Comments on Raman's "On Linking Alternative Representations To Enable Discovery And Publishing"

These are comments on the August 8 2006 draft of "On Linking Alternative 
Representations To Enable Discovery And Publishing"[1]  (note that its 
title page says the date is 01 08 2006, which is the non-US convention for 
dates).

I think it's mostly just fine and ready to go, except for some editorial 
suggestions listed below.  All of these are minor, and I'd leave it to 
Raman to decide which if any to accept. 

The two comments that are a bit more significant are:

Substantive comments
--------------------

Chapter 2.3.1:
Currently says:  "Pointers to alternative content are encoded as link 
elements, and the same mechanism is used within RSS/ATOM feeds to 
advertize permalinks and other pointers to make them discoverable."

First of all, I think you mean same mechanism "as within RSS/ATOM feeds". 
More importantly, this presumes that readers know how RSS and ATOM do 
this.  I'm not sure that's a fair assumption even today, and we don't know 
whether RSS and ATOM will be current technologies through the useful life 
of this finding.  I'd explain in a bit more detail.

Chapter 3:
When creating a multiplicity of URIs for a given canonical resource, 
ensure that the relationship amongst these multiple URIs is captured in a 
machine-readable form.

[I think you mean, make sure you can crawl your way through a fully 
connected link graph.  As written, it seems to imply that all sorts of 
machine readable descriptions of the relationship, including possibly 
nonstandard-format documentation, might be acceptable.  For example, if 
XYZNews documents that:

"To get the French version of any of our articles, append "-French" to the 
URI" isn't what you meant at all. 

Editorial comments:
-------------------

Chapter 1: Some hesitancy about the term "facets" of the Web, which seems 
to overemphasize separation into sub-Webs.

Chap 2.1: 
Current: If publishing URIs for the reesource and its various 
representations
Suggested: If publishing distinct URIs for the reesource and its various 
representations .   [Adds word distinct, fixes spelling of reesource]

Chap 2.1.1: 
Current: Serve a canonical representation of the content at 
http://example.com/ubiquity/resource
Suggested: If no content negotation is employed, serve a canonical 
representation of the content at http://example.com/ubiquity/resource
[Reason: you are actually serving all the representations through 
redirects from the main URI]

Current: Use HTTP content-negotiation, along with the correct HTTP VARY
Suggested: With that same URI, Use HTTP content-negotiation, along with 
the correct HTTP VARY

Current: This is a temporary redirect,  -> This is a temporary redirect; 
the [correct punctation is semicolon+; also added word "the"]

Chap 2.2.1:
Current: Each language version contains pointers to available languages.
Suggested: Each language version contains links to URI's that serve 
representations in all the other available languages. 

Current: Since these translations are typically for human consumption, 
they are encoded as HTML a elements so that they get displayed in 
browsers.
Suggeted: Since these translations are typically for human consumption, 
the links are encoded as HTML a elements so that they get displayed in 
browsers.  [it's the links that are encoded, not the translations I think]






--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------

Received on Tuesday, 29 August 2006 16:52:16 UTC