0 error in validation - Is it necessary?

So dividing by zero is OK as long as you get the answer you want?

 

It's more of the difference between theory and application. 99% of the time you should be able to have validated code, most programmers with a depth and breadth of experience would probably tell you that there are times when the desire for validation vs. the need to complete a project and the restriction of resources outweigh this desire. If you haven't experienced that you're luckier, or perhaps smarter than me.

 

User experience comes first. It is better to have 1 error than a site that looks worse.

 

I agree completely. I also think it's important to consider the validation level, if I'm validating against strict xhtml and there's one error that something else relies upon and would require other compromises I think that is acceptable, you still validate as transitional and your code and user experience may come up better. It's also worth considering than a handful of validation errors on a site has no affect on SEO. I really think it's about balance.

 

I agree with that in part. If a deadline requires just getting it done, and it works, then it's alright BUT the intention must be there to come back and fix it. Validation errors only lead to further complications later if ignored.

 

Valid markup never makes a page worse. Invalid markup almost always causes problems later.

 

I definitely believe clean code and markup are essential to a well built web site and for efficient long term maintenance.

 

CSS3 lands it will of course be 100% necessary to be valid
There are some circumstances where you can't avoid having invalid css, maybe just 1 error but still invalid, which is better to have 1 error.

 

That's my rule of thumb too. Not everything can be valid all the time, especially on larger projects. At the end of the day, if the user experience is perfect, the rest is secondary.

 

So MSN.com can't be valid all the time? I mean, it's too big and size matters? Now, granted, this morning they have 1 stupid HTML error but, generally, they have none and you think you and the rest of us can't do it?

 

Validation is like an examination of unnecessary thing. If your site has no errors - good work, but if it does - nothing bad too

 

Some common and accepted code like youtube video embed code makes your site full with errors in validation test.

 

Are you kidding me?

Click on any of the stories and run it on a validator - hundreds of mistakes!

 

But YouTube does not have accepted code; the <embed> tag has never been part of the specs (though I think I heard someone stuck it in the HTML5 proposal, yikes!) and there are many famous (well, I think they're pretty well-known in the web dev world) ways to code around browser issues with <object> tags.
BTW you can employ these methods yourself to put video from YouTube on your site, as far as I know. The methods I'm talking about: Flash Satay, Bye-Bye Embed, or even better, SWFObject (with its downside of Flash-enabled but JS-disabled folks only getting the alt content). Heck, there are even online tools that will take your crappy embed code and make it valid for you! Just copy-pasta! Voila! So while it sucks that YT starts folks out with crapola, that doesn't even matter if you care about validation. It can be done. Same goes for Adobe Flash suggesting target=_blank. I remove it easily.
YouTube will likely never have any compelling reason to have 100% valid code. I'd say this is due to <embed> being rather good cross-browser, stable, and likely to continue working in newer browsers of the future.

This makes no sense: Why and how would having an error make a site look BETTER than a valid site? Oh wait, I can think of one way: the guy who wrote the code wrote it *valid* but *terribly*. It certainly is possible for people who don't know how to write (good) code to come up with valid code (just plug away at those validator errors until you get that green Passed page).
This can only be fixed by MOAR LERNIN' and it comes with time.

However that aside, it's better to have 0 errors and a site that looks better. If you need that 1 error to make the site "look good" then ask yourself what you are doing wrong (this argument is not directed towards those poor souls who are forced to work with others' sh*tty code, such as those forced to work on pages that *must* work with poorly coded widgets, apps, and trackers). I still remember the kludge I had to do to get around an advertisers really really bad terrible smelly invalid-up-to-the-neck VB mixed with Javascript ad-targeting code. The boss insisted on doing business with the company; we were forced to code around that steaming pile. Arg, hate VB code in mah codez.

I suppose one could argue that most of the stories aren't msn.com, but the subdomains, like lifestyle.msn.com or health.msn.com... what's their excuse? The errors are often silly and stupid, something that, seriously, I can't explain where they even came from. What was the point of making the main page valid (except for the one error, which btw is still there and kinda serious/stupid) if every other page on the domain are complete and total car wrecks?

Maybe the plan is, the crap to slowly weed out and make them all valid or nearly so.

Certainly no sane person could believe that these are unavoidable errors, or errors that must simply be lived with. Why? They don't help the site in any way, are a lot of them sure look pretty easy to fix (I dunno what kinda of monster back-end they have, but if it makes valid code difficult then it should be junked, no questions).

 

And do what I said, go to the front page and you'll find zero (or one today). You claim it can't be done. I'll state it can be done cause I do it every day.

 

Yeah and it makes no sense. They're including it because people use it but they are using it erroneously their reason is based on an error. Sometimes I question their sanity and it's not the first time I have. (Mark Pilgrim called me an idiot but Roy Fielding defended me).

 

Ah, okay. I guess I misunderstood. You meant that one should only strive to validate the homepage of a project...

 

I did not say that and I wouldn't.

 

Now that everyone has worn this topic out ….

There is only one circumstance that I can bring to mind where valid markup to a strict DTD is not practical; an <iframe> requires a transitional DTD, and the only reason for an <iframe> is IE's really crappy implementation of <object>.

There is no other excuse.

cheers,

gary

 

Lawlz.

Actually I think they're thinking like CSS2.1, where some of the real-world applications became part of the specs because they worked and some of the other stuff was taken out because it wasn't being used/working. While that's okay, I think real progress was made with <object> and they should have stuck with it. Though IE wasn't the only browser to screw up a bit with objects, it just screws up the most obviously. Did IE8 even fix that?

Mark Pilgrim makes me lol. I read his "why the world will stop and we'll all die if real XHTML enters the web" story which I thought was highly entertaining. Except he forgot the zombies. That would have made it more awesome. Me, I'd be estatic if the whole web died unless it was zero-errors: that's pretty much the only way folks like me can force their bosses to agree to keep sh*t code in the trash where it belongs. It would force software to improve (comment widgets, templates, wysiwyg editors, anything Joes touch). It would bring in a glorious new future filled with rays of sunshine and giant greek statues proudly showing off their godlike muscles and of course the required little multicultural children holding hands and singing kumbaya. And we'd each have our own trained pet monkeys to fetch beers for us.

Actually I'm too cranky and depressive to enjoy a future like that and would prolly have to move to the cesspool of mobile applications instead, and trade my monkey in for something useless, like a guinea pig. Those just sit in their cages and go squeeeee.