- From: Al Gilman <asgilman@iamdigex.net>
- Date: Sat, 06 Apr 2002 08:12:16 -0500
- To: Graham Klyne <GK@ninebynine.org>, "Roy T. Fielding" <fielding@apache.org>, Mark Nottingham <mnot@mnot.net>
- Cc: "Tim Berners-Lee" <timbl@w3.org>, "Paul Prescod" <paul@prescod.net>, uri@w3.org
At 01:54 AM 2002-04-06 , Graham Klyne wrote: >[I noticed this pair of messages, one from Roy on WWW-tag@w3.org, the other >from Mark on uri@w3.org. to avoid cross-posting, I've continued this on >the URI list only.] > So what's the question? As a sometime purveyor of the "mailbox" interpretation, let me support Roy's point, with a minor PC wording change which I hope is acceptable as friendly amendment. To say a "composition window" is too GUI-flavored. The response to activating a mailto: URL is that a message object is created and returned open for editing. [Yes, with properties initialized from the URL.] The object can well be at the other end of a phone call as you answer your messages which you read by text to speech transformation. http://trace.wisc.edu/world/modtrans/ The integrity, the one-ness of the transaction launched with the mailto: URL activation has to do with the unit message, not a single recipient mailbox. I hope Roy can accept that as a 'yes' to his assertion. The message editing session binds the message object to a value [in a very strong preponderance of times of type message/rfc-822] and once that value is bound it is re-identified not with a mailto: URI but with the mid: URI with the identity-stamp created and associated with this message object by the initiating MTA. Is that so? Al PS: It would also be possible to say that 'window' is just an abstraction for the presence of the edit session as a sub-session of your HCI 'delivery context' session. But that carries too many GUI-built expectations of autonomy along to be safe, as well. The strong autonomy among windows that we experience in the GUI has to be modified to get graceful transformation to a blind-optimized delivery context, for example. The optimization criteria for blind HCI engagement call for a stronger "virtual table of contents" integration of the session constituents than what we get away with in the best-current-practice WIMP delivery context. So to be properly general it is better to talk to the state of engagemenet with the message object itself and not say things that imply any given degree of autonomy between the in-composition-process message and the referer context of the URI reference. Compare with the navigation concepts for the now-Standard [TaDa!] digital talking book format for what I mean by strong table of contents support in the session structure. The following reference is the epicenter, Truncate the URL a segment at a time for more context. Document Navigation Features List http://www.loc.gov/nls/z3986/background/navigation.htm PPS: It's not even misnamed. Just mis-explained. Names are heuristic, not definitive. The point is that you have to parse 'mailto' as 'mail to' in order to get the object type (mail) and the initialized properties of that object (To:) properly identified in the deconstruction of the name of the scheme. This is breaking the neologism down according to its natural language constituents, and so is good naming, not misnaming. We just hadn't migrated to camelCase yet to make the parse clear. >I think Mark's observation here rather underscores Roy's point. The >"resource" identified by a mailto: URI would appear to be, in Roy's words, >"a composition window with the following pre-filled attributes" rather than >a mailbox. > >#g >-- > >At 12:38 PM 4/4/02 -0800, Mark Nottingham wrote: >>The 'mailto' [1] and proposed 'sms' [2] schemes allow multiple resources >>to be identified by a single URI; e.g., >> >> mailto:bob@example.org,mary@example.com >> sms:+41796431851,+4116321035 >> >>Is this encouraged in new schemes? I.e., is it a good idea to have a >>one-to-many uri-to-resource mapping? >> >> >>1. http://rfc2368.x42.com/ >>2. http://www.ietf.org/internet-drafts/draft-wilde-sms-uri-02.txt > > > > > >At 03:52 PM 4/4/02 -0800, Roy T. Fielding wrote: >>>The mailto: schema name was badly chosen, but the concept is >>>sound as originally defined. It was intended simply to be a space >>>in which to put all the internet email addresses, which are called >>>mailboxes. A mailbox is an abstract thing, related to email messages >>>by (for example) To: From: and Cc: feilds but also used in many >>>other situations. It also normally has a relationship with the social >>>entity >>>-- typically a person or group --which owns it. >> >>Hmm, well, from my perspective outside the early CERN days, I'd have to >>disagree. >>Mailto has been consistently defined and implemented to mean "get a >>composition >>window with the following pre-filled attributes." I absolutely agree that >>what >>we should have is a URI for a mailbox that can be placed in a form for >>defining >>the action of a POST, but mailto was created before FORM was invented. >> >>I think it is critical that we not try to redefine the semantics of existing >>URI after they have been introduced to the Web. If someone wants a generic >>URI for mailbox, they are going to have to invent something other than mailto >>for that purpose. Wishful thinking is not interoperable. >> >>....Roy > >------------------- >Graham Klyne ><GK@NineByNine.org> >
Received on Saturday, 6 April 2002 08:12:55 UTC