On 05/09/2008 4:43 PM, Aristotle Pagaltzis wrote: > * Peter Saint-Andre <stpeter@xxxxxxxxxx> [2008-05-09 23:50]: >> From the definition in RFC 4287 I would say that "related" is >> intended for static resources that provide connected/associated >> information, such as background data, additional content of >> interest, etc. > > Disagree. On what basis do you disagree? >> Granted, the definition "identifies a resource related to the >> resource described by the containing element" is somewhat >> circular, but it doesn't quite capture what I had in mind for >> "discuss". > > You are reading too much into it. Basically `related` has no real > relation semantics at all. It’s pretty much the same thing as an > HTML `<a>` tag: a link you can follow that’s there because > someone deemed it relevant. No one includes a hyperlink unless they think it's relevant in *some* way. Saying that "related" indicates that someone thought the URL was relevant in some way strikes me as void for vagueness. Comments are related. Alternate locations are related. Licenses are related. Replies are related. Pages that identify the author (self) are related. Why do we need any link relations other than "related"? >> I think a link of type "discuss" would identify a venue where >> multiple people can exchange messages about (i.e., engage in a >> discussion or conversation about) the resource. Such a >> conversation could take place via email, Internet Relay Chat, >> XMPP groupchat, MSRP groupchat, web forum, voice or video >> conference, or whatever other technologies people come up with >> for social interaction and discussion over the Internet. The >> use case is that you don't just want to find related content, >> you want to actively engage in a conversation about the topic. >> This is different from "related" because a conversation happens >> at the link, and it is different from "comments" because you're >> not following a feed of comments on the entry but instead >> engaging in a conversation. So perhaps "conversation" would be >> better than "discuss", but be that as it may I think these use >> cases (multi-party interaction, not reading a static resource >> or following comments made by others) are sufficiently >> different to justify definition of a new link relation. > > Modulo my disagreement about what `related` means, you still > haven’t answered James’ question: > >> On 05/09/2008 3:07 PM, James Holderness wrote: >>> Why would a client need to be able to differentiate a >>> "discuss" link from any other link type? Because a human probably cares if he or she can converse with other humans about the topic. We *are* social beings, are we not? > Specifically I assume that all of the protocols you indicate > above have URI schemes that clearly identify their URIs as names > for places where users can converse, Why do you assume that? Is the typical human user even aware of URI schemes? And if so, do they differentiate between irc:, mailto:, xmpp:, sip:, msrp: (etc.) as identifying more socially-oriented technologies? That's not even to talk about http: URLs as pointing to static resources as opposed to more socially-alive spaces like web forums. > and if you wanted to use > them in an HTML document you’d put them in an `<a>` tag without > thinking twice about it. And on systems with the respective chat > software installed, browsers would indeed do something useful > with such links without need for a `<chat>` tag or some such. > > So why do you need more in the context of XMPP sessions? You and I don't need more because we know what those URI schemes mean. But let us not fall into the trap of thinking that people who frequent IETF protocol discussion lists are typical users. Peter -- Peter Saint-Andre https://stpeter.im/
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature