--On Tuesday, 03 January, 2006 06:47 +0200 Yaakov Stein
<yaakov_s@xxxxxxx> wrote:
> The downside is that when a group is working on a document
> in Word, anyone not having the SW would not be able to
> directly contribute - but joint work is not really practical
> using any system without tracking anyway.
That is an interesting observation. It must be true since you
say it is.
However, I wonder what it means that I have my name, as author
or editor, on, by rough count, 27 RFCs. And I've been a major
contributor (of text, not just ideas, for a few more). Of
those, all but about three were collaborative, with multiple
people contributing text. And, of that group, the number that
started in Word, including one of those non-listed-editor
collaborations, was two.
I guess the other 25+ documents must not have been practical to
produce.
At one level, I'm sympathetic to what you are saying and trying
to accomplish. While there are many characteristics of Word
that I dislike, I like its tracking facilities and ability to
turn tracking displays and marginal comments on and off. Used
properly, I find them much more satisfactory than anything I've
been able to do with CVS-like and Diff-like systems: the latter
are better at identifying what has been changed (although that
can be effectively simulated in Word by letting it generate
change tracking by comparing two documents) but far worse at
recording and identifying the reasons why the changes were made.
Unfortunately, even if one ignores most of the issues with
proprietary format and costs, it is hopeless for IETF use in
managing working documents and RFCs. Among the problems:
(1) Its template mechanism is very version-sensitive and
fairly fragile. If one makes template or option changes
to accommodate IETF needs, one cannot then go back and
forth with the formatting requirements of "day job"
documents without risking making a royal, and
essentially irreversible, mess.
(2) Its Style model is even more version-sensitive and
even more fragile. It is also badly documented and
idiosyncratic. But it, or major template changes, or
both, are necessary if one is going to produce IETF
documents that correspond to RFC Editor norms (and I'm
not just talking about ASCII output).
(3) Its cross-reference model is so "smart" in terms of
what can be referenced, what header styles it can be
used with, etc., that cross-references in RFC Style and
within RFC objects are not consistently possible. The
facilities of WordPerfect 4.2 in this area, nearly two
decades ago, were far superior, I believe because of
less of an attitude of "we know what you need and, if we
don't supply it, you don't need it".
(4) Others have pointed out the versioning problem in
terms of reading documents, but it is worse than that.
I've seen document formatting and structure destroyed
beyond recognition or recovery by the simple mechanism
of being moved back and forth among authors who are
using different version of Word, even when Word 2000 is
the oldest of them. I know how to mitigate that
problem but it would require far more drastic changes in
how the IETF does business than merely switching working
document formats.
(5) Other than change-tracking, Word has very little
built-in collaboration machinery. I have the impression
that there are even fewer such facilities in recent
versions than in earlier ones. Instead, the strong
collaboration facilities require even more expensive
versions of Office, use of Outlook for email, and other
client and server conventions that would be far more
problematic for the IETF.
(6) While I had high hopes for the XML output from Word
(again, available only with Office Professional and
above, if there is an "above), the 2003 version turns
out to be one of the stranger things I have every seen.
The XML output from Office 12 is supposed to be much
better -- I haven't seen it-- but it is, again,
incompatible.
That is by no means my entire list, but it is indicative. And
others may have other lists or entries.
For all of those reasons and others -- and I am still concerned
about costs and share the concerns of others about proprietary
and changing formats -- actually IETF use of Word seems to be to
be a non-starter.
I do believe that, if you want to do initial document
preparation in Word, you should be able to do that. As others
have suggested, no one I know of is really interested in
standardizing on or requiring a particular editor. But, to do
so, you need to be able to produce an editable format that is no
worse than ASCII. You may have better ideas, but, as I have
explored that range of options, I've come to the conclusion that
there ought to be two ways to accomplish that end. They are:
(1) Development of an "IETF printer driver" that can be
distributed as freeware or with minimal costs and
restrictions and that would produce lines and pages of
the right layouts _and_ would handle "smart character"
to ASCII conversions, generation of appropriate
line-ending sequences, etc. Whomever developed this
thing would need to make a long-term commitment to
producing and maintaining versions for every version of
Windows from, I think, Win98 through the indefinite
future. The generic printer and the conventions of RFC
3285 are demonstrably not good enough.
(2) Development of a converter between the MS-XML output
of Word Pro 2003 and the XML input of RFC 2629bis so
that xml2rfc and its friends could take responsibility
for final formatting. Note that, if the converter were
two-way, you could edit happily in Word and others could
edit happily in XML and both could interwork. However,
as with the above, I think this solution would rapidly
deteriorate into uselessness unless there were a
commitment to produce new versions as new versions of
Office appeared -- at least until Microsoft stabilizes
and documents their XML formats.
To the extent to which the trend toward a requirement --not just
an option-- for XML input to the RFC Editor process --
continues, the first of these options might easily become
unavailable.
john
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf