Discussion in 'XenForo Pre-Sales Questions' started by Puntocom, Aug 28, 2010.
3 Errors, 1 warning(s) checking HTML5 validation at http://validator.w3.org
All related to Facebook HTML. At this point it a case of accept the minor validation errors, or lose Facebook integration.
Surely you can fix these?
The transparency on frameborder thing is part of Facebook's iframe, AFAIK there's nothing we can do about that.
The date format thing is saying that our <input type="date" /> does not have a date entered. This is ridiculous, and I'm confident that the spec and or the validator will be updated to allow empty values at a time in the near future.
The Facebook ones are hosted on Facebook itself, or at least dictated by their code. AFAIK, without going directly to the Facebook dev team (good luck with that!), there's really no way to change it.
Direct link to the results, for those not savvy:
Funny you should mention that, Scott will be at my place on Wednesday, I'll see if I can convince him to get it changed by plying him with chinese food and southern comfort
hahaha. You can do it, I know you can.
Hehehe, see if he can figure out why I can't use my $50 Ads credits. The system goes in a loop.
It's now 5 errors, and 2 are that meta tag:
<meta property="og:site_name" content="XenForo Community" />
I realize people are trying to be helpful, but realize that it is humans that read websites, using web browsers, not HTML validator scripts. I guess there might be a few people browsing XenForo with lynx.
Validation, unlike SEO, does matter for the humans reading it. I agree that some errors are pretty pointless and don't matter (like these damn OpenGraph meta tags), but making sure you're valid ensures that all [standards-compliant] web browsers view your site the same, and also makes it accessible for those who have disabilities.
All of the invalid code shown won't do anything for humans reading it, sorry.
Show me any browser that renders the pages differently after dealing with these particular validation errors described.
Browser-specific CSS (such as border-radius) don't validate in the CSS validation, but we know that the CSS is valid.
Validation is good for checking for errors that you do not know about. Things like these properties I'm sure Kier/Mike were aware of and it doesn't affect the way the page is displayed AFAIK.
I'm not saying THESE error codes will change rendering. From what I understand, his post was bashing validating HTML in general, which is what I was referring to. My post did mention that these errors are pointless.
Should start typing in big bold letters from now on.
Fair enough. I agree with you, standards compliance is important.
Of course, sometimes the W3C just gets it plain wrong. The CSS box model is a prime example. Whoever came up with the the idea that 'width' should not include padding, border and outline was a bumbling fool, and their error has crippled the development of the web. It's not encouraging to see new idiotic errors creeping into specs either, such as the input:date requirement to contain a value. It's just plain wrong and lacks foresight.
I guess I agree. I recently got that "input:date" error and was going insane figuring out why it was an error. While some validation is important to the growth of the web, a lot of what we have now is just a hindrance.
For example, take a look at a lot of major corporation sites that you may visit on a frequent basis. Online stores, communities, anything. Almost none of them are 100% W3C certified (at least in HTML/XHTML). I just checked Wal-Mart for the heck of it. 318 errors, 30 warnings. Site works fine in major browsers, probably works with screen readers and magnifiers, and they even have an auto-detecting mobile site. Works fine from every angle, but doesn't have W3C's seal of approval.
I guess the important thing is to build your site right, and not worry about the validation.
No... Checking for validation is actually quite important, it's just that there are *some* errors that you may have to simply live with. But getting as close as possible to complete validation can make your life a LOT easier, it has a tendency to eliminate a lot of the display quirks that make life hell when you get into cross browser support (with the exception of the pile of POO that is IE6, that's just a black hole of wasted time).
As Kier said, there's a few standards that just don't make sense, but for the most part the validator is a major help if you use it *frequently*.
Sometimes I do think that IE5 got the box-model right. Including the padding and borders into calculating the width of an element makes so much more sense than W3C's way of not including it by default.
CSS Box model tweaking - QuirksMode
Did Internet Explorer get the Box Model Right? - net.tuts+
IE5 got padding right. w3c got it horribly wrong - WebmasterWorld
Validation is not the be all, end all for SEO, as some think it is. Validation is only useful to exact HTML tag specifics, correct nesting, etc. As long as you validate the actual HTML, then in essence validation is true. If you left tags out, incorrectly nested, etc... sure, you will get browser issues.
If you want to live in the 90's with boring to non-existent features, then sure... rely on 100% validation, as you're certainly not going to have anything cool and new within your site.
Separate names with a comma.