RE: Comments on arch doc draft

I agree with almost all of TimB's comments.  I've cut all the ones that I
agree with out, so I can focus on the far smaller set of ones that I'm not
sure about.  That means that I agree with TB1-32 with addtional comments on
18-30.

> TB18. Defined term for "http:" URIs
>
> We use these a lot and talk about them a lot, so let's bless some
> generic term.  Why not just HTTP URI, or possibly HURI or hURI?
>

I agree in general, but it doesn't seem to go far enough.  There are at
least 2 different usages for hURIs, location and identity.  Example,
namespace names are currently for identity, though we are debating whether
to make them for a location.  My concern is that we are using http: URIs,
which look an awful lot like URLs, when sometimes they mean identity.  We
have no classification for http: URIs that are specifically for identity and
not for location (ie dereferencable).  I guess I'd like to see something
like hURIs, hURLs, and hUR?s; where an hURI could be either hURLs or hUR?s.
It's perhaps unfortunate in this regard that the URI spec used the
"identifier" to include both identity and location, as it makes it hard to
talk about identity and not location.  My off the cuff suggestion: hiURIs
(:-) for http identity URIs.  I don't think we should use hURNs to indicate
http: schemed names, because URNs have a particular scheme.

I imagine this notion will spark a fair bit of debate.  I'm sure I will be
informed of the irrationality of my ways ;-)

> TB30. Protocols
>
> Please copy in Orchard's text with what copy-editing is appropriate.
> We've had it for some time, we should use it.
>

Do you think my text on shared information space should go into a more
introductory section instead of here?  And we could use the REST write up in
the protocols section, or the REST write up could go in a more introductory
section as well.

Cheers,
Dave

Received on Thursday, 27 June 2002 18:13:36 UTC