Dave Long wrote:
> > These guys really need a database to start from, ...
> Nonsense. The database solution requires storage, it requires
> maintenance, it requires a good deal of computation to determine how
> items are disseminated along the web, and it probably requires a
> certain amount of human intervention in tuning precision and recall.
So engage the humans ala dmoz. There's nothing like taking ownership
to help build a community, as long as the UI is mind numbingly simple.
> There's a much simpler implementation: feed your subscribers what
> your paying vendors (not random strangers!) wish them to see. After
> all, some animals are more equal than others...
Not sure I follow. How does that differ from a shopping portal?
Where's the metadata? Surely the vendors won't be providing all of
it?! I can just see it now, a TPSML (This Product Sucks Markup
> (cynicism aside, I *do* think this a neat idea, although I'm not
> sure how much of an improvement it will prove to be upon our current
> messaging+wetware architecture. I'd be more interested in a
> peer-to-peer munchkinized run at this particular fence.)
Until then, can we do much better than centralizing the metadata
and publishing all of it in an open format? What if we treat the
database as a cache for what metadata would normally be encapsulated
in a munchkin and cached everywhere (ie. the company would not own the
data). But yah, that's not gonna happen, at least not until somebody
can make a business case for it, which I wouldn't quite rule out;
where there's a will, there's a way.
Content-Type: text/x-vcard; charset=us-ascii;
Content-Description: Card for Mark Baker
org:Sun Microsystems Inc.;Consumer & Embedded
adr:;;126 York St., Suite 325;Ottawa;Ontario;K1N 5T5;Canada
title:Personal Apps Lead