[dustinleblanc]

Its up to the person who wrote the site stylesheet to make it behave in this instance. It's not Safari, or Chrome, or FF, or Dolphin that made me declare that my text should be this big or that my div should be that wide no matter the context. I am capable of writing css that adapts; if I choose not to, my site is less readable. Writing a web browser that makes those decisions for me would make my job of designing the site harder, not easier.

Perhaps an alternative would be some declaration that says "Hey render me this way if you feel like it, but use these all powerful browser supplied styles if I become unreadable"

Some browser's have implemented that reading mode (safari?) and it seems like a cool idea, but I think it only works if you provide really good HTML5 markup.

[jacquesm]

Essentially you are saying that older websites should not be expected to be readable on more modern devices because the burden of change lies with the provider of the style sheet to be compatible with devices that don't exist yet. That's not the intention of the web, it's not a display protocol, it's an information transfer protocol.

if it were a display protocol then we could have kept things a lot simpler, specify display width, height and DPI in the request, get back a compressed image and a bunch of 'hot zones' or 'fields' in response. Such protocols existed and they did not make the cut, the one that did explicitly left the rendering decisions to the client receiving the data.

[dustinleblanc]

No those websites should be expected to be readable at all times, but it up to the provider of the content to make that happen. Old sites read fine, as long as their stylesheet didn't make decisions to try to look a certain way at a certain resolution. If you never made the decision to write your styles a certain way, you don't have to worry about wierd rendering issues. Unless we want to force folks to use Lynx, it is up to the content provider to say what their content looks like. A browser can not predict the particulars of my design well enough to make those decisions for me.

[jacquesm]

You utterly mis-understand the nature of the relationship between the server provided stream of information and the browser.

Websites are not pixel perfect renditions created by dictates issued from servers, they are streams of information decorated with meta-data that give hints on how the information could be rendered.

http://en.wikipedia.org/wiki/Hypertext

> A browser can not predict the particulars of my design well enough to make those decisions for me.

On the contrary, you are not capable of predicting the context in which the information you supply will be used, so less design is better. If you want pixel perfect eye candy write an application but don't use the web, it's lossy by design, for all you know your information will be presented in an audible rather than a readable form, will be printed out on a printer that does not support colour and/or will be read with a device 10,000 years into the future.

Your design is not relevant, the information is what is relevant.

[dustinleblanc]

This is exactly my point, the browser doesn't make me choose comic sans and wingdings. I hold in my hands, the ability to not make a site look like garbage...by using less crazy css. Information is what is relavent, it's the designer's choice to make it look like poop. This is exactly the argument that it's your job as a developer/designer to apply good css. If you do some crazy positioning and lock down your text size so it looks great on a 27inch iMac, its not the browser's fault when it obeys your styles and renders like a turd on an Galaxy S1, 2, or 50 million.

[jekrb]

Are you saying that sites that don't provide a positive experience across screen sizes shouldn't worry? Because browsers aren't supposed to do anything besides present information?

[jacquesm]

Sites don't worry, people do and plenty of sites are divorced from the people that created them. In the interest of keeping that information accessible the onus is first and foremost on the makers of the consumption devices and associated browsers to render that content as accurately as possible. Regressions (where content that rendered just fine on previous generation devices but where future devices for whatever reason fail to display this content) should be avoided where possible.

Note that in the case of content such as 'flash', 'silverlight' and other plug-in related content this is going to be a major problem.

Finally, in the interest of commerce and user experience website owners are free to improve their sites to enhance the experience for their users but they should not be required to get basic functionality working.