Re: A Parable of Two Carpenters

From: Alan Connor (zzzzzz_at_xxx.yyy)
Date: 12/12/03 

Date: Fri, 12 Dec 2003 08:28:44 GMT

On Fri, 12 Dec 2003 01:10:28 -0600, Bx. C <null@the.void> wrote:
>
>
>
> "Alan Connor" <zzzzzz@xxx.yyy> wrote in message
> news:DGdCb.10947$rP6.1251@newsread2.news.pas.earthlink.net...
>> On Thu, 11 Dec 2003 11:03:21 -0000, Beth
><BethStone21@hotmail.NOSPICEDHAM.com> wrote:
>> >
>> What do you think of the idea of starting a free 'search' engine? One that

<snip>

>> cannot be bought like google and that wouldn't allow people to manipulate
>> their ratings like google et al do?
>>
>> I put search in quotes because I don't think there's any NEED to search
> the
>> Net.
>>
>> All you'd have to do is ask anyone that wanted their site listed to fill
>> out a form and return a challenge-response from the address given, all
>> by email. Run a whois just to be sure and send the form, compressed and
>> converted to a format the search tools can deal with to every box with
> that
>> segment of the entire distributed database.
>>
>> So it would be a directory service, really. Be easy to distribute the
> database
>> among 10,000 or so Linux users, with just a few servers routing search
> requests
>> to available boxes, where it would drop out of the picture, and the
> database
>> box would send the files directly to the requesting party.
>>
>> Also, you could have most of the search software on people's home boxes,
>> permitting the app on the database boxes to be very small and simple.
>>
>> So someone would send an email to the one of the "routers" which would
> already
>> be formatted by the user's software, and it would be sent to whatever
> available
>> "nodes" were relevant, and that box woud send the forms, compressed to
> whoever
>> made the request, directly.
>>
>> I'm getting REAL sick of Google and that bunch!
>>
>> This would be the perfect project for all of us Linux rebels.
>
> how would that work for people who search for "sex", "hardcore", "porn",
> etc.
>
> i may be misunderstanding what you're saying.. but.. if i'm not.. from what
> you're describing, can we say "liquified spam"?? (aka. spambot... or in this
> case... more like... "spamnet"
>
> i'm not meaning to put down the idea... it sounds decent... but... almost
> anything free gets misused and abused (ICQ,AIM,Y!,MSN,IRC,.. dare i say
> LINUX...) just as much as it gets properly used....
>
> there's no bad programs.. just.. misunderstood... even virii have
> feelings... they need a proper home... one where they can procreate at will
> without someone coming along to kill them... but virii need to learn to
> behave first.. and do as they should, instead of as they were told... it's
> the users that are bad or good...
>
> ummm.. throw out that last paragraph.. i think someone spiked my applejuice
> tonight...
>
> anyways...
>
> Bx.C
>
> PS. NT
> PPS. NT
> PPPS. one of these days i'll remember what it was i was gonna say...
>
>

:-)

Hmmmmm....I send a mail that has been strictly formatted by the software on
my box at home to a "routing node" which accepts mail in no other format.
It's compressed too.

The ONLY thing that is done with that mail is to pass it on to whatever
part of the distributed db is available, and the only thing it can do there
is tell the local program to send *descriptions* of websites that have
vetted forms on record.

The initial request would include a unique identifier for that transaction
which would include the IP address of the sending party, a specific port
to send the response to, and a serial number.

The first thing whatever box in the distributed db that received the request
would do, is send a UDP datagram to that box and if it didn't receive the
transaction serial number back from that IP address and that port, then it
would be dumped.

See? No possible way for spammers to do anything but whimper.

As for porn and such? Free Speech is the Law. The InterNetCardCatalog would
be completely neutral.

(with one exception: commercial sites would be in one broad category, and
non-commercial in another)

Any webmaster submitting a listing-form would have to give their real
address, and return a challenge response to prove it.

The form would contain information about the site that would be verifiable
with a simple program that would visit the site and create a report that
would be compared with the information in the form. If they didn't match,
the request for listing would be denied.

(this software would be on the distributed nodes, which would do the
vetting, thus decentralizing this phase. The address that the form was
sent to would be just another "router" like the ones used to send in
requests. Not pop or imap servers in either case. It's just convenient
for the user app to send the requests with the email programs they
already have)

That serial number would be a password. If it wasn't the current password
for that port, then the communication would be turned away except for
an echo-request type packet to tell the calling machine that it had been
reached. No connection would be allowed.

If anyone complained that a site wasn't as advertised, (a form for this
would be incorporated in the userside software) the site would be checked
again against its description in the form, and if they didn't match, then
the site would be dropped like a hot potato.

(once again, this re-check would be done by one of the distributed dbs.)

Webmasters would have to submit an up-date form whenever they made significant
changes to their site.

There. Covered all the bases?

And thanks for making me think harder. It's looking more and more like a
viable idea. The decentralization of the whole system would make abuse
almost impossible.

AC