Forum Moderators: martinibuster
Use only 1 search ?
I do, apart from my "search results" page where I have two. But if you are allowed two, then google should have tested it for such things.
Ask for permission to change the code so it validates (they -used to- give that relatively easy)
LOL, You're funny.
I contacted them about two months ago asking permission to alter the code for something similar. I first received the world famous automated email that tells you to check the (spammy) forums.
Then, after a few days I received a "How did we do?" email. Needless to say I didn't give them a good score.
Not heard back since. Fine customer service they turned out to be. Not.
Very few pages on the internet actually validate, especially the larger sites. google dot com returns 65 Errors, 8 warning(s) on the w3 validator. I guess they just don't take it very seriously. If Google doesn't care, why should you?
Because I take pride in my work. Out of the 10000 page site, there are about 3 pages that do not validate. Two of them are google result pages.
Validation is your only weapon in a fight to get your site to render properly on all browsers out there.
Nope. Your only weapon is called testing. I've had pages that validate render differently on different browsers.
Running a validation program is only one sanity check. If something doesn't validate, I check it out, but I don't fix it if it ain't really broke. It's like getting warnings back from a compiler, it is simply telling you that something might be wrong.
Is that what you'd tell a client or someone interviewing you for a job?! Yikes.
Yeah, I would. I would also explain why they shouldn't care. Being willing to tell them such things, and being able to explain why, is the reason that I was usually the highest paid contract programmer at those companies.
[edited by: Atomic at 7:25 am (utc) on Jan. 19, 2009]
...Google as a good example of why it's not that vital to have your pages validate.
No, I wouldn't say that to a client because the client might be ignorant of the reality and explaining reality to prospective clients is a fool's errand. That's one reason why I'd never design web sites for other people--or be in sales for that matter.
But if I were applying for a web-related job I definitely WOULD express such an opinion in a job interview because working for an employer who was ignorant enough to hold this against me would be an ill-fated venture, I'd be quitting or get fired somewhere down the line anyway.
[edited by: OnlyToday at 2:14 pm (utc) on Jan. 19, 2009]
As we are not allowed to alter the code, what are we to do ?
Sounds like a pretty simple decision. Which is most important to you? Having two search boxes or having your code validate? The choice is yours.
having the final result validate is not necessary
It would matter a great deal exactly WHY one's page did not pass validation.
Some validation errors cause no known harm except to annoy the purists, but it's possible for the wrong kind of error to cause major problems for some users, including to stop a spider in its tracks.
It is vital to understand the difference, and clean up accordingly.
No, you could use Google as a good example of why it's not that vital to have your pages validate.
Just because something is not *vital* doesn't mean we shouldn't at least try to do it. It's not vital to comment your code when programming but it's a good habit. A page that normally validates makes it a lot easier to spot problems or errors if it suddenly stops being validated. Google has bigger issues than validation, like trying to compress their html files as much as possible (they avoid quotes in attributes for this reason) since every byte cost them thousands. If it wasn't for some of these special needs, I'm sure they'd try to validate their pages too.
If I modify my pages and I've left out the body tags I'll want to know about it. The W3C validator is handy for that. If I had 176 errors already I might not notice the other one or two that might be something important.
The old custom search engine didn't validate either so already I have 2 errors. Those are annoying.
I don't care if Google validates, but if they release scripts for me to use I do care that those validate.
That's why I personally gave up trying to validate a long time ago.
It's not that hard to just learn to code HTML properly, after that, the validation error is the exception, not the rule, which helps you spot problems on your page.
I learned to accept that Google Search will produce errors on my page (along with some affiliate links and banner ads) and I don't mind, although if Google could do better, I'd be happy.
To say you won't try to have valid html code because the big boys aren't either, well that's like saying you won't try to lose weight because all the big shot CEOs also have extra pounds anyway. Do it for your own good, not to emulate the big companies.
Validation is your only weapon in a fight to get your site to render properly on all browsers out there.Nope. Your only weapon is called testing. I've had pages that validate render differently on different browsers.
Running a validation program is only one sanity check. If something doesn't validate, I check it out, but I don't fix it if it ain't really broke. It's like getting warnings back from a compiler, it is simply telling you that something might be wrong.
So that means you feel you can test in all browsers ?
E.g. you have all models of all phones in active use around the world operating ?
I'm sure you can also test future browsers ;-)
I'm not just talking about the handful of browsers on computers (including the poor excuse for a browser called MSIE), *all* of them is a _lot_ harder.
E.g. if you're in the US or Europe or most of the rest of the world, try testing your site on a Japanese iMode phone... you don't get to buy them, and don't have a network to test them on.
So why can big names like Google do this differently ? Well first they have a global presence, far larger budgets and most importantly browser makers will test their contraptions against sites like Google's (not against our stuff)
Really testing is useful, but it's a small part if you want it to work always, validating code is also not going to become less important (xhtml content e.g. should not be rendered at all if not valid)
I do not worry that a tiny percentage of my visitors may view a slightly less aestheticly pleasing layout.
That said, I spend the greatest portion of my scarce time working on making my actual content more valuable.
IOW, fussing with validation is a low ROI activity.
Validation has to take a back seat.
No it doesn't. In fact, I'm not going to sit there and track which errors would be allowed because someone doesn't have 5-10 minutes to do some minor repairs. Nope, that won't happen. It ain't no a la carte proposition either. Do it, or sign a waiver that I am no longer responsible for anything that may occur from this point forward. :)
Google, how about just fixing the borken code and we can be done with it? Ya'll should know better than to use that ID name more than once, it is one of the first errors many will encounter and probably one of the easier ones to fix.
For now, you're best bet is to run one CSE until they fix it which I'm sure they will. Wait, I should say, I hope they will after reading this topic. All it takes is a topic to make WebmasterWorld Home Page and things start happening quickly. It has already started making its rounds through the community and Google surely doesn't want the negative press. ;)