PDA

View Full Version : Google not valid



shield
10-29-2005, 09:52 AM
You would have thought that a huge site like google.com which receivews millions of searches each day would make their site 100% code compliant...

But no... http://validator.w3.org/check?uri=http%3A%2F%2Fwww.google.com&charset=%28detect+automatically%29&doctype=Inline


Surely theyve got the staff and money to do it...

CannonBallGuy
10-29-2005, 10:12 AM
I totally agree. It's pretty bone-idle of them to be honest.

shield
10-29-2005, 11:44 AM
same goes for Yahoo.com - Their homepage isnt compliant either

martian2k4
10-29-2005, 11:50 AM
Why would they need to do it? What would they gain from it?

shield
10-29-2005, 12:24 PM
Well I dont know the exact figures but lets say for example they get 5 millions searches per day.... Out of those 5 millions there has got to be some accessibility issues with regards to browsers...

Its not guarateed that every single person who visits the site is using a browser capabale of viewing it as it should be...

Nuvo
10-29-2005, 01:35 PM
I noticed this a long time ago, and while it's something I'd like to see rectified, it's a simple case of "if it's not broken, don't fix it".
The Google site is simple enough that it should be usable in most of the major browsers and quite a few lesser ones too.
Since most graphical browsers are based on one of 3 main renderer groups, if it works with one browser, it'll probably display the same in other browsers which use that renderer.
The main rendering groups are:
MSHTML "Trident" or "Tasman" on Mac (MS Internet Explorer), which is available as part of Windows or as an additional tool for Mac (both differ, but they shouldn't differ too much).
Opera "presto" (Opera, Adobe Macromedia Dreamweaver as of MX / V7), which is used by, well, Opera (not available for other developers like IE or Mozilla renderers).
Gecko (Mozilla suite, Firefox and so on and so on...) is the most widely used renderer in terms of how many applications use it.
Gecko can be used without actually having a Mozilla browser installed and it's highly portable with versions on Windows, Linux and other OS's.
KHTML is another one for Linux and other *nix OS's, but it's not as widely used as Gecko and it's not as good :/
Browsers like Lynx don't take much notice of how the page is built as they only display text and a few other elements, which means it'd be like putting opera into text mode :|

A good list of browsers is available at: http://en.wikipedia.org/wiki/List_of_web_browsers and renderers: http://en.wikipedia.org/wiki/List_of_layout_engines
I'm pretty sure Google makes sure that their sites work in all but the most obscure browsers, even if their compliance is poor, and I haven't used a browser which didn't render google correctly yet (the PSP browser doesn't support the AJAX version of Google mail / Gmail, but Google also provide a basic interface that's almost the same, IE, Opera, Mozilla suite and Firefox work as they should).

ryan
09-05-2006, 03:15 AM
google probally dont care because they get enough money anyway

DeluxeNames
09-05-2006, 09:27 AM
same goes for Yahoo.com - Their homepage isnt compliant either

This is very suprising about yahoo and Google. If the big dogs like them don't feel compliance is necessary and yet are still successful, it makes me think they are right. \

Can anyone find a major search portal that is compliant?

Nuvo
09-05-2006, 01:20 PM
It's more a case of it not being an issue when they started out with HTML 4.01 Transitional really.
If they didn't need it in the old days and nobody can challenge their massive cut of the market without being bought out, they don't feel threatened enough to care if their site will please the super validation mongers like me.
MySpace isn't anywhere near valid (you can't make a valid page because the wrapper code is awful), but because they are a big name, it's unlikely that they will take this into account, just like they haven't taken any other flaw in their system that isn't losing them users seriously.
It's a sad fact that most of the internet is using invalid or only partly valid code when you can make fully valid sites, but it's often because people are more bothered about a site looking good than about how it will actually function at code level.
Many pin invalid coding on web scripts like integrated CMS's, but this is a really poor excuse as even dynamically generated code can be 100% valid by the most recent standards (I have a CMS running on my test server which is valid XHTML 1.0 and CSS and it looks decent and uses AJAX in places.

Valid code isn't imperitive to a site's success, but this is no excuse for lax coding abilities, not even where Google is envolved.

xPureNLx
10-12-2006, 12:22 PM
Off-topic: this site is really awsome :O. I've never known there was a site like this. Many thanks Python.

On-topic: Partly agree to Ryan. The only problem huge companies like google and yahoo probably have is that there source is way to big to simply edit + basic scripting languages keep improving and renewing, which basically says the same as my first line: their source is way the big.

WatchOut
10-29-2006, 05:39 AM
Yeah i totally agree

Nuvo
10-30-2006, 04:25 PM
Off-topic: this site is really awsome :O. I've never known there was a site like this. Many thanks Python.

On-topic: Partly agree to Ryan. The only problem huge companies like google and yahoo probably have is that there source is way to big to simply edit + basic scripting languages keep improving and renewing, which basically says the same as my first line: their source is way the big.

Actually, the HTML code wouldn't be overly complex, it's only an issue of wether or not the server side scripting is mixed in with it or not.
If Google's using something like MVC design plans in their applications, there should be minimal logic within the HTML, which would be easy to edit if you're skilled with HTML or XHTML.
Changing <font><% result_count %></font to <span class="result_count"><% result_count %></span> isn't exactly rocket science.
Because of how such pages work, the designer would only need to edit the code used to generate an item listing once as it's used multiple times in a FOR \ FOREACH loop to make life easier.
In the event that Google is mixing logic into HTML, or equally likely, generating all HTML through a script, it's only going to make a ro for their own back should they wish to change anything in the future as it could require a massive recode to do it.