Vodafone Move Invites Web Development Chaos 192
hoagiecat writes "Web developers want mobile phone users to be able to access their sites, but mobile browsers generally choke on heavyweight HTML put together for traditional Web browsers. A host of services have sprung up that allow two sites — one for mobile users, one for PC users — to coexist at the same URL, with the browser's user agent string distinguishing between the two. Vodafone has come at the problem from the other end, offering a new service that translates traditional Web pages into mobile-friendly ones on the fly — but it strips out the user agent in the process, breaking sites designed around the other strategy. And Web developers are mad. Will similar moves by other carriers disrupt this nascent Web development ecosystem?"
Re:User-Agent = breakage (Score:5, Interesting)
I clean the ua's and apply a Jaro Winkler similarity algorithm. This approach results in a 90% successful match, and in the cases where the match is incorrect it return a sibling phone.
As far as the mobile world is concerned UA's are great.
It also breaks the Accept header (Score:2, Interesting)
Revenue stream (Score:4, Interesting)
By breaking the functionality that allows operators to display the mobile optimised pages, they are forcing people to download more content. Even if they only charge for the amount transmitted to the mobile after they've processed it, that's still likely to be significantly more data than people would have had with the optimised pages. And if they charge for the size of the original page (and I wouldn't put it past them), they really are ripping people off.
Either way, I would not be happy with this change if I was on a limited data tarrif.
User-Agent not reliable anyway (Score:2, Interesting)
Other than V managing to content block me from one of my own sites in Australia for a while (even though I have content blocking turned off and there's nothing dodgy about the site) their service seems to work quite well. I don't like stuff being redirected via a 'transparent' proxy, but maybe I could use SSL to prevent that if I really cared I guess.
Rgds
Damon
Re:User-Agent = breakage (Score:3, Interesting)
Otherwise you're just restricting access to a site someone's specifically navigated to. Offering a second site provides all the benefits you mention of image scaling, etc. without this downside. I'm sick of having to change user agent strings just to view websites.
Re:User-Agent = breakage (Score:2, Interesting)
Re:Isn't the real problem... (Score:4, Interesting)
Re:Neat! Can I access the cell-page with a compute (Score:2, Interesting)
If you're using large images for products (which should be the only items that take up more than 30KB per image), then why would you want it smaller anyway? I would rather the full version especially since if I'm loading it that means I want to look at it.
It seems everyone is missing the big picture. Relying on User Agent strings for ANYTHING is like adding concrete to a bridge that's falling apart. If you don't fix it and fix it the correct way, you're going to have tons of patches and eventually the whole thing is going to just fall. The services that make file sizes smaller are also quick fixes that don't really resolve anything and just allows web developers to become more sloppy.
I've been building web sites for over 10 years now and there is no reason you can't have a really complex and graphical website the size of Slashdot or CNN and have the entire page be under 100-150kb. Compression and file formats have come a long way and knowing when to use which one is key.
Branching based on User-Agent is a BAD IDEA (Score:1, Interesting)
If history has shown us anything, it's that the vast majority of "web developers" cannot be trusted with checking the User-Agent header. How many times have those of us who use non-mainstream browsers or platforms seen websites choke because some moron is doing this incorrectly?
Example: My bank's website works with Firefox on Windows and Linux and Mac, but not on any other platform Firefox runs on (like BSD or Solaris). Why? They've hard coded a few Firefox user-agents into their scripts. Result? To check my bank statement I have to find the nearest Win/Linux/Mac box.