[Top] [All Lists]

RE: Alternative formats for IDs

2006-01-02 15:22:33
Lets go ahead and ask then -

  Does anyone else think that IETF should allow documents which
  format/structure is not publicly known as one of the ways to
  distribute IETF specifications?

For the record, my answer is "absolutely not".

And why do all the other SDOs get along with non-ASCII formats?
On my intranet I have a list of 120+ SDOs in the communications
and computer-science fields, and although I haven't gone through
them all (I have asked someone to do so) I haven't found another
group that uses ASCII files.

First off, there's a big difference between using, say, an appropriately
constrained subset of HTML and using MS Word. Both may be non-ASCII, but one is
publicly specified whereas the other is not.

Second, your assumption that other SDOs have been able to blissfully make use
of private formats like MS Word without incident is simply untrue. One obvious
counterexample I know of is the CCITT/ITU, which has in the past used MS Word
as a distribution format for many of it's documents. I have quite a few of
these documents on hand and occasionally need to refer to old versions of them,
but when I try and read them using modern tools the results are rarely good.
Many of these documents simply refuse to open, sometimes crashing the tool I'm
using, while others do open but are misformatted, sometimes to the point of
being illegible.

Like it or not, compatibility with future versions is just not something you
can count on. Of course this is also true with formats whose specifications are
public, but at least with those you have the option of writing something to the
specification that will read the stuff.

    > the state of online collaboration and editing that we have been at for
    > 20 or 30 years.
    > Finally, there is a longstanding and more or less explicit decision in
    > the IETF community to keep the costs of participation as low as possible

Quite true. I don't like thinking about how much $$$ I've spent trying to deal
with proprietary formats of one sort or another.

There's one other thing, also tied to the IETF (and its predecessor's) long
existence, which is the long-term accessability of online documents - again,
another facet in which our experience is pretty unique. Is MS-Word (or
anything else) going to be 30 years from now?

Not if the past is any indicator.

In case you think this is a silly question, I just recently finished scanning
/ OCR'ing / proofing the oft-cited IEN-19 (Shoch, "Inter-Network Naming,
Addressing, and Routing"), from January 1978 - 28 years ago.

Access to old versions of specifications is often quite important. Even if you
ignore the frequent direct use of such specifications, it is not at all
uncommon for a modern RFC to have normative references to other RFCs written 25
years or more ago.

It is also worth noting that access to old drafts can be quite useful. And this
isn't confined to the IETF. To use the CCITT/ITU as an example again, back in
the day there was a widely used X.400 body part format that was only ever
specified in a particular draft of ISO 10021. (It never appears in any ITU
version of the specification, draft or otherwise.) You either had to have a
copy of this draft or you were reduced to dumping the ASN.1 and guessing what
the various fields were.

The original of
this document was presumably in some Bravo format, and the printing version
was in PRESS - and I somehow doubt either is supported anywhere in the world
now. I only had a hardcopy, so the question's a bit moot, but I very much
doubt a machine-readable version of either form would have done me much good.

Well, I don't know about PRESS specifically, but I have a book describing the
Interpress format in a fair amount of detail. If it is the same it might
actually be possible to process it, probably most easily by converting it to
PostScript (which in many ways is it's descendant). So once again, it's all
about whether or not the details of the format are public.

ASCII may be pretty lobotomized, but it *is* timeless.

Or as close to it as we can reasonably get.

(Not that I'm per-se against allowing more powerful forms, mind, but any
proprietary option is just not viable, IMO.)

I completely agree.


Ietf mailing list

<Prev in Thread] Current Thread [Next in Thread>