Comments on draft 1x of the Guidelines


Thanks for the new draft, Jo!

I do not have any comment on the changes themselves.
I just wanted to reactivate the comments I sent to the list last 
December [1].

These comments are made while reading the spec from a test suite (one 
could perhaps say "algorithmic") perspective, to create conformance 
tests out of it. Here are the comments:

Normative statement in an example
In [2], the "For example" statement contains a normative 
statement. It is correct but it seems awkward to find a normative 
statement in the middle of an example. I think we should make the 
example informative.

Duplicate guideline in 4.1.6 and
In 4.1.6 [3], the normative bullet point:
[[ proxies MUST include a Via HTTP header field]]
is repeated in

No need to repeat the same guideline twice.

"Splitted" guideline between 4.1.5 and
My initial comment was saying that the guidelines on reconstructing the 
original User Agent originated header fields was a duplicate of 
Eduardo pointed out that it is not entirely true. I agree.
I think we should still re-write the statement to action the proxies 
instead of using a passive form. In other words, I suggest something of 
the form: "Proxies must copy verbatim the values of the original User 
Agent originated header fields in the corresponding X-Device header 
field values", possibly completed with the rationale: "to make it 
possible for the server to reconstruct the initial values", and also 
possibly moved to as these two guidelines are highly related.

Could we discuss these points next week?



Received on Friday, 29 January 2010 16:19:59 UTC