On Wed, Mar 2, 2011 at 9:25 AM, Dave Cridland <[email protected]> wrote: > Actually, if you make the document creation be a simple message, and you > ignore completion, then what you end up with is the ability to correct > messages after they've been sent, which can be extremely useful.
Retroactive message editing is not included in my spec, so this cannot be done. However, my spec is extensible to support retroactive editing: That is part of the reason I included a "msg" attribute, so that retroactive editing is possible, by referencing the msg number. This would obviously not be recommended in normal "chat" conversations. Retroactive editing is only necessary in some contexts such as meeting transcripts and real time captioning. There can be a workflow where messages are delivered from one system to another. One person is a caption operator, and the next person is a caption editor, and 5 seconds later, the captions come out. (two persons doing captioning -- one transcribing, one editing, all in a near-real-time manner) I have a spouse who is a court reporter, and I am considering future projects that take advantage of a "workflow"-based correctness-critical captioning system. But we're now getting off the topic. I want to keep the standard simple, so for <rtt> tags I am even considering removing the "msg" attribute at this time. (seq and type is still necessary) Comments?
