X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil]
	[nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil])
Date:         Fri, 4 May 90 11:30:19 GMT
Reply-To: LaTeX-L Mailing list <LATEX-L%DHDURZ1@unido.bitnet>
From: CA_ROWLEY@VAX.ACS.OPEN.AC.UK
Subject:      Forwarded comment: on upward compatibility
To: Rainer Schoepf <BK4@DHDURZ1.uucp>
Status: R
X-Status: 
X-Keywords:                  
X-UID: 98

The message below was sent to me by David Rhead.

He has also sent me a file of suggested enhancements, each
backed up by evidence from such sources as "The Chicago Manual of
Style" and ISO Standards documents.  Thus I cannot submit it "over the
ether" (even that part of it which is also available as a LaTeX
document would get corrupted leaving the UK).


chris

----------------------------------------------------------------------------
Orginally from: David_Rhead@uk.ac.nottingham.ccc.vme


                              BACKWARDS COMPATIBILITY

I see that, in TUGboat vol. 10 (p. 689), Frank Mittelbach says (of LaTeX 3.0)
that "it would be possible to process any existing document with the
addition of a style option".

While compatibility is always nice, I would like to suggest that priority
should be given to getting things right (or more right than LaTeX 2.09 did).
I think that it would be better for sites to run LaTeX 2.10 and LaTeX 3.0
in parallel for 12 months (while LaTeX 2.10 people finish off their theses,
etc.) than for LaTeX 3.0 to not "get things right".

It might be worth considering analogues of the transition aids that Fortran
committees/vendors supply:
*  a "deprecated features module", to allow LaTeX 2.10 constructs (e.g.
   \chapter*) in LaTeX 3.0, while giving notice that there are better ways
   to achieve the same effects (e.g. prelimpages and endpages environments)
   and that, if there is ever a LaTeX 4.0, the deprecated features will
   probably be absent
*  something analogous to the VMS Fortran's /NOF77, to make LaTeX 3.0 act
   like LaTeX 2.10.

However, I shouldn't let the question of backwards compatibility inhibit
you from "getting things right".  We already deal with new versions of
other packages e.g. SPSS that phase out "the old way of doing things" by,
if necessary, having both old and new versions available.  Sites have
already dealt with incompatible BibTeX 0.98 and 0.99.  We could,
if necessary, do the same with LaTeX 2.10 and 3.0.

                                 ---------------

                                                                        David