W3C home > Mailing lists > Public > public-bpwg-comments@w3.org > July to September 2008

Re: Comments on Content Transformation Guidelines?

From: Luca Passani <passani@eunet.no>
Date: Tue, 05 Aug 2008 23:30:58 +0200
Message-ID: <4898C692.4020909@eunet.no>
To: public-bpwg-comments@w3.org

Sean Owen wrote:
>
> I understand you don't think sites should have to bother. But isn't
> that hurting those very developers in the real world? it's a shame if
> they're being hurt by a transcoder and could solve it with a line of
> markup and you're not helping them figure that out.
>   
so, here is the picture from a developer viewpoint:
- a company called novarra thinks they are so freaking cool they can 
reinvent the mobile web, or even better, declare the end of the mobile 
web as we have known it so far and promote reformatted web pages to the 
rank of mobile web.
- Novarra hires expensively some salesman in the UK to go and play golf 
with the Vodafone UK execs and convince them that they have a solution 
to all of their issues with people not using the mobile web enough.
- from one day to the next, VFUK places Novarra in the middle of all 
HTTP requests. Developers lose money, they complain, and the message 
from the Vodafone-Novarra gang is "change your application. We have 
placed the UA string in a different header and the UAprof we don't give 
a damn about"
- Everyone gets really mad and rightly asks how on earth something so 
abusive and so non-standard can be implemented by an operator.
- Novarra joins this W3C CT task force and gets very close to getting a 
W3C stamp of some kind for their totally hacky product which everyone is 
so mad about.

How can you ask developers to change their applications in this 
condition? what will prevent novarra, vodafone and others from being 
equally abusive in the future? answer: nothing will. We have to stop 
this now.

it is obvious that this is a battle I and others will fight to the end, 
even if I had to go write  VodafoneUK sucks with a spray in London metro 
stations at night

>
> Of course it was a joke. I think this phrase is as silly as anyone.
> Yes it's a profit-seeking corporation, though I do see a lot of
> genuine desire to "do the right thing." But here, thankfully often the
> "right thing" by users is the same as what is right for the business.
> If our transcoder broke out user's experience, they wouldn't use
> Google, and we don't make our ad money.
>
> But we do, so something is right about transcoders -- done right.
>   
Google is *the* search engine. As such, it drives tons of traffic.  This 
explains the money.  It's a bit like microsoft, though. The fact that 
Microsoft made money does not mean that others can do it the same way. 
Anyway, I digress. Google is not an operator and they are not in the 
position to control HTTP. Operators are and they are abusing this 
possibility. It is immoral for W3C to support this.
>
>   
>> so, you are suggesting that every mobile and web site should opt-out of each
>> and every operator trascoder out there, instead?
>>     
>
> Nope. I think they should opt-out once -- with standard mechanisms
> like those HTTP already defines for dealing with transforming proxies.
> That's why I am glad this recommendation exists!
>   
The standard mechanisms are those in the Manifesto. I don't understand 
why it's so hard for W3C to adopt them (particularly because they have 
unconditioned support by developers and operators are using them as the 
requirement for their transcoders)
>
>   
>> the CTG agrees, but it won't agree to spell out clearly "DO NOT CHANGE THE
>> USER-AGENT".
>> Well, you know what, if you change the user agent, mobile sites do not get a
>> chance to even understand that it is a mobile phone they are talking to.
>>     
>
> They are not talking to a mobile phone!! that's the point. Again, see
> my other email about the problem this causes and tell me how
> developers should really handle it. An answer is not "well, I wish
> transcoders weren't here."
>   
Transcoders are a hack, so it is not unreasonable to require that they 
keep using hacks also to preserve mobile site. So, call it a hack if you 
want but add the rule "Do not change the User-Agent string no matter what".

Luca
Received on Tuesday, 5 August 2008 21:31:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:01:50 UTC