Re: [URN] URI documents -- "# fragment"

Sam Sun (ssun@CNRI.Reston.VA.US)
Tue, 27 Jan 1998 10:29:58 -0500

From: "Sam Sun" <ssun@CNRI.Reston.VA.US>
To: "Foteos Macrides" <>
Cc: <uri@Bunyip.Com>, <urn-ietf@Bunyip.Com>
Date: Tue, 27 Jan 1998 10:29:58 -0500
Message-ID: <01bd2b38$6d026b80$29019784@ssun.CNRI.Reston.Va.US>
Subject: Re: [URN] URI documents -- "# fragment"

From: Foteos Macrides <>
>The #fragment was specified as a way of providing instructions to a
>client via a URI-reference, is positional, and is not considered part
>of the URI.

I think it might help to address the terminology first.

If we don't call

a URI or URL, what do we call it?  The draft
call it a URI-reference. But I think most people will call it a URI
(actually URL).

And the mere different between having "#..." or not  is that:

      identifies the file 'filename.html' at ''.

     identifies the section 'section' at ''.

They are both used to location a (portion) of web resource by location.
And I don't see the benefit of defining the additional term 'URI-Reference'
vs 'URI' here.

Further, I think it would help if we address the question that whether the
generic URI parser (which handles URI-Reference, as in HTParse.c
in libwww) should handle the portion "#......" regardless of the URI
scheme or not.

While it makes sense for 'http' URL or 'ftp' URL to cut the "#..." off
at the client side, it's not so appropriate for other kind of URI
schemes (eg. mailto, telnet, ldap, ... ... ... ... ... ... and new ones
are keep coming :).