Re: Semantic mail

Greg FitzPatrick <greg.fitzpatrick@metamatrix.se> wrote:

> I argue that before we can have a semantic web we should at least have a
> semantic mail.

While it may not be a prerequisite, I certainly think it's useful and may
actually be one of the killer apps for the Semantic Web.

We've had this discussion before on XML-DIST-APP, search for "Time for
XMail?"

> Though I am completely serious about this, I will warn you there is an
> undertone in this proposal:  Would we do it?  Would we really insert these
> tags into our mails?

What tags? I see no more tagging than already exists -- the rest is just the
sharing of data.

>> http://www.mozilla.org/blue-sky/misc/199805/intertwingle.html
> Greg: good ideas from Jamie Z but as I understand it, this deals with mail
> as defined in RFC 821-2 (as is), without thought to extensions.

But it does mention RDF and XML. ;-) Anyways, if this hasn't been done yet,
and requires no changes to the email infrastructure, what makes you think an
entire infrastructure overhaul will take off?

>> Re the details, I would propose a distinction between representations of
>> mail in RDF and re-formaliting the body of mail messages in XML/RDF. The

Of course.

>> latter is harder as you'd need rich mail clients. Somewhere in
>> http://www.w3.org/Submission/ there was a proposal a year or so back for
>> such a format.

I can't seem to find it, but for the former:

http://www.w3.org/Submission/2000/07/

looks somewhat useful. In general, this may have some overlap with the XML
Protocol Activity. Personally, I'd recommend my changedPage spec, for its
utter simplicity:

http://my.theinfo.org/changed/

It may require some application-specific additions, but it was designed with
exactly this type of application in mind.

>> Who sent it, which messages it is in reply to, perhaps an RDF
>> representation of autoclassified categories, plus annotations / comments
>> added after the fact. That'd be plenty to be going on with.

Definitely.

>> We wouldn’t have to worry about the wisdom of dividing up the RDF list into
>> interest and logic or debate the pros and cons of a SW list.  We would no
>> longer have to sort through all those hundreds of mails on our desktops but
>> rather follow the flow of ideas as interpreted and presented by parsers and
>> applications of our choosing.

This is, of course, the value of URIs. One URI per message, categories
(mailing lists) added as RDF assertions.

> we replace the addressee with wwsm@w3c.org signifiying that
> the mail was wwsm hunky-dory and to be considered as part of the big picture,
> we would have to add a tag for the core list name ex. – RDF logic.

Hmm, rather than trying to build this kind of extensional mess on the
current system, I say start anew.

> 2. We should formalise our way of quoting.

If we used HTML, we could use <q cite="{x}">.

> mail was stamped with its archive URL.

Definitely should happen -- mail should _be_ its archive URL! This is the
URI world, after all.

> I would be prepared to offer 10 hours of my short life on this if anybody
> else was interested.

Only 10? I'm interested, perhaps even more than 10 hours worth.

> Perhaps an unofficial BOF in San Diego?

Sorry, I won't be in San Diego.

> There might be some tie-ins with WebDav and XML-mail as well.

XML-Mail? You mean:

http://www.oasis-open.org/cover/xmtp.html

? Looks like Jonathan's site is down.

-- 
[ Aaron Swartz | me@aaronsw.com | http://www.aaronsw.com ]

Received on Wednesday, 8 November 2000 18:46:37 UTC