Is validation important?

[Tim]

I've been briefly debating with Mike and Jabber (Notes on new design) about the importance, or not, of valid HTML. I don't want to distract from the main thread of that post but I do think the topic warrants further discussion and would like to continue and get feedback here.

As I see it, clean/valid code is important for several reasons:

• Valid code reduces incorrect rendering and browser nuances

• Valid code makes for more accessible pages for visually impaired people

• Valid code minimises loading time

Mike and Jabber seem to think validation isn't very important provided that pages render well for sighted people in the top 2 or 3 browsers. Mike thinks validation is more about giving the developer a 'warm glowy feeling' as they see the W3C thumbs up. Jabber thinks that 'Validation can be quite restrictive to design' and that 'the prequisite is that the browsers display the pages properly and that we don't get penalised by the search engines for non-validating code'. Mike thinks that because mainstream sites like Yahoo, MSN, Google, Amazon, Dell have huge numbers of errors in their pages, it's fine for the rest of us to make non-valid pages.

My contention is that errors are just sloppy coding. Most people use a spell checker and, after reviewing, trust the results, so why not an HTML checker that is much more accurate? Curious.

If I paid a proffesional: electrician, plumber, copywriter, carpet layer, gas installer, child minder or accountant etc. I would not be too pleased if they told me that they had made loads of mistakes but it doesn't really matter. Why should it be different for web developers?

I don't know of any valid HTML/XHTML code that requires a non-valid variant to correct an issue in any browser or platform. Of course, there are some valid codes that IE doesn't recognise and they require bug fixes but these are often cured by separate style sheets or minor (and irritating) tinkering with spacing to account for variable text-size rendering for instance. Equally, there are proprietry IE styles that do not render in other browsers at all because they are not valid and my advice is stay clear of them.

During the browser wars between Netscape and Explorer both browsers developed tolerances, difference-over-lookers or 'bug' fixes, for the other's interpretations of some HTML. I contend that it is largely because those browsers have built-in tolerances that developers today get the impression that it is OK if they make mistakes. JavaScript and Perl coders, amongst many others, would surely be envious of the flexibility.

At last there is an apparent agreement by the browser programmers to comply with W3C recommendations. This is something to celebrate. All but IE6 are in compliance and that browser really has become the new NS4. Hopefully IE7 will be a lot better, we'll find out later this year.

So why don't the mainstream sites validate? The answer is because like all of us they don't have to. W3C is an advisory body not a law maker, hence 'recommendations'. Imagine what would happen if Google, Yahoo and MSN decided to add validation to their algorithms and that only valid and accessible sites ranked highly... everyone would clean up their code very quickly. Funny how self interest works!

Now that the browsers have reached a level of agreement, developers can focus on design and provided that the code is clean, pages should look more or less the same in all browsers and platforms. No more weak excuses for sloppy coding. Surely, it should only be a matter of time before the search engines catch up.

What do you think?

[sands]

Read: http://www.thenoodleincident.com/tut...alidation.html

I contend that it is largely because those browsers have built-in tolerances that developers today get the impression that it is OK if they make mistakes. JavaScript and Perl coders, amongst many others, would surely be envious of the flexibility.

Absolutely. I think that if you are a coder, validating pages wouldn't seem like a burden to you. Valid code is one way to prepare for the future.

Surely, it should only be a matter of time before the search engines catch up.

True, and now it is becoming increasingly common to have folks ask for their sites to have valid code. But it is a matter of choice. Best not to knock people on the head and lecture them about valid code. Do it and care not what others do.

[sands]

Also if we were to keep aside for a moment the merits or need for valid code, at least no one can say that validating code is harmful in any way, can they? But not validating code could lead to problems. So the best bet would be to validate.

[Jabber_uk]

Thanks Tim - a great argument and well put too!

As I think you may have realised already, for me validated code is something to work towards not a be all and end all.

The problem I have with the whole validation thing is the importance placed on validation by the big players in the industry:

The biggest sites on the Internet do not validate including www.microsoft.com; www.netscape.com; www.yahoo.com; www.google.com (omg the page only has one image, a search form and a dozen or so links!). Anyway we could go on and on about that. I commend www.firefox.com and www.opera.com of course :-)

I should also point out that web page creation programs such as Dreamweaver allow and encourage non validating HTML (such as the height property of tables).

