[FoRK] portals vs time
dl at silcom.com
Wed Feb 2 00:59:51 PST 2005
> Perfect email would have some sort of personal workflow system -- you
> could send out 'ping' messages to folks that owe you stuff and get
> their responses checked off and filed appropriately.
Winograd and Flores tried to commercialize
something along these lines twenty years
ago -- but maybe their structure was too
(James, this is the kind of thing I'd been
thinking of in when ranting about a double
entry-like system for workflow/PIM)
> You almost always
> end up with email backed by a decent archiving system, rather than a
> discussion system with email notifications.
The best systems I've used have had both a
structured database-y side and a good email
side -- so email conversations are rather
freeform, and after some kind of consensus
is reached in discussion, someone takes the
effort of hassling the update to the formal
side -- kind of like code, which combines a
strict format for machine-meaningful stuff
with an informal commentary attached.
> If they only expected 80% success, it would be
> relatively easy. Cf. text constraints from Lapis by Rob Miller:
Nearly a decade ago, when Miller was in
the CMU HCI group, I put hooks into our
web editor so he could try to get some
semantic information from selections and
deletions. It's nice to see how much
progress he's made since then, and a bit
humbling to realize how long it's taken.
:: :: ::
A related not-necessarily-high-accuracy
technique for recovering trees from flat
Cilibrasi & Vitanyi, "Automatic Meaning Discovery Using Google"
> The approach is novel in its unrestricted problem domain, simplicity
> of implementation, and manifestly ontological underpinnings.
(shades of Aristotle's Categories, in the
attempt to make use of an extant corpus of
statements about the world to derive more
More information about the FoRK