ietf-822
[Top] [All Lists]

Re: New Internet Draft: draft-duerst-archived-at-00.txt

2004-02-25 10:12:31

Bruce Lilly writes:

RFC 2017 provides for URIs in parameters. RFC 2557 uses URIs in a structured
field. 2557 syntax is broken because it allows comments (this has been
discussed on the MHTML list). RFC 2369 uses URIs in various List- (structured)
fields.  2369 uses a quoting mechanism which permits comments outside of the
quoted URIs.

RFC 2017 deals with URIs as parameters, insists that they be in the form
of quoted-strings, and then allows you to insert FWS at selected places,
e.g.:

                   URL="ftp://ftp.deepdirs.org/1/2/3/4/5/6/7/
                        8/9/10/11/12/13/14/15/16/17/18/20/21/
                        file.html"

The broken-ness of RFC 2557 seems to arise from the fact that it does not
insist that any parentheses within the URI get %-quoted. Also from the
fact that it tells you to use the RFC 2017 mechanism for long URIs,
whereas that mechanism only applies to parameters. Also, it makes life
hard for itself by not providing delimiters (whether "..." or <...>)
around the URI. But, worryingly, Martin's proposed syntax also suffers
from some of these problems.

RFC 2369 appears to set the best precedents, including delimiting within
<...>, allowing FWS within those <...> (to be eliminiated before use as a
URI), and provision for comma-separated lists of such URIs.

Since the List-Archive header and the proposed [List-]Archived-At header
are quite likley to occur in the same message, it would be a great pity if
they did not follow the same syntactic conventions.

And I would much prefer a comma-separated list of URIs than multiple
Archived-At headers. For a start, it gives you a way to express some
priority between the various alternatives (as in the List-Archive header).
And then, multiple headers with the same name are Bad Thing for other
reasons (for example if ever you come to have digital signatures of
headers, such as have been mentioned recently in other threads on this
list).


Whitespace and CRLF are not permitted in URIs.  So folding isn't a problem.
See RFC 2396 Appendix E.

Not quite so.

<http://gbiv.com/protocols/uri/rev-2002/rfc2396bis.html>

seems to recommend, in its Appendix C, that you should "just put spaces"
into your URI (all genuine SP having already been replaced by %20) and
then fold, and it especially recommends delimiting within <...> at the
same time. So if that is the way the URI people are thinking, then that is
the way we should be going.

One wishes, however, that the provision had been made in the body of their
draft, rather than as an afterthought in an Appendix.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: chl(_at_)clerew(_dot_)man(_dot_)ac(_dot_)uk      Snail: 5 Clerewood Ave, 
CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5