W3C home > Mailing lists > Public > uri@w3.org > November 2004

small edit Re: draft-hoffman-news-nntp-uri-02.txt

From: Al Gilman <Alfred.S.Gilman@IEEE.org>
Date: Fri, 5 Nov 2004 19:15:20 -0500
Message-Id: <p06110401bdb1c5527c5b@[10.0.1.2]>
To: uri@w3.org

Small edit:

Where it says

>
>    Observe, in contradistinction to the news scheme, that the <news-server>
>    is not optional here, because the mapping from <article-numbers> to
>    actual articles is unique to each server.
>

[in lieu of
"is unique to each server."
substitute:]

varies from server to server.

... or ..

is established independently by each server.

[why?]

'unique' is stronger and not guaranteed.

Al

At 1:44 PM +0000 11/5/04, Charles Lindsey wrote:
>This draft seems pretty good as regards the news scheme, so my 
>quibbles are just for tidying things up.
>
>2.  Scheme Definition
>
>Perhaps that should be "The news URL scheme", and keep Section 3 for 
>the nntp scheme.
>
>    The news and nntp URL schemes are used to refer to either news groups
>    or individual articles of USENET news, as specified in RFC 1036.
>
>    The news URL takes the form:
>
>       newsURL     = "news"  ":" [ news-server ]
>                        ( newsgroup-name | '*' | message-id )
>       news-server =  "//" server "/"
>       message-id  = id-left "@" id-right
>
>    <server> is defined in [2396bis].  Note that there is widespread use
>    of the username-password authentication in news servers, so news
>    clients should implement this part of the <server> syntax.
>
>Maybe that "should" is a bit too strong, and there should be some 
>warning that username-password authentication has security risks. 
>Also, Port numbers need to be mentioned. So how about:
>
><server> is defined in [2396bis], and provides for a <host>, a 
><port> (defaulting to 119 in this scheme) and possibly a 
>username/password. This latter is usually deprecated on security 
>grounds, but is still in widespread use on current news servers and 
>so clients MAY implement it.
>
>{There are currently drafts in hand for SASL and TLS for the NNTP 
>protocol, so the practice should disappear gradually, but I am still 
>concerned that there seems to be no way for URL schemes to cope with 
>SASL.}
>
>Then I think you need to say:
>    If no <server> is specified, the resources are to be retrieved from
>    whatever server has been configured for local use.
>just to save the bother of repeating it three times in what follows.
>
>2.1  The newsURL has a <newsgroup-name>
>
>2.2  The newsURL has a *
>
>2.3  The newsURL has a <message-id>
>
>Those three sections are mostly fine, if you move the repeated 
>sentence as I suggested above. However, under the <message-id> case 
>you need to say:
>
>    A <message-id> corresponds to the <msg-id> of RFC 2822 and to the
>    Message-ID of section 2.1.5 of RFC 1036, but without the enclosing
>    "<" and ">".
>
>which comes almost verbatim from RFC 1738. There is often confusion 
>as to whether those enclosing "<" and ">" are meant to be present, 
>and some implementations of this scheme are tolerant enough to allow 
>them.
>
>
>
>    For example, neither
>
>       news:"abcd"@example.com
>
>    nor
>
>       "ab\cd"@example.com
>      ^
>      news:
>
>    is in canonical form, because the form
>
>       abcd@example.com
>      ^
>      news:
>
>    is available.
>
>2.4  The nntp URL scheme
>
>So I would call this section 3. Here is a suggested text:
>
>    The nntp URL schemes is used to refer to individual articles of USENET
>    news, as specified in RFC 1036.
>
>    The nntp URL takes the form:
>
>
>       nntpURL     = "nntp"  ":" news-server newsgroup-name "/" range
>       news-server =  "//" server "/"
>       range       = article-number ["-" [article-number]]
>       article-number = 1*DIGIT
>
>{The syntax for <range> is taken from 
>draft-ietf-nntpext-base-21.txt, so you could instead refer to that. 
>The syntax for <news server> probably does not need to be repeated 
>here.}
>
>    Observe, in contradistinction to the news scheme, that the <news-server>
>    is not optional here, because the mapping from <article-numbers> to
>    actual articles is unique to each server.
>
>3.1  The range is a single <article-number>
>
>    The resource retrieved by this URL is the Netnews article numbered
>    by the given <article-number> in the given <newsgroup-name> on the
>    given <server>.
>
>3.2  The range encompasses more than a single <article-number>
>
>    The resource retrieved by this URL is some means to gain access to
>    the articles numbered within the given <range> of <article-
>    number>s in the given <newsgroup-name> on the given <server>
>    (usually by invoking a suitable news reading agent initialized to
>    access that range). A <range> of the form "nnnn-" provides access to
>    all articles numbered "nnnn" and above.
>
>{Having said all that, is the form "nnnn-mmmm" actually implemented 
>anywhere? RFC 1738 provided just a single <article-number>. OTOH, I 
>can see that the form "nnnn-" could be useful to people for keeping 
>track of the last <article-number> read who now want to see 
>everything newly arrived since then. I even do this myself on 
>servers that do not support the NEWNEWS command (but by making the 
>direct NNTP calls, not by using any URL).}
>
>{And finally, if we all agree that the nntp scheme is now properl;y 
>defined, we can go back to the question of whether the two schemes 
>should be combined, though I think it would probably be better not 
>to.}
>
>--
>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@clerew.man.ac.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
Received on Saturday, 6 November 2004 01:55:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:35 GMT