Re: "resource locator strings"

Mark Nottingham wrote:
> 
> I'd be interested to see what URI people think of:
>   http://www.gnu.org/projects/dotgnu/spec/rls.htm

I would suggest that the syntax be modified to fit
into the URI generic syntax[1].

for example,

   dotgnu.org:spell:"I wnt to chke my, speeling"

becomes

  rsl://dotgnu.org:spell:"I%20wnt%20to%20chke%20my,%20speeling"


I would suggest it directly to the authors of the rls.htm
document if that document made their email addresses handy ;-)
I'd appreciate it if anybody who is in contact with them
would pass my suggestion along to them.

I was briefed on some aspects of the .NET design by
one of the designers a while ago; when he got to the
part where one "assembly" points to another, I immediately
asked if they're using URIs for that. He reported, with regret,
that the pointer is just a string. That leaves open the possibility
of using URIs, but does not mandate it. He had argued to
use URIs explicitly, but accepted a decision to use plain strings.

I think the .NET design has a lot of good features... the
Modula-3 guys got a lot of things right; Java borrowed some,
but .NET borrows even more of the right things.

.NET seems to be less constraining than Java and Corba
in some important ways. (in particular, in the
design of the root of the object system, which see
my unfinished article
  "Corba is not Mimally Constraining" 
  $Id: oop-min.html,v 1.4 1999/10/02 13:35:20 connolly Exp $ 
  http://www.w3.org/Architecture/1998/12/oop-min.html
and some nearby notes
  The Web Object Model and Type System
  http://www.w3.org/DesignIssues/DistObjApps
)

Hence I'm encouraged to see .NET go open source.
But I hope that the result will be integrated with URIs
and the Web, not in competition.


[1] August 1998 
     Uniform Resource Identifiers (URI): Generic Syntax
     (RFC 2396) T. Berners-Lee, R. Fielding, L. Masinter 
http://www.ietf.org/rfc/rfc2396.txt

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 17 July 2001 13:28:54 UTC