Some reasons for being lazy (as Tim might call it) are down to the immensity of the field and the incredible variation in the ways we can create our product (skills and programs used) and the way the product is interpretated. As a for instance:

Take my situation years ago - I didn't know much about web sites or the Internet and I started creating websites with Netobjects Fusion (I think version 3 or 4 when I started). So I happily created pages, I had no idea about what proper code was or about validation, I assumed (wrongly) that the program I was using would be able to deal with all that scary code type stuff. One of the first sites I created (http://www.creativetheatre.co.uk/) is still running to this day but it does not validate!

Imagine the uproar if Google decided to drop non validating websites from it's index! It would have to clean up its own code first!

The bottom line is these are recommendations from W3c only. You cannot compare these recommendations to rules and regulations regarding plumbing, electrical work, Gas installers, builders etc because those are more than recommendations, they are there for the safety of the user. There might be recommendations on how an electrician does his job too and they would be just that - recommendations.

From the HSE (The Health and Safety Executive of the UK):

Note: The inspection and testing intervals for electrical equipment given in these
publications are recommendations and are not legal requirements.

Taken from the HSE booklet "Electrical Safety and You"

As you can see even something as important as safety of electrical equipment can have recommendations only.

Finally, because I feel i am going on a bit:

In principle I agree with Tim in that it would be good to create validated websites, but in reality this cannot and perhaps should not be the most important thing....?

[Jabber_uk]

Originally Posted by Tim

Valid code reduces incorrect rendering and browser nuances

You are probably right 'reduces' being the operative word ;-)

Originally Posted by Tim

Valid code makes for more accessible pages for visually impaired people

Perhaps you are right again but, If I have a page that uses the height attribute in a table, how does that adversely affect visually impaired people from seeing the site properly (I genuinely want to know).

Originally Posted by Tim

Valid code minimises loading time

Correct (for the most part) but load time depends on so many other aspects. One web designer could create a site in HTML with embedded styles and javascript while another uses links to the external stylesheets and javascript files. Which one loads faster? Probably the one with external styles and javascript. But it could be that the one with internal code is the validating one! In which case it probably takes longer to load (splitting hairs I know) :-)

[mikesmith76]

If I paid a proffesional: electrician, plumber, copywriter, carpet layer, gas installer, child minder or accountant etc. I would not be too pleased if they told me that they had made loads of mistakes but it doesn't really matter. Why should it be different for web developers?

I have to agree with this comment. While valid HTML is only one factor to consider when creating a website I think it's a matter of professional pride to make sure a site validates. If an electrician left a cable running across the floor your tv may still work, but it isn't a professional job.

However as stated by previous posters there may be times when you need to use a feature that isn't allowed by the DTD you are using. Do you want to rewrite the page from scratch or should a few mistakes be ignored? I don't think valid html is the most important factor when developing a website but it should be a serious consideration

[frogmanandy]

I just want to remind you ... most people i know say that we write mark-up or at best script ... coding is for the people writng oop or better ... just a bit of levity here ... smiles

I think maybe I should add ... valid mark-up ie w3c valid html is not code or coding in these people's eyes

[Narasinha]

Originally Posted by Tim

My contention is that errors are just sloppy coding. Most people use a spell checker and, after reviewing, trust the results, so why not an HTML checker that is much more accurate? Curious.

If I paid a proffesional: electrician, plumber, copywriter, carpet layer, gas installer, child minder or accountant etc. I would not be too pleased if they told me that they had made loads of mistakes but it doesn't really matter. Why should it be different for web developers?

Great work, Tim. This is often the analogy that I use to impress upon people the importance of complying with web standards.

My post from back in November of 2003, Creating Standards-Compliant HTML and XHTML, is one of several where I explained to people why they should try to comply with current standards.

In a perfect world, we wouldn't be continuing to have these conversations, but, alas, it's definitely not a perfect world. Many people are still turning out non-standard code. WYSIWYG editors still turn out non-standard code. Some of us, though, enjoy creating fine, hand-crafted XHTML/HTML.

Browsers can be like an old television set: very forgiving. Even with a poor signal you can still see a picture. But browsers, like televisions, continue to evolve. Will the next version of browser X display non-standard code properly? If you leave the closing tag off of a paragraph, will the paragraph perhaps not show up at all? Who can say? Crafting your code according to the appropriate standards will help to ensure that your site will still be as readable by the browsers of 2012 as they are by today's browsers.

