From: Adam Rifkin (adam@KnowNow.com)
Date: Thu Sep 07 2000 - 17:32:01 PDT
The welcome wagon continues faster than I can keep the FAQ up-to-date...
Welcome, new FoRKers! Number of people on this list just hit 180.
Let's flood the mailing list while Kragen's on vacation so he has
something to do in Ohio... :)
> From: email@example.com
> I'd like to be an official FoRK!
> Thanks, Jeff;
> Jeff Barr - Home: 425-836-5624 Office: 425-936-3098
> 4610 191st Place NE. Redmond, WA
Jeff: Your wish is our command.
> From: firstname.lastname@example.org (Justin Mason)
> Hi Adam/Rohit/whoever's reading this, I think Kragen forwarded on a mail
> from me a while back), and occasional reportings of rather colourful
> jargon coinages from this side of the pond.
> I've been putting off joining FoRK for a few months now, as I'm already
> on too many high-volume mailing lists, but what the hell -- it's good
> discussion. ;)
Justin: tomwhore has more jargon coinages coming out of his buttockses
than most mailing lists have in their most colorful flamewars.
BTW, for extra fun, say something about Lloyd Wood in the middle of a
post *just* to see if he's still listening. Hey Lloyd, AT&T Wireless is
thinking about merging with Nextel; two McCaw tastes that taste great
together, right? :)
> From email@example.com Wed Sep 6 22:18:43 2000
Fresh from Deepleap.com ... forkers_in_austin_texas++;
> From: Bill Humphries <firstname.lastname@example.org>
> Subject: subscribe
The world needs more people who know XSLT. :)
http://whump.com/ there it is...
> From: Rahul Dave <email@example.com>
> How do I get on this list? sending subscribe to firstname.lastname@example.org
> went into a black hole and I am not able to locate subscription info.
Rahul: Any friend of Dave Farber's is a friend of ours...
Welcome to FoRK, a place where bits go to die... :)
> (more from Rahul Dave on the decentralization list today)
> Consider Freenet. Every message sent out waits needs to be replied to
> before another message can be sent. If the original respondent has the
> data, it is sent over, otherwise its not. When the data is sent over
> it is cached along the way.
> Consider Gnutella. Queries are in parallel, and are separated from
> the actual obtaining of the data. Because 20 nodes are queried in
> parallel, querying is fast. But data is stagnant, on the original
> Now, when I query, I either know what I want(shakespeare.hamlet) or I
> want to ask(shakespeare.story("some danish prince")). In the former case
> freenet is faster, and in the latter, gnutella. Freenet mitigates this
> by its caching.
> So what if we combine the two mechanisms. Metadata needs to be separate from
> data for easy querying. Gnutella like mechanisms can be used to do this
> querying. However, gnutella leads to exponential number of messages, how
> to reduce this. Cache the results of queries, and who has answers to the
> queries. In other words, we can cache the metadata, either purposely or
> on response to queries, or, in Yenta style, and possibly costing less
> in latency, the answer to the question: who can answer the query. This
> is the notion of a rumour cache and when a gnutelle style node is
> asked a question, it will not forward to all its peers, but only those
> in its rumour cache which match (what match means is a whole different
> story) the query.
Rohit, read his original rant on Web and P2P infrastructures at:
Pedal to the medal, blow by blow, You're running with the devil, it's touch-and-go. -- Emerson, Lake, and Palmer, "Touch and Go"
(Man am I gonna miss Napster if it goes away; Gnutella just doesn't let me probe around other peoples' *collections* like Napster does, and Dave's right, *that* kind of sharing & swapping is why I keep going back... though I'm still waiting for www.dankohn.com to make his whole collection available... :)
This archive was generated by hypermail 2b29 : Thu Sep 07 2000 - 17:35:21 PDT