs).   [The reason why they almost all start with a dollar sign is a story that dates back to the very early 1970's VM/CMS if anyone wants to hear about it.]   The few REXX programs that don't having that naming convention are the few PC/REXX and Personal REXX extended BIFs (built-in functions) that I wrote for other REXXes that didn't support them: LINESIZE, SCRSIZE, DOSCD, DOSDISK, DOSDRIVE, DOSISDIR, SOUND, CHANGESTR, JUSTIFY, DELAY, and some others.]   It makes it much easier for me (and anybody who may take on that job later) to maintain and update those files if they have the same name everywhere.   Also, the HELp files mention the program (and how to invoke it) by name.   Ditto for the flow (FLOw) and sample (SAMple) documentation files --- which I haven't placed on Rosetta Code yet (and may not).   I've been ''thinking'' (with reservations) about also "uploading" additional (two) help-type files:   one is for a general flow-type visual diagram (each 21 lines only, one is included below, just for show-n-tell),   along with a sample-use documentation   (i.e.,   how to use/invoke the program with various options and whatnot), but I'll wait on that as I don't want to start down a road and then have to rename them all, update each of those pieces to refer to their changed names (or just remove all of 'em).   I don't want to have to maintain yet another time-consuming off-site system that has a different naming system and/or requirements.   I realize that Rosetta Code may not be the proper forum for large programs (its all relative, of course, of course);   programs that do much more that only address tasks with few (or limited) requirements.   '''$SPELL#'''   is one such REXX program, it goes beyond what is required, but then, I wrote that program long before Rosetta Code was even thought about, indeed, even before the internet (as I know it).   I'd like to do that with a few others (REXX programs), such as the Morse code solution, drawing a clock, and a couple of others.   A lot of the subroutines (which are programs in the truest sense of the word) and are very robust and more general-purpose than what is needed to solve the various task's requirements, but as I mentioned elsewhere, I'd rather show the whole program than ripping out the guts of a program (and testing it) and include that anemic version on Rosetta Code.   It would be much easier to just include the whole shebang in the first place, preserving the robustness and including other features and options, which, for the most part, would be too much to ask as general requirements, not to mention just too difficult.   Also, I don't want a special version of the various HELp (documentation) files just for Rosetta Code, updating them is a chore by itself; I wish I could upload the files directly instead of cut-n-paste (and doing that for a larger file is a pain in the neckhole and prone to mishaps).   A thought about ''child'' storage of programs:   I understand why the sudoku REXX program would/should/could/maybe be renamed to reflect the task name, but what about the various subroutines that it (and other REXX programs) invoke?   In addition, the subroutines are used elsewhere, so by themselves, they aren't children of any particular Rosetta Code task solving examples.   It's a lot easier to have other people take on that chore, but it's becoming a chore.   I'd rather spend time on solving tasks than making more work for myself. -- [[User:Gerard Schildberger|Gerard Schildberger]] 15:44, 31 January 2013 (UTC)

 

An example of the FLOw documentation   ('''$T.FLO''')   for the   '''$T'''   REXX program:

<pre>