[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: if SGML is so great...



> -----Original Message-----
> From: Kevin Turner [mailto:Kevin.Turner@oberlin.edu]
> Sent: Wednesday, May 03, 2000 12:39 AM
> To: ldp-discuss@lists.linuxdoc.org
> Subject: Re: if SGML is so great...
> 
> On Wed, May 03, 2000 at 02:58:23AM -0400, Gary Lawrence Murphy wrote:
> > >>>>> "K" == Kevin Turner <Kevin.Turner@oberlin.edu> writes:
> > 
> >     K> The last time someone asked for help with this on the Open
> >     K> Source Writer's Group list [1] (I think it was for the
> >     K> "Replacing NT with Linux HOWTO"), no less than four 
> volunteers
> >     K> signed up to help with the transition.
> > 
> > And what happens with the updates?  As I've said before, the Linux
> > kernel will welcome a really clever hack in VBasic, but they won't
> > tolerate continual updates in that form (unless it is blazingly
> > brilliant)
> > 
> > I think the LDP needs to be exactly the same.
> 
> Do you?  
> 
> I guess I haven't paid any attention to the submission queue, 
> but it was
> not my impression that you had so many good authors that you were at
> leisure to pick and choose among them.  Rather, my impression of the
> atmosphere was that the free software as a community as a whole needs
> more documentation desperately, and so I would think you would welcome
> good content in whatever form you could get it.
> 
> If, on the other hand, you believe "Most of what we want has already
> been written, it's just so poorly organized that we need structure to
> find it," *then* you have reason to be picky about the syntax, rather
> than the content.
> 
> Could this bit of matinance be viewed as an acceptible cost?  
> Many other
> roles in software development have seen similar divisions of labour.
> When the programmer of FooApp releases the new version, the package
> maintainer curses and swears that he's changed the pathnames 
> again, and
> the translator groans at all the little changes in punctuation to the
> strings he has to re-translate, but the three of them end up 
> with a much
> more useful package than the programmer had when he started, and it's
> certainly nothing the package maintainer could have done on his own.

It's hard to say if it would be an acceptable cost.  For example, if they're
all writing like me, and not keeping track of a revision history in any
meaningful way, then it's going to require completely starting over on the
markup.  If, instead, they keep track of which sections they've made changes
to, and what kinds of changes they've made there, then keeping the document
in a useable SGML format shouldn't be too tough.  Perhaps strongly
encouraging authors to use SGML for their updates (I've had good resposnses
from the authors of 2 mini-HOWTOs that I've contacted about this) if we can
get the document into SGML for a first revision.  We will, of course, have
"sticks-in-the-mud" who refuse to use SGML, and for them we can only ask
that they keep a good record of the changes that they make, so that our job
is easier.  But please keep in mind that going from anything other than a
structured SGML language (such as plain text or HTML) to a DocBook type
language cannot be automatic, and that I'm not proposing that we make it
automatic, just that we get it done.

> And seriously, someone who writes plain-text doing in-place edits on a
> docbook file is probably going to do very little damage.  
> Unlike putting
> VB code in the kernel, it'll probably still compile, and even be
> readable at that.

Aye, that one is probably correct, especially if we can teach them to put
some sort of a tag around that section that says that it hasn't been
properly DocBookized yet, or that changes have been made, and it needs to be
looked at for re-DocBookizing.  (I love making up words).  Thanks,
	Greg


--  
To UNSUBSCRIBE, email to ldp-discuss-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org