Re: [Groff] refer question

[Peter Schaffter]

On Sat, Apr 02, 2005, Jorgen Grahn wrote:
> On Fri Apr  1 23:42:51 2005, address@hidden wrote:
> > On 01-Apr-05 Peter Schaffter wrote:
> ...
> > > 
> > > In contemporary North American bibliographies, about five
> > > styles are "standard", though by no means the only ones used:
> > > 
> > >     AMA      (American Medical Association)
> > >     APA      (American Psychological Association)
> > >     MLA      (Modern Language Association)
> > >     Turabian (from Kate Turabian's _A Manual for Writers of Term Papers,
> > >               Theses, and Dissertations_)
> > >     Chicago  (The Chicago Manual of Style)

> > I understand that the MLA style is the general "norm" for
> > publications in the Humanities, to the extent that many
> > Humanities journals either strongly recommend it or insist
> > on it.
> 
> Humanities probably drives the evolution in this area; you refer
> to a lot of works and you need to do it correctly.

I don't have a preferred ref/cite style, or one I use most often.
Years in typesetting shops acquainted me with about a dozen styles,
eleven of which could probably have been eliminated.  And the one
left standing would have been neither better nor worse than any of
the others, functionally or typographically.

For refer support in mom, I'm using MLA for the reason Jorgen
gives, above.  Additionally

- its rules are simple: every field ends with a period and two
  spaces

- additional marks (colons, semi-colons, commas, parens) are kept to
  a minimum

- I have an aversion to Author/Date biblio styles :)
  
> > ...JAMA.  There are so many!  One of the principal merits of
> > specialised bibliography programs is that they can be configured
> > (using a syntax-like "style file") to adhere to a specified
> > standard.
> > 
> > This is one reason I've never much liked using {g|t}roffs
> > "refer". While it has the feature that you can set up a
> > database using standard tags, and refer to items in this
> > in various natural ways in your text, it is distinctly
> > inflexible when it comes to changing style.
> > 
> > This is not helped by the fact that the implementation
> > of the formatting is done by specialised macros in whatever
> > macro package you happen to be using.

To my surprise, I find refer *isn't* inflexible.  On the contrary,
precisely because formatting is set up in macros, you can design
any biblio style you like.  But that comes at a price: implementing
new styles is fussy.  Not difficult--not by a long shot--just time
consuming.

I think, if there's a problem with refer, it's the documentation.
There's very little of it other than the manpage, and the manpage is
a pure example of what can be both good and bad about manpages: it
contains all the information you need, but in a form so terse you
have to know the program before you can make sense of it.  Example:

   Macro interface
       Each  reference  starts  with  a call to the macro ]-.

The manpage doesn't say you have to create ]- yourself, which,
though self-explanatory once you grasp how refer works, would make
the grasping of it faster.

As I said in a previous post, I've considered putting several
styles in mom and letting users choose.  I'd prefer not to.  It
would a lot of finicky setting up and testing, and for what?  So
some poor user, quite possibly a groff newbie, can write the list
asking how to generate references of style XXX, which mom doesn't
include?

I'm thinking it makes more sense to have just one style, with a
section in the mom docs explaining how to set up others.  I have
to write that section anyway, regardless of whether mom has a
couple of styles to choose from (another reason, time-related, for
not offering a choice of styles).
 
Any thoughts on this?

-- 
Peter Schaffter
  Author of _The Schumann Proof_ (RendezVous Press, Canada)
  http://www.golden.net/~ptpi/theschumannproof.html