Best Regards,
Dan

Next time can we start some debates about Accessibility? How about CSS and separating content from visual display?

[Tim]

Originally Posted by Jabber_uk

If I have a page that uses the height attribute in a table, how does that adversely affect visually impaired people from seeing the site properly (I genuinely want to know).

I don't think using the height attribute would affect a visually impaired person at all, someone please correct me if I'm wrong. I'm mainly referring to missing alt, title, longdesc, label and table summary tags. There's also tab index which to be honest I think I've only used once but I don't believe that's in the W3C list.

Originally Posted by sands

Best not to knock people on the head and lecture them about valid code. Do it and care not what others do.

I'm sorry if I came over as lecturing or judgemental. That wasn't my intention. I just want to advocate best practise. What's the point of the browsers getting their act together if we, the suppliers of their 'meat' don't 'sing from the same hymn book' as Jabber might say? Sorry for the strange mix of metaphors!

[mike]

Originally Posted by Tim

I've been briefly debating with Mike and Jabber (Notes on new design) about the importance, or not, of valid HTML. I don't want to distract from the main thread of that post but I do think the topic warrants further discussion and would like to continue and get feedback here.

Cool.

Originally Posted by Tim

Mike thinks validation is more about giving the developer a 'warm glowy feeling' as they see the W3C thumbs up.

Well, in all fairness, that does seem to be something of an oversimplifaction. The warm glowy thing was, after all, just a guess.

Originally Posted by Tim

Jabber thinks that 'Validation can be quite restrictive to design' and that 'the prequisite is that the browsers display the pages properly and that we don't get penalised by the search engines for non-validating code'.

I'd put myself more in this camp if I were picking a camp. I'm very much a bottom-line kind of guy. 'At the end of the day, what are we looking at here' is one of my personal mantras

Originally Posted by Tim

Mike thinks that because mainstream sites like Yahoo, MSN, Google, Amazon, Dell have huge numbers of errors in their pages, it's fine for the rest of us to make non-valid pages.

I didn't say it's ok because they do it. I said they don't care and offered as evidence of their indifference towards the subject the fact that their stuff won't validate either.

Originally Posted by Tim

My contention is that errors are just sloppy coding. Most people use a spell checker and, after reviewing, trust the results, so why not an HTML checker that is much more accurate? Curious.

I don't think that errors are simply 'sloppy coding'. Though I'm satisfied that some errors are the result of sloppy coding, I would submit that the vast majority of these things are more a matter of practicality than laziness, ennui, or ineptitude.

If I want my page to look like 'this' and behave like 'that' and I can do it in 2 hours if I'm not concerned about validation or 8 hours with validation, for me, it's an easy call.

Originally Posted by Tim

If I paid a proffesional: electrician, plumber, copywriter, carpet layer, gas installer, child minder or accountant etc. I would not be too pleased if they told me that they had made loads of mistakes but it doesn't really matter. Why should it be different for web developers?

I think this comparison relies on the supposition that code that doesn't validate is a 'mistake'. I don't necessarily agree with that assessment. These are,as has been observed, guidelines not rules, correct?

If I'm washing my hair (snicker) and I fail to rinse and repeat (as necessary) I'm not making a mistake if I don't rinse and repeat, I'm just not adhering to the guidelines. Does that mean my hair would be dirty or that I 'washed it wrong' since I didn't rinse and repeat?

Originally Posted by Tim

At last there is an apparent agreement by the browser programmers to comply with W3C recommendations. This is something to celebrate. All but IE6 are in compliance and that browser really has become the new NS4. Hopefully IE7 will be a lot better, we'll find out later this year.

I'm all for folks being on the same page and having guidelines. My contention is that the W3C guidelines are archaic and needlessly nitpicky. I don't need a 3 page manual to assemble a pencil sharpener.

Originally Posted by Tim

So why don't the mainstream sites validate? The answer is because like all of us they don't have to.

Then, again the question becomes 'then why do I'? If it works in the browser and the search engines can find it just fine, why should I be concerned with the fact that I use a height attribute and it fails validation?

