I don't know who should be the more embarrassed by a finding I've
made as I've been working on the necessary changes to MIME to ensure
SGML compatibility: I, who missed a one-word change, or the authors of
the draft, who made it, and then forgot about it.
The present draft, draft-ietf-822ext-messagebodies-03.txt, which I
had not read through completely when I raised the issue of nested
comments, does in fact specify that comments nest. This must come as
a surprise to a lot of people, including the authors.
In message <sdOUctK0M2YtA_0OoM(_at_)thumper(_dot_)bellcore(_dot_)com> of
1992-01-07
15:14:19 -0500, Nathaniel Borenstein
<nsb(_at_)thumper(_dot_)bellcore(_dot_)com> writes:
|
| The new draft is very much like the last one -- the changes are
| all textual and not substantive, with the exception of the
| "local-file" access-type for message/external-body. You can
| retreive the full document in the usual way, anonymous ftp from
| thumper.bellcore.com, in pub/nsb/BodyFormats.{ps,txt,ez}. Or you
| can wait for the internet draft publication.
|
| Here is a summary of the changes: [...]
|
| 7. Fixed the richtext-to-text translator to handle nested
| comments properly.
In message <AdYecem0M2Yt8Y_kop(_at_)thumper(_dot_)bellcore(_dot_)com> of
1992-02-07
10:47:54 -0500, Nathaniel Borenstein
<nsb(_at_)thumper(_dot_)bellcore(_dot_)com> writes:
|
| The comment issue is a real red herring. This has been argued
| over and over in programming language circles. The current
| richtext comment syntax is a lot like C comments. You want
| something more like -- well, I can't think offhand of any
| programming languages in which comments nest, actually, although
| I'm sure there must be some. But the point is, the model is not
| so weird. Big chunks of the world program in it all the time.
These messages are almost exactly one month apart, and clearly
self-contradictory.
The paragraph which made me see red in the second draft was this
one in 7.1.2 The Text/richtext subtype
Richtext conformance: A minimal richtext implementation is one
that simply converts "<lt>" to "<", converts CRLFs to SPACE,
converts <nl> to a newline according to local newline convention,
removes everything between a <comment> command and the next
| following </comment> command, and removes all other formatting
commands (all text enclosed in angle brackets).
This has been changed to the following without any notification, and
the authors don't even remember they did any changes to it:
Richtext conformance: A minimal richtext implementation is one
that simply converts "<lt>" to "<", converts CRLFs to SPACE,
converts <nl> to a newline according to local newline convention,
removes everything between a <comment> command and the next
| balancing </comment> command, and removes all other formatting
commands (all text enclosed in angle brackets).
The word "following" on the line with the change bar (which is *not*
present in the drafts!) has been replaced with the word "balancing".
This is a very important modification, much more important than the
authors knew, or cared. I'm not impressed.
I noticed this quiet modification when I looked at the code
example in Appendix D. A one-word change like that is damn easy to
miss, yet it has such an enormous impact on the language.
What _really_ amazes me about this, was that NOT ONE SINGLE READER
of the draft noticed that this had changed. We've spent something
like a week fighting over it, and NO ONE REMEMBERED it had changed,
including those responsible for the change. _Totally_ amazing!
Let me at this point, partly because I feel a little embarrassed
at not knowing about this, partly because I'm pissed at the authors,
say just one thing about drafts: I *hate* modifications to documents
which do not include change-bars, and aren't diff'able. By inference,
I *hate* to have to reformat documents so they can be worked with.
(For those who wish to work with the present draft, a plain-text
version, *much* easier to work with than some half-formatted document,
can be found in /pub/tmp/draft-ietf-822ext-messagebodies-03.txt at
ftp.ifi.uio.no. It contains the same text as the distributed version,
only it's readable and editable in an 80-column editor, as opposed to
a damn print-out. Page breaks have been eliminated to ease working
with it.)
An impression I've had since I read the first draft has been
strongly reinforced: This stuff is not standards material. No wonder
nobody have cared to scrutinize it. God knows how many other quiet
changes with massive impact have sneaked into this third draft.
The current draft still has problems with the comment "element,"
since it ignores anything found within it, including other unmatched
tags, and that's not good. I'm still working on the mods.
Good night,
</Erik>
--
Erik Naggum | +47-295-0313 | ISO 8879 SGML | Memento,
Naggum Software | | DIS 10744 HyTime | terrigena.
Boks 1570, Vika | <erik(_at_)naggum(_dot_)no> | JTC 1/SC 18/WG 8 |
Memento,
0118 OSLO, NORWAY | <enag(_at_)ifi(_dot_)uio(_dot_)no> | SGML UG SIGhyper |
vita brevis.