W3C home > Mailing lists > Public > public-bpwg@w3.org > June 2010

Re: Mandatory conformance statement for CT guidelines

From: Francois Daoust <fd@w3.org>
Date: Mon, 07 Jun 2010 13:54:59 +0200
Message-ID: <4C0CDE13.2010802@w3.org>
To: Eduardo Casais <casays@yahoo.com>
CC: public-bpwg@w3.org
On 06/07/2010 12:27 PM, Eduardo Casais wrote:
> I do not quite get the point of this. Of course the guidelines mandate disclosure of an ICS for those deployments that claim conformance. Those that do not want to claim conformance are not obliged to publish an ICS -- whether they actually conform to the guidelines or not.
>
> The probable scenario regarding this aspect is a CT vendor claiming conformance to a customer (i.e. an operator), but not wanting to make public the corresponding ICS. This is a tantamount to letting some unscrupulous CT vendor install its software and then telling its customer "The deployment conforms to the W3C guidelines. Here is the ICS. Trust us." This would not be acceptable because:
> a) In the absence of a standard, W3C-defined and controlled comprehensive test suite, customers cannot verify the claims of the vendor according to a standard specification -- even while keeping the entire procedure confidential.
> b) In the absence of a published ICS, the community of developers cannot test and verify that the deployment actually fulfils the requirements of the guidelines.

I don't think we disagree. The absence of a test suite developed by the 
group is a problem, and the ICS merely reduces that problem, in that it 
allows the community to check a few things at least. I have no proposal 
here apart from "the group needs to work on a test suite" which doesn't 
seem to meet a lot of enthusiasm. The group resolved last time not to 
bind the Candidate Recommendation exit criteria to the existence of a 
test suite.


On the ICS statement itself, I'm not sure I understand how the change I 
propose weakens what we have, but I may well have missed something. As 
you said, the only use cases that are of interest for the group are 
public claims. For these claims, a public ICS would be mandatory, no 
change here.

The "probable scenario" you mention is indeed not what brought us 
together to work on the guidelines. We surely cannot control the 
deployments of transformation proxies that do not follow the guidelines 
or that follow the guidelines but don't want to make a claim of 
conformance in public, but why prevent the use of the guidelines as a 
contractual basis for negotiation between two parties willing to follow 
them? As soon as the operator say "we've deployed a transcoding solution 
that conforms to the guidelines", the claim becomes public and thus 
becomes invalid without an ICS.

All other W3C standards let implementors claim conformance on a private 
basis, used for B2B. I am not saying this change brings anything for our 
use case, I just think it doesn't affect our use case, allows other uses 
of the spec, and addresses a concern raised during while preparing the 
transition.


>
> In any case, conformant CT disclose their presence in the HTTP header field "Via" -- so what is exactly the objective of denying its existence when it cannot keep it confidential?

The disclosure in the HTTP header field "Via" does not mean "I conform 
to the guidelines". It simply means "I am able to transcode content":
http://www.w3.org/TR/2010/WD-ct-guidelines-20100211/#sec-via-headers

Francois.

>
>
> E.Casais
>
>
>
>
>
>
Received on Monday, 7 June 2010 11:55:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:43:02 UTC