Re: Changing User Agent Header Value (was Re: transcoders bad)

Jo, I think that what you write is not consistent with the fact that the 
Manifesto exists, it comes from developers and that it is widely adopted 
by everyone in the industry (including operators!).
Now, if you disagree with this, let's talk about the Manifesto's 
success, but if you don't, please see in-line:
> > well, given the current status, I think it is better discard the CTGs
> > completely.  Here are the shortcomings of the CTGs:
>
> That's not a conclusion I draw from this exchange. In my view it is 
> not possible to write any set of guidelines that do not draw criticism 
> from some quarters in some respects.
wrong. The Manifesto is a set of guidelines and it has met very little 
public criticism (in fact, most public criticism has come from you and 
Sean Owen). If you know of a different specimen of criticism to the 
Manifesto, please send a pointer...

>
> 1. Many of us can think of ways in which the world we live could be 
> improved. The fact that we can't stop people doing things we don't 
> like is, well, a fact of life.
The Manifesto did stop operators from installing transcoder abusively, 
so I am happy that, in this particular case, wrongdoers were effectively 
stopped and a lot of wrongdoing was pre-empted.
> The fact that CT guidelines don't have some kind of legislative force 
> is not a criticism and is no different to anything else on the Web or 
> Internet.
The Manifesto has no legislative force either, but letting operators 
know that they may enrage every single mobile developer in their 
ecosystem has been a strong deterrent.
>
> 2. Given the opportunity for "an improvement to some degree" vs 
> "stalemate" most of us would choose the former.
The Manifesto has been an imporvement to a very large degree and it took 
a lot less time than CTGs are taking.
>
> 3. Nothing is for ever. Improvement now can be followed by improvement 
> later.
correct. Infact, I am playing with the idea of having a release two of 
the Manifesto (explicitly stopping abuses which we did not quite foresee 
because they were unthinkable: breaking HTTPS, adding extra ads, adding 
operator navbars,...)
>
> 4. In order to move things forward now, compromise, however hard, is 
> needed. CT vendors have compromised in agreeing the CT guidelines as 
> they stand.
CT vendors have agreed to the Manifesto and even signed it. Those who 
have not are telling operators (their customers) that they can support 
the Manifesto rules.

> 5. Improvements have been offered on this list and I hope will 
> continue to be offered which on the face of it seem likely to be 
> accepted.
I am curious to see the WG's decisions, because right now, I see a 
paradox: the world is shouting that the UA string should not be broken 
and HTTPS pages should not be reformatted, and you and Sean are arguing 
with against them with the most fanciful arguments possible. Did you 
guys consider a future in politics?
>
> 6. Some aspects like "user can choose a transformed desktop 
> experience", "this must not be the default experience" and "servers 
> that don't have to accept the deal don't have to", are actually in my 
> view not that hard to run with.
as long as you don't make the "economics" of reformatting part of your 
spec, your rules will always be so wide that abusive behavior by 
transcoders is possible. Just ask yourself: How do I make a rule that 
avoids that VodafoneUK launches novarra the way they did last year AND 
claim CTGs compliance at the same time?
> The spec specifically tries makes sure that the user gets to see the 
> server's choice of experience, means that developers ought easily to 
> be able to show that their experiences far out-perform transformed 
> desktop experiences. Let's rise to that challenge.
"let's"? are you a mobile developer? Mobile development is already hard 
enough that you can't expect developers from around the planet to go 
after each and every carrier around the planet again. This would be the 
final blow on top of device fragmentation.
>
> 7. There is nothing any one can do to force malicious people to "do 
> the right thing". However the CT Guidelines provide a framework within 
> which testing can be carried out and questions can be asked about the 
> interpretation.
I am sorry, but what is the point of releasing guidelines which beg for 
interpretation? once again, look at the Manifesto, there is very little 
to interpret there. Everyone, developers and CT vendors alike, got it 
and implemented it.
>
> I can't speak for what the BP Group will do about a conformance test 
> suite but my opinion is that such a suite would be very useful and all 
> the SHOULD clauses, when not met, need to have a proper justification 
> in such a suite.
the way it works today, I an see Novarra Vision (as launched by VodaUK) 
to be fully compliant to CTGs. This is not acceptable.
>
> 8. It's actually the deployments that we should be interested in, not 
> the products. Network operators need to be convinced that their 
> interests are served by the products they deploy.
The Manifesto already achieves this.
>
> 9. A lack of consensus of any kind in the industry means that 
> operators will continue to deploy regardless.
Again, there is consensus around the Manifesto
> Industry agreement about the parts that can be agreed upon is 
> therefore essential, in order to avoid throwing out the baby with the 
> bath water. Saying that it should be discarded completely is, well, a 
> guarantee that nothing you want will be achieved.
The Manifesto has already achieved much better than the CTGs will ever 
achieve in its wildest dreams.
>
> A lot of useful things have come out of the discussion on this list 
> over the last few days. Call me hopelessly optimistic, if you want, 
> but I think that technical disagreements are reduced to quite a small 
> level.
yes, you are hopelessly optimistic.
>
> What we are left with is an opportunity to move things forward, to 
> nobody's complete satisfaction but to the benefit of everyone.
What you are left with is an opportunity to bring confusion to the 
industry and give abusive transcoder vendors a leaf fig.
Also, if CTGs are released officially in this condition, I will have no 
choice but to blog hard against them.

