- From: Al Gilman <Alfred.S.Gilman@IEEE.org>
- Date: Wed, 22 Sep 2004 15:59:49 -0400
- To: uri@w3.org
+1 to all Charles's assertions. I think he has the story straight. This includes but is not limited to "the 'nntp' scheme should not go away." On the maybe's that have passed through this thread: a. Allow <server> in URIs where the scheme is 'news'? I believe that this is implemented in multiple browsers, so should be added in as a friendly amendment above and beyond 1738. b. Is a 'news' URI by message-id identifying a single resource? Yes. In the more current thinking about URIs and the integrity and identity of resources, this fits well into the concept of an identifier as a sufficient key to isolate something from its peer resources. The language 'not enough to locate' is actually saying 'not enough to address a server for recovery operations' and is different from the current use of 'identifier' as distinct from a recovery address in current writings about URIs. c. '*' as a group wildcard? This can be found 'in the wild' in running code in Lynx. http://lynx.isc.org/lynx2.8.5/lynx2-8-5/lynx_help/lynx_url_support.html#news_url Hoping I held heat:light under 2:1 in this, Al At 2:16 PM +0100 9/22/04, Charles Lindsey wrote: >>2. Scheme Definition >> >> 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 > >OK, that syntax is correct now, but you need a normative reference >to RFC 2822 for <id-left> and <id-right> (or maybe to the >definitions in the Usefor draft, if that manages to become an RFC by >the time this draft is ready - but don't hold your breath). >Presumably also to RFC 2396bis for <server>, and it still is not >clear to me whether <server> could include user+password >information, and if so what one does if the authentication required >by the server is SASL based, which will soon become the norm. >> >> A <newsgroup-name> is a period-delimited hierarchical name, such as >> "comp.lang.perl.modules". A <message-id> corresponds to the >> Message-ID of section 2.1.5 of RFC 1036 [RFC1036], without the >> enclosing "<" and ">"; it takes the form <unique>@<full_domain_name>. >> <unique> cannot be quoted text or have escaping characters. > >No, it is not as simple as that, because you can also have an IP >address after the '@' and this is covered, in RFC 2822, under the >guise of a <no-fold-literal> and that brings in quoted text and >escaping characters again. > >So I think what you have to say is something like: > > "The <id-left> and the <id-right> MUST be in a canonical form in which no > <quoted-string> or <quoted-pair> is used in a context where the same > semantic meaning could have been rendered without such quoting; > moreover, no whitespace may be included, whether %-encoded or not and/or > quoted or not. > > For example, neither > news:"abcd"@example.com > nor > "ab\cd"@example.com > is in canonical form, because the form > abcd@example.com > is available." > >Yes, there are indeed email systems around that will happily treat >email message identifiers using those three forms as being >identical, which is of course a complete no-no in News. >> >> If <newsgroup-name> is "*" (as in <URL:news:*>), it is used to refer >> to "all available news groups". > >OK, that feature has been available since RFC 1738, but I am not at >all sure what it is meant to DO. I just tried it on my browser, and >it was totally confused, telling me "411 Invalid group name (not in >active).". I would be quite happy to see it simply dropped (unless >someone can point me to a system that does something useful with it). >Next, we really need some text to explain what resource is supposed >to be retrieved by this URL. Something like: > > "The resource retrieved by this URL is the Netnews article with the > given <message-id>. In a properly working Netnews system, the same > article will be obtained whatever server is accessed for the purpose > (assuming the server in question carried that article in the first > place and that it has not expired). If no <server> is specified, the > article is to be retrieved from whatever server has been configured > for local use." > >One may then need wording as to whether this is truly a global >resource, such as people have been discussing regarding the file >scheme. I am keeping out of that one myself - the point may become >more relevant if we bring back the nntp scheme. I notice that RFC >1738 contained the paragraph: > >"The news URLs are unusual in that by themselves, they do not >contain sufficient information to locate a single resource, but, >rather, are location-independent." > >I am far from clear what that actually means (if anything), but >maybe it is related to what I was trying to say in my suggested >paragraph above. AFAICS, they do indeed locate a single resource (if >you count different copies of the same article as "single"). >> >> The nntp URL defined in RFC 1738 is deprectated. > >No, I don't think we ever agreed that, and a couple of people have >pointed out places where it is implemented. I have also seen it as >the recommended method, in Opera, to force the system to reload an >article from the server if the client has lost it somehow. > >So I would be in favour of bringing it back. I might even be >persuaded to combine it with the news scheme as originally proposed, >but only if we establish first exactly what it is meant to do on its >own. > >-- >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 Wednesday, 22 September 2004 20:36:12 UTC