[Top] [All Lists]

Re: Why the 822bis grammar is so painful

1999-02-10 07:26:24
At 01:54 AM 2/7/99 +0000, D. J. Bernstein wrote:
Pete Resnick, over the objections of several implementors on DRUMS,
threw away the RFC 822 tokenizer. He wrote a new ABNF grammar that

Just to clarify this particular point: It is not uncommon to have objections from one or more participants. So the fact of such dissension is not automatically significant.

What matters is rough consensus, meaning a sufficient base of support for a decision.

On the matter of removing the tokenizer approach, there was VERY strong working group consensus.

The tokenizer was a nice idea for RFC733, attempting to provide a nice, academically clean model. I thoroughly enjoyed the time I spent researching the topic while we were writing 733.

But it turned out not to work very well, because the mail header environment has too much context sensitivity. All that academic purity simply added confusion and, therefore, errors.

The DRUMS working group deliberated on this point quite carefully -- considering both retention and removal -- and the agreement to do the removal was quite strong.

I'll also add that it is not only counter-productive to put the issue in ad hominem terms, it is just plain wrong.



Dave Crocker                                       Tel: +60 (19) 3299 445
<mailto:dcrocker(_at_)brandenburg(_dot_)com>             Post Office Box 296, 
                                        Serdang, Selangor 43400  MALAYSIA
Brandenburg Consulting
<>                       Tel: +1 (408) 246 8253
Fax: +1(408)273 6464             675 Spruce Dr., Sunnyvale, CA 94086  USA