Re: Navigator standard change proposal

Le 13 févr. 2014 à 20:09, Predrag Stojadinovic <predrags@nvidia.com> a écrit :
> One example is the Opera 9.8 issue you mentioned, where it simply makes no sense to mess up the navigator.appVersion attribute value in order to accommodate bad sniffers. That decision is doing more harm than good. But again, I digress…

It's not digression. It's exactly the core of the issue. :)
A highly competitive ecosystem partly regulated by market shares. People do not pay their browsers. So the revenue stream is often based on the fact that the product will be used by many people. A direct effect of this is that the browser has to work with Web sites. 

You can spend a lot of times trying to contact Web sites and trying to get them fixed. Opera until last year had a 12 persons team doing that on a daily basis (I was part of it). At Mozilla, we do it now with a team of 4 and a growing community for Firefox OS. Contacting Web sites, fixing UA sniffing frameworks, etc. It is costly, takes a lot of time, and often people don't care. It's not necessary because people are bad programmers, but sometimes they don't realize what they do OR/AND they have a project manager with business constraints. They are dependent of framework which are not up to date. 

On top of that, you can add a daily dose of legacy Web sites which exist but will not be fixed. Ever. That's the business reality. When a browser had a hack to be identified in the user agent string for being identified, it's not with the flower between the teeth. It's always a painful choice, but one which is sometimes necessary to survive: The case of Opera with the version number.

When people are trying to do better "Firefox OS" by simplifying the user agent string and making it meaningful again or IE11 recently. You see plenty of developers arguing that browser vendors are breaking their sniffing algorithms. Search for "ua sniffing" on twitter. 

I perfectly understand where you are coming with the desire to do good things with the feature. But as a matter of fact and with years of experience, we have seen the damage of hooks based on out-of-band of information for user agents. So it's a question of balance. 

You will always find good examples for using user agent sniffing. It's normal.
But I can give you day long examples of Web sites doing the wrong things and not only small Web sites, big high profiles Web sites. Currently it is mostly used for redirection to mobile Web sites, and to deliver tier1 experience. The issue being that even capable browsers do not receive the tier1 experience just because the known company doesn't want to spend time on it. That's the sad reality.

It's even worse than that, because it creates a barrier of entry for new players. Not being in the UA database, not matching the algorithm of the site, the browser is getting the wrong version of the site, or worse being denied the entrance. So bad for the user, so bad for the business of the browser. Only well known products get more market shares. In fact, most of the time, the users think that it is the browser which is crappy, not the Web site. Because the Web site is working on the shiny toy of his friends.


Hope it helps to understand.


-- 
Karl Dubost 🐄
http://www.la-grange.net/karl/

Received on Thursday, 13 February 2014 12:45:18 UTC