Re: [URN] Checksums in URNs

Ryan Moats (jayhawk@ds.internic.net)
Fri, 20 Dec 1996 08:32:08 -0600


Message-Id: <32BAA368.4FD2@ds.internic.net>
Date: Fri, 20 Dec 1996 08:32:08 -0600
From: Ryan Moats <jayhawk@ds.internic.net>
To: "Martin J. Duerst" <mduerst@ifi.unizh.ch>
Cc: Towsner <tows@earthlink.net>, Larry Masinter <masinter@parc.xerox.com>,
Subject: Re: [URN] Checksums in URNs

My turn to add my two cents worth...

Martin J. Duerst wrote:
> 
> On Thu, 19 Dec 1996, Towsner wrote:
> 
> > >Checksums and other kinds of decorations can be done outside the
> > >URL. Think of it as a metascheme:
> > >          ck:<checksum><url>
> > > e.g.     ck:A3Bd:http://www.parc.xerox.com/masinter
> >
> >       I like the idea of having the checksum be optional, as a
> > meta-scheme.
> 
> Same for me.

I'll make three...

> >I also think the checksum should be limited to alpha-numeric
> > characters, to avoid confusion.
> 
> What do you mean by "limited to alpha-numeric characters"? Can you be
> more precise.

How about just hex characters to keep it simple...

> >Perhaps this should be part of the URL
> > scheme as well.
> 
> If it is a metascheme, then it is a new URI scheme. It doesn't
> have to be discussed either in URL nor in URN syntax. It would
> go into a separate draft. It has to make sure, on its own,
> that it is compatible with the URL/URN syntax.

I agree with Martin, and if we use ck:<hex sequence> then we
are already compatible with the URL/URN character sets.  How to
generate the hex sequence is another headache altogether...

> Perhaps the URL syntax draft, where it speaks about URIs,
> could be generalized in the direction of metaschemes.
> From the syntax viewpoint, we already have two metaschemes
> (urn: and ck:), which can even be combined together,
> as ck:urn:namespace:NSS.

Martin is correct here as well.

Ryan