[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnupedia] From the "Department of Wacky Ideas"

From: Bob Dodd
Subject: Re: [Bug-gnupedia] From the "Department of Wacky Ideas"
Date: Sat, 20 Jan 2001 14:42:00 -0800 (PST)

> I think this would somehow defeat the point of the
> Encyclopedia, because although the idea is to have a
> large spectrum of opinions and subjects, it is also
> partly to manage to do this in one centralised
> resource.   

Ah... but distributing the database isn't the same as decentralising
control of the database. It's only where it's put, not who puts things
in it. There's lots of valid objections you can make to the scheme (and
Tom found a few :-)) but that isn't one of them.


> And your points supporting it are quite good, though,
> have you seen gnutella recently?  Not very good
> anymore.   I dont think a distributed approach has
> enough control.   And if someone suddenly decides to
> take their server offline, we've lost some entries.
> And it would be slow.

Speed and the possibility of losing parts of the database are
definitely negatives to it. I don't quite know what happened to
gnutella, it probably just goes to show that jou can't distribute a
database without considering the content properly. I also suspect
problems with their indexing approach.

Taking a server off-line wouldn't such a problem: the model is designed
for massive redundancy: if you only have a few "untrusted" servers out
there "in the wild" so to speak, the approach would fail miserably.

> Also, we would have to develop a system which people
> could use.   The advantage of our current idea is it
> can be as confusing as hell (well nearly) to
> understand how to set it up, but thankfully only about
> 7 people in the world would have to.
> Otherwise we would have to worry about making it
> usable, and then as you say we have the trouble of
> porting it to other OSs.

I think I've already replied to this point for Tom: it's us with the
problem, not the users. They shouldn't see the problem. Actually
whatever solution we have, the users should be blissfully unaware of it

> Interesting, though im not sure its quite what this
> projects all about.

To some extent, Tom was right when he sort of said that it was a
solution looking for a problem, just not our problem. I think I would
say it's more a solution anticipating a problem, though I also accept
that there are many more(expensive) safe, reliable, conventional
solutions than this crackpot scheme...

It does have its good points though. For one (asuuming that it
worked...) we would have a cheaper, more reliable solution than any
conventional approach. Of couse it could also fail miserably.

What I would say, is that we need an honest list of advantages and
disadvantages to any architecture being put forward. This "wacky idea"
got shot down pretty fast, and it probably deserved it, but at least it
had its list of pro's and cons, which I hope was reasonably honest.
There are already too many assumptions in our postings as to what the
architecture will be, what platform it will run on, and what
applications we will use. We don't even have a server we can store code
on yet.

Anyway, thanks for the comments Rob, I appreciate them.
/Bob Dodd

Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. 

reply via email to

[Prev in Thread] Current Thread [Next in Thread]