To: uri@bunyip.com
Subject: [m.koster@nexor.co.uk: Re: X.500 external attributes]
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <93Dec22.104722pst.2732@golden.parc.xerox.com>
Date: Wed, 22 Dec 1993 10:47:16 PST
You might be interested in the following discussion, which I think
raises some issues about the line between URLs and URNs, the
extensibility of the name space, etc.
(apologies to those on the wg-nap list for the duplication).
================================================================
To: Tim Howes <tim@terminator.rs.itd.umich.edu>
Cc: Martijn Koster <m.koster@nexor.co.uk>, pays@faugeres.inria.fr,
wg-nap@rare.nl
Subject: Re: X.500 external attributes
In-Reply-To: Your message of "Wed, 22 Dec 1993 11:30:03 EST." <199312221630.LAA23065@terminator.rs.itd.umich.edu>
Date: Wed, 22 Dec 1993 08:45:22 -0800
From: Martijn Koster <m.koster@nexor.co.uk>
> > However, I do have a problem with the current proposals: They are too
> > specific . I'd hate to see a separate "dap" "solo", "ldap" URL's.
> > Which am I going to use the display my phone number?
> > I think it would be far more useful to have an X.500 URL.
> > Compare for example the "news" URL, it is not a "nntp://..." URL.
> > X.500, like Usenet News, and unlike the WWW/gopher models, has global
> > naming already, why limit ourselves to a specific access protocols and
> > host?
>
> > This can be done basically by replacing the ldap://<hostname>:<port>/
> > bit in Tim's proposal by "x500". A browser can then decide which
> > protocol to use, which machine to retrieve it from, and how to display it.
> >
> > If we don't do this we are going to get deifferent references to the same
> > X.500 data in incompatible schemes. For gateways this may be acceptable,
> > but for a URL scheme it isn't, IMHO.
>
> Good point. As long as dap, ldap, and solo all give access to X.500
> and only X.500, this is true. But there may come a day when some things
> are accessible through ldap that are not actually in X.500. They would
> be accessible only through ldap.
>
> Maybe we should define a url for general "x500", as you suggest, but
> also one for ldap (or dap, or solo if needed). This way, things in x500
> would be referenced by an "x500" url, while things only in ldap could
> still be referenced by an "ldap" url.
Agreed. I think we should wait with extra standards until there is a need,
ie until ldap/solo have non-ldap functionality.
> The only problem I see with your proposal for an x500 url (removing the
> host and port) is that it no longer gives a complete picture of how to
> access the resource. It depends on the client to "know" the location of
> some dap/ldap/solo server it can contact to resolve the url. Maybe this
> is not really a problem, but it seems contrary to the way other url's
> work. Is there an example of something else that works like this?
Yes, take my News example above. There is also a mailto: URL that uses
an SMTP server that is externally configured.
It makes sense: Say you give a link to your X.500 info on your WWW page.
You don't want users to use _your_ local LDAP server, but _their_ local
ldap server. Same for News, SMTP.
The reason gopher/http/ftp are different is that they haven't got a global
namespace, there the server/port are very much part of the location, so it
is required info. In X.500 the LDAP gateway is not part of the location of
a DN (I could substitute servers).
> > BTW, is this the right forum to discuss this? What has the URI wg
> > to say about this? They must have touched it..
>
> I don't know. We should probably find out. -- Tim
I think we should cc the uri working group on the X.500-in-a-URL debate,
and leave them out of the URL-in-X.500 debate. Now, where do they live...
-- Martijn
__________
Internet: m.koster@nexor.co.uk
X-400: C=GB; A= ; P=Nexor; O=Nexor; S=koster; I=M
X-500: c=GB@o=NEXOR Ltd@cn=Martijn Koster
WWW: http://web.nexor.co.uk/mak/mak.html