[rfc-i] URL checking

Carl Malamud carl at media.org
Mon Oct 17 23:55:58 PDT 2005

Hi Joe -

> > At 2:40 PM -0700 10/17/05, Joe Touch wrote:
> My point is that liveness during that period isn't a guarantee of
> liveness in the next day when it becomes an RFC. Even if the author
> checks it because the editor flags it, it's just as likely to die in the
> time to publication.
> >> It's just as likely to break by the time it
> >>goes from the queue to final
> > 

Er, Joe, that's some bad science you're practicing.  If the probability
of breaking was evenly distributed, this might be the case.

In this case, however, it seems like a valuable check.

> > In this particular case, it didn't break: there was a typo in the draft.
> Checking for liveness of all URLs in an RFC should be a fairly simple
> script; checking that they result in the intended page, however,
> requires manual verification.

A frequent issue for me when writing a draft is that I mean to point to someplace
authoritative and end up citing some blog instead.  Needless to say, I use
the AUTH48 to make sure that links to IEEE didn't get changed to Engadget.

> I don't believe in long-lived URLs; they're exactly as long-lived as the
> organization that they're registered with, and no more - in all cases.

I think this goes to the point.  Seems that most folks are pretty conservative
on their choice of references, particularly in normative sections.  I like to
think groups like the IEEE, W3C, and particular the RFC Editor are long-lived
institutions and are worth citing.  Granted, all organizations might disappear 
at the same time, but then we have other problems.  For now, at least, if I cite
a source, I'd be very pleased if the RFC Editor did that one extra check to make
sure my citation didn't get mangled.

> IMO, given search engines, it's not clear that they serve a useful
> purpose in archival documents anymore.

Joe, you seem to have adopted some odd-ball geek religion which believes Google 
is going to solve all problems, the web will go away soon, and all life, including 
the writing of RFCs, are ultimately pointless.

Despite your hopes in the Church of the Search and the second coming of
the holy trinity (the Larry, the Serge, and the Holy Schmidt), I think 
it is worth taking those extra two seconds to make our current RFCs just
a tad bit better. 


More information about the rfc-interest mailing list