Re: [unrev-II] Moving beyond the web...

From: Jack Park (
Date: Thu May 04 2000 - 07:19:21 PDT

  • Next message: Bill Bearden: "[unrev-II] Eugene and Paul..."

    One thought, no, make that two thoughts come to mind here. The first one is
    bandwidth. It is not now and never will be unlimited. One of the charms of
    distributed computing is that you can pass searches off to some other cpu
    rather than tie up bandwidth downloading files just to search them.

    The second one is security...

    From: Jon Winters <>

    > Jack Park wrote:
    > > > > <snip>
    > >
    > > The thesis behind your statements is certainly valid. Indeed, our brains
    > > highly distributed as Gnutella would be. But, our brains depend on
    > > connectivity, and your home computer and mine are not necessarily
    > > 24x7 as I presume the individual "nodes" in my brain to be (after a
    > > of course).
    > Actually all of my computers at home are connected 24x7, they have been
    > for years. (but I live in a geek compound and I run a web hosting
    > business in my spare time :-)
    > I also try to disconnect my brain every night when I sleep. ;-)
    > > So, to take a file-shuffling approach seriously (and I think we should
    > > at it seriously), we ought to take a bigger picture view. Here's mine.
    > >
    > > I see this functionally like a neural net. I see a gui in which one
    > > examines some kind of log of actions related to selected "threads of
    > > thought". I see one responding to those actions, not unlike an email
    > > thread, but perhaps cleaner. Here's why. In the background, there is
    > > processing going on (at each node), making sense of the actions. When
    > > sense is made, it is rendered presentable. That way, we don't deal with
    > > the verbage present in emails, but rather, get presentations (views
    > > constructed according to view preferences), and we react accordingly.
    > > Periodically, each node makes contact with other nodes -- perhaps those
    > > which there is a thread subscription (wiring, in the neurophysiological
    > > sense), and the nodes synch up (an action). Also periodically, each
    > > gets synched with the big bertha node "out there" that keeps all threads
    > > (where you go to subscribe to a thread).
    > Good points. I was thinking about it on my drive home. This is what
    > went thru my mind...
    > Each node is kinda like a book and Gnutalla is a library. When you go
    > to the library you don't have the ability to do a full text search of
    > every book in the library. You search by title, author or topic.
    > Searching on Gnutella is much the same.
    > But then I wondered.... Hmmm... when you search the web you do have full
    > text search a lot of the time. Would moving to a Gnutella based system
    > be a step backwards? There are those who would not want to give up full
    > text searching.
    > Then I had an idea. What if we coded a Gnutella client that could then
    > search the contents of your local dataset? Folks using our client could
    > do full text searches Gnutella style. (I only skimmed over the
    > protocol.. don't know if this is possible)
    > The other alternative would be to retrieve the information and then
    > conduct your full text search on your machine.
    > > Bertha performs analogical discovery by crosslinking threads.
    > > enough, some of this may be done out in local nodes when one is
    > > to more than one thread, or when one downloads a new thread for study.
    > > the long run, all of this finds its way back to big Bertha.
    > >
    > > What fun. Not sure it's even close to what Doug has in mind, however.
    > Doug, you'll chime in and correct us if we stray too far won't you?
    > --
    > Jon Winters
    > "Everybody Loves The GIMP!"

    Avoid the lines and visit for quick and easy online
    reservations. Enjoy a compact car nationwide for only $29 a day!
    Click here for more details.

    Community email addresses:
      Post message:
      List owner:

    Shortcut URL to this page:

    This archive was generated by hypermail 2b29 : Thu May 04 2000 - 07:26:52 PDT