Hi! On 9/6/22 18:23, Ingo Schwarze wrote:
Hi Alejandro & Branden,G. Branden Robinson wrote on Tue, Sep 06, 2022 at 07:49:58AM -0500:At 2022-09-06T13:37:39+0200, Alejandro Colomar wrote:I was wondering if tbl(1) wouldn't be better split into tbl(1) and groff_tbl(7)[...]I'd like to be able to refer to tbl(7) as a language when talking about it as a language.That's a reasonable request. I think Ingo already does this in mandoc.Not quite. The mandoc package does provide a tbl(7) language manual, but it does not provide a tbl(1) command line command, and consequently no tbl(1) manual page either. [...]And, I think it also makes sense to separate documentation about the command and its options from documentation about the language.The GNU tbl(1) command has no options, which is actually laudable. Well, there is the rather obscure -C option. I say obscure because even formatting the oldest manual pages that ever used tbl(7), those in PWB and v7, does not need -C. I think if a command has no options except a compatibility option that is almost never needed even in extreme cases where one might expect that they might possibly require compatibility, it is fair to say that for practical purposes, it has no option. Doug is right that a manual for only the command line command would be almost empty.
Then, it might make sense to have tbl(7) with no tbl(1). But, of course, every command needs a manual page, even if it's just to know that the command is as simple as it can be. I think I prefer having both tbl(1) and tbl(7), even if tbl(1) is very short.
We have for example nologin(1) in the Linux man-pages. It's 16 lines of source code, and 25 lines of formatted output. I wish more commands had only that problem.
Cheers, Alex -- <http://www.alejandro-colomar.es/>
OpenPGP_signature
Description: OpenPGP digital signature
