Re: [Groff] Correct protocol for making changes

[Eric S. Raymond]

Werner LEMBERG <address@hidden>:
> Some other observations:
> 
>   . The proper way to write an ellipsis is `.\|.\|.\&', optionally
>     starting with `\&'.  Please don't omit the `\|' -- it looks quite
>     ugly in PostScript output if the dots don't have enough horizontal
>     separation.

OK.  Do you think it's worth adding an ellipsis definition to an-ext.tmac? 
 
>   . Don't use real tabs in tables; use the `tab' keyword to substitute
>     them with, say, `@'.

Can do.  Is there any technical reason for this, other than the
"future editor settings could silently mess you up" I'm already aware
of?

>     BTW, trailing and leading whitespace in tbl fields *do* matter and
>     should be avoided in general.  You need the `nospaces' keyword (a
>     GNU extension) to make tbl ignore them.

And we need to avoid GNU extensions in this context.  I hear you.

But I believe you may be worrying a bit too much here.  I think trailing
whitespace, which I inserted at some points to make the table sources
more readable, could only mess up the rendering onder very strange 
circumstances -- fonts scaled up too large and tabs set in a unit that 
doesn't scale with font size.

Or do you know of some subtle trap that I don't?

>   . If you use a table within a man page, the first line should be
> 
>       .\" t
> 
>     Similar key letters exist for refer and eqn.  This is documented
>     in the `groff_man' man page.

I was aware of this -- but last night when I tried to be a good
citizen and applied this first line to chem.man, it actually *broke*
the table rendering.  I thought I'd fooed up something else, and spent
about fifteen puzzled minutes before I got it through my head that the
effect of .\" t was the reverse of what I expected and all I had to do
was remove it to get my table back (!).

One of my cleanup tasks for after I get the big stuff done is to
figure out why this happened.

(BTW, the list markup that I replaced with a table was a truly classic example
of the-author-should-be-hung-by-his-thumbs .TP abuse.)

>   . I've further refined chem.man, using .SY/.YS within the man page
>     also (this is, outside of the synopsis section) -- they are quite
>     handy.  Do you see problems if I do so?

No, not at all.
 
>   . It's better to say
> 
>       .B "foo bar baz"
> 
>     instead of
> 
>       .B foo bar baz
> 
>     Reasons:
> 
>       (a) it's processed faster (no issue today, but...)
>       (b) it works with old troff also (which has a limit of
>           9 macro arguments)

OK.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>