[rfc-i] LaTeX proposal misunderstood

Phillip Hallam-Baker hallam at gmail.com
Mon Apr 2 05:07:56 PDT 2012

I find this concern for a single version of the RFC being the
authoritative one to be completely bogus.

Almost none of the specs produced by the IETF are even STANDARD status
for a start. And even if they were, the idea is that the real standard
is set by the running and deployed code. When we have had problems
caused by widely deployed code being incorrect with respect to the
specification we have almost invariably fixed the spec to match the

I cannot remember a single case where a bug in running code has
resulted in an ambiguity that arises from a difference between an HTML
and plaintext version of an RFC. So just what does it mean when people
say the plaintext version is the authoritative one?

I do not read or review the plaintext versions, not my own, not other
people's. If you want to see the text I was working on when I was
editing you need to read the HTML version.

The stuff about plaintext is just people demanding I kiss their ring
which is why it is never happening.

It is really simple, the only things that really care about document
formats are the tools that process them. It makes a lot of sense to
have a single canonical format for the drafts and RFCs at the point of
publication so that other tools can hook into them.

There are three uses that tools make of documents, presentation,
authoring, extraction.

There are only two widely supported presentation formats, HTML and
PDF/A. Only HTML works well as an authoring format. XML is not
anywhere near as well supported as an authoring format. It is possible
to recover text from PDF/A but the vast majority of tools to do that
loose any semantic markup.

Extraction of text or code from a document is something that requires
a semantic markup which again limits us to XML based formats.

On Mon, Apr 2, 2012 at 5:24 AM, "Martin J. Dürst"
<duerst at it.aoyama.ac.jp> wrote:
> Hello Yaakov,
> Sorry to be late in replying. I haven't listened to your talk, but I have
> read all the followup to this posting. I also want to make it clear that I
> in no way dislike (La)TeX; I use it myself and have my students use it.
> I just want to make three comments:
> 1) Various versions of (La)TeX can handle various ranges of
> "internationalized" characters, but it's one of the technologies that is
> moving to Unicode at a rather slow pace. As far as I am aware of, there's
> not yet a version of (La)TeX where I can input a wide range of Unicode text
> and just get it formatted and printed well. We would also have to include
> various fonts with the source to make sure the output is reproducible. [If
> you know better, pointers would be appreciated.]
> 2) Mathematical formulae are definitely easier to hand-author in (La)TeX
> than in MathML. But in a different thread, it was pointed out that using
> complicated formulae is only necessary/important for a rather small
> percentage of RFCs. Also, there are good (La)TeX -> MathML converters.
> 3) On a higher level, you write "there would be a suite of converters" and
> "Perhaps someone would .. develop .. tools". Lot's of woulds. Well, for
> XML2RFC, some people *did* develop tools. It's running code, which is one of
> the tenets of the IETF. So if you'd develop some tools to produce IDs and
> RFCs from (La)TeX, then I think nobody would oppose that. Eventually, there
> might be enough users that it could be adopted as the "canonical" format.
> But even XML2RFC, which is widely used and has something like a 10-year
> history, isn't yet the "canonical" format, so I'd expect that to take some
> time.
> Regards,   Martin.
> On 2012/03/28 15:38, Yaakov Stein wrote:
>> Hi all,
>> After hearing from several people after the BOF yesterday,
>> I believe that perhaps I was not clear enough.
>> What I suggested was to move from XML to LaTeX as input format,
>> and to make the LaTeX source code the normative (and everlasting) version.
>> There would be a suite of converters (most of which exist today and run on
>> every conceivable platform)
>> that convert on-the-fly LaTeX source into HTML, PDF, plain text, epub,
>> etc. for viewing and/or printing.
>> By moving from XML to LaTeX we automatically get beautiful equations,
>> various graphics, international characters, metadata, etc.
>> It is also MUCH easier and more intuitive to use than the present XML,
>> produces much nicer documents,
>> and there are dozens of (introductory to advanced) books describing its
>> use.
>> The only "special" thing we would need to maintain is an RFC style file
>> (with the defaults, boiler plate, headers/footers, etc.).
>> We would not need to continue coding tcl or xslt for our own proprietary
>> typesetting formats
>> as everything else is taken care of by the large and mostly Open Source
>> TeX ecosystem.
>> Since TeX is used by all major publication houses and academic journals,
>> this ecosystem is not going away any time soon.
>> Perhaps someone would be able to develop semi-automated tools to convert
>> old text
>> or xml documents to the new format.
>> Although a nontrivial job, that is no harder than creating a text to xml
>> or text to epub converter.
>> Y(J)S
>> _______________________________________________
>> rfc-interest mailing list
>> rfc-interest at rfc-editor.org
>> https://www.rfc-editor.org/mailman/listinfo/rfc-interest
> _______________________________________________
> rfc-interest mailing list
> rfc-interest at rfc-editor.org
> https://www.rfc-editor.org/mailman/listinfo/rfc-interest

Website: http://hallambaker.com/

More information about the rfc-interest mailing list