Luca

>
> Jo
>
>
> On 06/08/2008 22:38, Luca Passani wrote:
>>
>> Sean Owen wrote:
>>> I think one of Jo's points is that the W3C is not the government or
>>> any kind of enforcement agency. Transcoder vendors can do whatever
>>> they like, period, regardless of what any of us write. One can only
>>> recommend.
>>>   
>> well, given the current status, I think it is better discard the CTGs 
>> completely. Here are the shortcomings of the CTGs:
>>
>> - as you say, CTGs are not binding for transcoders (no W3C police 
>> around)
>> - CTGs compliance does not bring enough protection for content owners 
>> (as the discussion you have triggered on WMLProgramming is 
>> demonstrating)
>> - CTGs can still be used by vendors and operators to justify their 
>> totally non-standard practices and abusive business practices
>> - CTGs conflict in some important parts with the Manifesto for 
>> Responsible Reformatting (which already has incredibly wide industry 
>> support by the developer community, which has already been adopted by 
>> key transcoder vendors such Infogin and Openwave, and is being used 
>> by operators across the globe as the basis for their requirements for 
>> transcoding)
>>
>> I realize it is a pity to discard all the work done so far, but why 
>> struggle for a different balance than what already achieved through 
>> the bloodshed we witnessed last April?
>>> How about this part?
>>> http://www.w3.org/TR/2008/WD-ct-guidelines-20080801/#sec-request-no-transform 
>>>
>>>  A transcoder that follows this recommendation does not force
>>> transcoding on all requests, which sounds like something everyone
>>> agrees on.
>>>   
>> I still think CTGs should simply be discarded, but if you really want 
>> to write something I am happy with, what about?
>>
>> "Network operators should not install transcoders as the default 
>> gateways across which all standard WAP/WEB traffic is routed"
>>
>> Luca
>>> This is the place to submit a proposed change, perhaps related to
>>> these sections.
>>>   On Wed, Aug 6, 2008 at 4:50 PM, Luca Passani <passani@eunet.no> 
>>> wrote:
>>>  
>>>> the problem is that the rule is big enough that transcoder vendors 
>>>> can run a
>>>> train through it. They just need to claim that it's "full web on a 
>>>> mobile
>>>> phone" they are launching, and there you go, everyone gets 
>>>> trascoded (this
>>>> is exactly what VodaUK did, by the way).
>>>>
>>>> What about having a rule that says that Network operators are not 
>>>> supposed
>>>> to make transcoders manage all HTTP requests for their main/default 
>>>> WAP
>>>> configuration on devices?
>>>>
>>>> How do I submit a proposed change?
>>>>     
>>>
>>>
>>>   
>>
>>
>
>

Received on Thursday, 7 August 2008 08:29:04 UTC