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

Re: [syndication] Re: Using NNTP



One of the fallacies of these lists is that somehow we get a sense that we
have a say in what people do. I think it's a huge mistake. We're watching
the wave, at best we can give it ideas, see which ones catch, and invest in
those. And then do it again, see what happens and take a few more steps and
see how it goes. Repeat until retirement.

It's a fact, some people can't run Web servers on their desktops. But that
doesn't mean they can't FTP from their desktop, or make SOAP/XML-RPC client
calls from their desktop. Most can. So (almost) everyone theoretically can
participate in what I'm calling The Two-Way-Web.

Further, yes I think we do have to take responsibility for backing up our
data. As an industry we haven't yet had a complete meltdown. But what
happened to NNTP archives last week when Google acquired Deja? I haven't
really been following it. How safe are our archives on Yahoo? I have no
idea. I thought we were using eGroups and then one day we're using Yahoo. I
lost my calendar last year because I kept it on a centralized (free) server.
Who do I call to get a copy of it? I have no idea. It's lost. That won't
happen again, because I've take responsibility for backing up my data. I
always *was* responsible for this, but like many (most) Internet users I
forgot that.

Anyway, I went through a similar thought process doing the <enclosure> stuff
in 0.92. Why not just use email? Well, I worked it out. It wouldn't take too
long before I was stuck downloading a big video while trying to get a few
short email messages. Email is already used for one kind of content,
grafting another kind onto it would piss users off. So I decided to let the
XML carry a pointer to the data, and let a new piece of software do the
download under script control so it could be highly configurable. It's not
email because email clients weren't designed to handle it. I don't use any
NNTP clients, so I don't know, but it wouldn't surprise me if they didn't
handle weblog type stuff very well. We already know that the Web does it
well, with a little help from aggregators and shared agents like
weblogs.com.

And some people have pure Internet connections on their desktop. Back in the
PC era (80s) I liked to make software for the high end, and I'm doing the
same now. It takes a few years for a software product to mature and find its
niche. In that period, Moore's Law is putting yesterday's high-end machine
on today's average desktop. I think that's going to work for network
connectivity as well. I design for a full peer, one with no firewalls or
proxy logic. I am optimistic that this decision will prove to be right a few
years from now. In the meantime I want to be as inclusive as possible. But
some organizations won't allow their users to run servers. Gotta respect
that, one step at a time, convince them that it's worthwhile to add power to
the machines on their LANs.

Dave