Validation

[Medyman]

I was browsing another forum today and there was a dicussion about validation etiquette, among other things. I'm certainly guilty of this and I wanted to get everyone's thoughts on it.

So, when is it proper to validate someone's code? Jeff Croft, a sizeable name in XHTML/CSS, put it this way (http://jeffcroft.com/blog/2008/feb/24/your-markup-validator/). Sometimes people post here asking for help with various aspects of their site, perhaps unrelated to their HTML at all. Just the other day I told someone that they should be using standards compliant techniques in a thread about a PHP-related problem. In the context of a web design/technology related forum such as Dynamic Drive, maybe it's ok.

So what do you guys think? Is it rude to inform someone that they're using invalid code? To what extent is validation even required? Is validation, in your opinion, a sign of quality? After all, we all have our own preferred techniques and workflows and coding styles and validation guarantees very little in the end.

[Xenocide001]

well i'm newb but i think that a validator comes useful when you use it to make progress and help others
recognizing errors that you had before, not to be a jerk/self proclamed "god of code"

[Medyman]

I didn't mean that you would point out coding flaws in an uncouth manner. That wouldn't be right under any circumstance. The question is, if someone asks a question asks a non-HTML/CSS related question, how appropriate is it for us to point out validation flaws in their markup?

The essential question is what does valid code reflect about a website and how pertinent is it to its success? Is it enough that the markup works? Or does the markup have to follow the standards to the letter?

There really isn't any real answer to this, I suppose. Just something I was thinking about.

[djr33]

I don't like validation in theory-- just whining to someone that their code is invalid seems petty.

However, if this lack of validation and errors on the page are tied together, then I think you have every right to tell them, and they should want you to, if you, as the visitor, are put off from viewing their page again.

[rangana]

I believe it's not rude to inform someone that his codes are invalid, cause i've encountered a lot of times that most of the errors are from the errors shown the validator (http://validator.w3.org).

This link had been giving tough inputs too:
http://diveintomark.org/archives/2003/05/05/why_we_wont_help_you

...I don't see validation as a requirement, but then in this big world of web designing, it's good to set standards...and the world wide consortium stands to standardize the codes :)

...I'm not an avid fan of validation, but then, I see that in this vast changing field that we are in (web designing). Coding basing with the standards is a good way of sharpening your wits with what's instore for the next standard. May it 5.0 for HTML or 2.0 for XHTML.

...Anyway, just my 2cents worth ;)

[boxxertrumps]

while having invalid code, you have to test in each browser and fix each error without breaking the markup for other browsers.
if you start out with valid code you can assume your web pages look nearly identical in most major web browsers.
if someone gives me advice, trying to make my life easier in the long run, i would never call them rude.

[djr33]

If someone chooses to have entirely invalid code that works and causes me no problems, I don't see any reason to complain, and as I said, if you do, you're just being annoying.

BUT if there ARE problems, then I think recommending valid code is quite reasonable. But only if there are problems.

[Xenocide001]

well.. if the code runs out of problems i guess it would be kind of rude to tell them they code is invalid.... personally i would be glad if someone tell me my code.. or even my english is wrong and help me fix it... that's a way to learn and update knowledge.. and transmit it

[molendijk]

The sole purpose of valid code should be to not have to test in all browsers if everything is functioning well on your site. It's a means, not a goal. The rest in totally irrelevant.
---
Arie M.

[boogyman]

The sole purpose of valid code should be to not have to test in all browsers if everything is functioning well on your site. It's a means, not a goal. The rest in totally irrelevant.
---
Arie M.

unless the browser doesn't support the standards the code validators are based upon (cough cough IE), but yes very good observation

[hyk]

Originally Posted by molendijk
The sole purpose of valid code should be to not have to test in all browsers if everything is functioning well on your site. It's a means, not a goal. The rest in totally irrelevant.


that a excellent answer, I totally supported it * raising both hands & legs* :D

Boggyman:
unless the browser doesn't support the standards the code validators are based upon (cough cough IE), but yes very good observation

Is IE very eh troublesome?

[djr33]

And IE is exactly why the valid code is absolutely unimportant if the code works in every browser-- "valid" code in many cases WON'T work in IE (and maybe some other browsers, but certainly IE is the worst).

So, yep, a tool only.

[hyk]

I am just wondering. IE, firefox etc are all used to surf the internet,i suppose the infrastructure/system(not too sure what you experts called it) should be the same since they serve the same purpose.why is that all the browsers work fine expect IE? did microsoft used a cheaper way to create IE :D

[jscheuer1]

I am just wondering. IE, firefox etc are all used to surf the internet,i suppose the infrastructure/system(not too sure what you experts called it) should be the same since they serve the same purpose.why is that all the browsers work fine expect IE? did microsoft used a cheaper way to create IE :D

That's not quite the situation. There was a time when IE was actually the closest thing to a standards compliant browser, admittedly with a number of proprietary add-ons that tend to confuse the issue of compliance. This was around version 5.5. But then, while IE (relatively speaking) stood still in that regard, others caught up and surpassed it. In IE 7, MS did a lot of catching up, but there are still problems, as there are in virtually all of the others.

No browser I am aware of is 100% standards compliant, though most modern versions of Opera, FF, Safari, etc. are pretty good. IE 8, if you go over the advance material for its rendering engine will be like 99.99% if not 100% standards compliant, it will however also still have backward compliance for as many of the quirky features of the installed base of web pages out there as possible.

In short, none of us has had the experience of coding for a 100% standards compliant browser, unless we've worked with one that isn't in wide release.

So we are often relying on the quirks of browsers, and their error correcting routines, which can only correct to the level of compliance built into the browser. It's just that most of the browsers are similar, except for IE, which to be fair isn't really like being from another planet, more like being from the other side of town.