Re: barriers to providing a web presence (Re: "tag" scheme rationale)

> 
> > > It may be that one is unwilling to maintain a Web presence even
> > > though one would very much like to allow people to metaphorically
> > > follow their nose. There are barriers between many people and the
> > > maintenance of a proper Web presence.>
> >
> > Like what?  I think many people vastly over-estimate those barriers,
> > in this context.  Of course you've got to have (or hire) some
> > expertise and a little time and money, but especially if you use
> > services like purl.org and thing-described-by.org, what is so
> > difficult?   Maybe it depends what you mean by "proper" web presense;
> > a little RDF page is all that's needed for this application.
> 
> Hi Sandro:
> 
> I'm guessing this is all really tongue in cheek. ;) In a regular
> production environment we are looking at 1m identifiers, or 10's of
> millions and upwards, or sometimes up to the billions of identifiers.
> We're not going to gift those away. Or necessarily promise them on a
> sometime (here today, gone tomorrow) DNS address. Let's differentiate
> (right now) production environments from sandboxes. The numbers are
> different. The committment is different.

I'm not at all tongue in cheek, but maybe there's some big
misunderstanding.  I have no idea what you mean about "gift those
away".

When I mentioned perl.org and thing-described-by.org, I thought you
were worried about low-end solutions, where the cost of maintaining a
web presense is prohibitive.  But I guess you're thinking about
high-end, high-stakes situations.  What's the challenge there?

The idea that DNS names are "here today, gone tomorrow" is like the
local news anchor saying, "New evidence suggests your shoes could be
causing you cancer!  Stay tuned for more details later after all this
boring stuff an commercials."  Or like saying a new store shouldn't
advertize its location because maybe the building will burn down and
it will have to move.

The investment in some domain names is enormous, maybe in the hundreds
of millions of dollars, if you count everything that would have to
change if the domain name were somehow to be "gone tomorrow".  Think
of the big dot-coms.  Think of the mass-market software that calls
home.  There have been some well-publicized situations where domain
names changed hands inappropriately, but those are clearly the
exception.  I think people building web-oriented businesses have much
better things to stay up all night worrying about than whether their
domain name is somehow going to be taken away from them.

And how does the number of identifiers matter?  Managing 10^5 or 10^80
URIs starting with the same prefix is pretty much the same (fairly
trivial) problem, assuming you actually have some infrastructure to
pull up relevant information about that many distinct things.

      -- sandro

Received on Sunday, 16 October 2005 04:25:52 UTC