Re: [Groff] Status of the portability work, and plans for the future

[Eric S. Raymond]

Jon Snader <address@hidden>:
> As a Vim user, I use the K command to pop up a man page in an
> editor window when I need to check the exact usage or parameters.
> I know that emacs users can do similar things with the M-x man
> command.  Most serious developers use either emacs or Vim, so I'd
> be willing to bet that they prefer to read most of their man
> pages this way. 

For me, it varies.  I sometimes use M-x man, and other times man(1) in
an xterm.  I think I'd use HTML presentation more often than either if
the things that should be hyperlinks really worked, and if
ht://dig-style text search were available.

> As a long-time reader of ESR, I know that he prefers to have a
> *single* window on his desktop with an emacs window and
> (presumably) a browser side by side.

Actually, it's usually emacs to the right and a big xterm to the left,
browser called up from the toolbar as needed and minmized afterwards.

Right at the moment I have a third 80x25 xterm open in root, it's
where I'm repeatedly doing "make install" to test groff man-page
modifications.

>                                     With this arrangement,
> having the man page render in the browser exacts no extra cost
> and has all the benefits that we've been discussing. 

Your thinking is intelligent and cogent -- but your factual premise is
wrong, leading you to an incorrect model of my assumptions.  On my
usual desktop arrangement, rendering man pages in a browser *would* in
fact have the cost of popping up a browser.  It is indicative that I
don't have one up right now.

>                   One could argue, I suppose, that this
> just shows that ESR is correct and we should stick to a single
> window, but there's a whole bunch of us who disagree.

Heh.  For the record, "ESR" himself wouldn't make that argument. :-)

>                      The important thing is that the behavior
> be configurable for each user,

Agreed.  I designed the changes to man(1) very carefully to achieve that.

>                             and that we don't dismiss a desire
> for the current behavior as indicative of arrested development.

You skipped a step.  You have a good point about calling up manual
pages within an editor, but not all character-cell displays are
equivalent; it doesn't follow from this that man(1) through xterm has
any value that lynx(1) through xterm wouldn't.  I'll be interested to
see if you can make that argument, especially since Gunnar ducked it.

Until somebody meets that challenge, I'll stick with my diagnosis of 
arrested development.  It's not one I make casually -- I've been thinking
*hard* about Unix UI, from a position deep within Unix culture, for
half a decade now.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>