Originally Posted by Tim

Imagine what would happen if Google, Yahoo and MSN decided to add validation to their algorithms and that only valid and accessible sites ranked highly... everyone would clean up their code very quickly. Funny how self interest works!

I guess, but in response to this and the post earlier about being able to read my page in 2012, I would say they're never ever, ever going to require validation because there is nothing to be gained from requiring validation either on their part or on the part of the people doing searches, so I don't look for that to happen.

As for people in 2012 not being able to read pages coded today because they failed to validate: That supposes that browsers are going to get worse at code interpretation in the future. If I am thinking about upgrading to a new browser who's only shortfall is the fact that it doesn't display sites correctly that don't validate W3C, I'm going to go out on a limb and say I'll probably stick with my 'old' browser.

The browser of tomorrow that can't accomplish what the browser of today handles just fine, is gonna have a helluva time getting much traction in the userbase.

[mike]

Originally Posted by sands

Also if we were to keep aside for a moment the merits or need for valid code, at least no one can say that validating code is harmful in any way, can they? But not validating code could lead to problems. So the best bet would be to validate.

Nobody (at least I'm not anyway) is saying that validating your code is a bad thing. I love all valid sites and wish a hearty heigh-ho to all of their authors.

Code on fearless ones. I toast your pereseverance!! I applaud your aplomb! And I sleep better at night knowing you're out there. We want, no... we need you on that wall!

[sands]

Originally Posted by mike

Nobody (at least I'm not anyway) is saying that validating your code is a bad thing. I love all valid sites and wish a hearty heigh-ho to all of their authors.

Hey.... wait a minute.... I thought that I was the first out here to say it was a matter of choice.

Originally Posted by mike

Code on fearless ones. I toast your pereseverance!! I applaud your aplomb! And I sleep better at night knowing you're out there. We want, no... we need you on that wall!

Very droll!

[Easywebdev]

Originally Posted by Tim

I contend that it is largely because those browsers have built-in tolerances that developers today get the impression that it is OK if they make mistakes. JavaScript and Perl coders, amongst many others, would surely be envious of the flexibility.

Heh, perl coders or any programmer would not be envious of that "flexibility", they like things to be standard and spit out warnings when you write erroneous code.

The problem with non standard markup is that browsers try and interpret bad "malformed" markup and display "something" it might not be exactly what the designer intended but would be "good enough" (ie6).

A lot of web designers are not programmers and use wysiwyg tools without knowing html. If they ever used a C compiler or php with error reporting cranked up and their code either refused to compile or spits out a stream of errors and warnings to the browser then they would not be able to pass the work off as "good enough" an error is an error and should be eradicated.

The w3c validator is the next best thing to a html "compiler" it shows where your markup is wrong, anyone who gets paid to develop a website should be using it. But until the browsers start to drop support for deprecated tags or display warnings/errors then the "good enough" brigade will be able to excuse their sloppiness.

Writing valid html markup is not rocket science, heck it is not even programming. You layout a document and style it with css. How many tags are used on most sites? 8-10 on normal pages maybe 18 on a page with a form. How long can it take someone to learn what each does and how to style them?

People are lazy, why read up on "how to" when you can buy an editor that spits out "good enough" code?

Valid markup has become a buzzword and people are now requesting websites be built with it. Designers are advertising the fact that their work is valid. Whether it matters or not, it is here and it will only grow.

There is no valid argument for not using valid markup.

[mike]

Originally Posted by Easywebdev

The problem with non standard markup is that browsers try and interpret bad "malformed" markup and display "something" it might not be exactly what the designer intended but would be "good enough" (ie6).

Which is why you check your page in a few browsers.

Originally Posted by Easywebdev

A lot of web designers are not programmers and use wysiwyg tools without knowing html.

You say that like it's a bad thing. I couldn't disagree more. The easier it is for Joe Normal to make himself a web page, the better says I.

Originally Posted by Easywebdev

If they ever used a C compiler or php with error reporting cranked up and their code either refused to compile or spits out a stream of errors and warnings to the browser then they would not be able to pass the work off as "good enough" an error is an error and should be eradicated.

I don't believe the the aim on much anybody's part here is to 'pass something off' as 'good enough'. They build their pages, the pages work in the browsers they check them in, they are indexed in the search engines, and they say "cool, my page is done". I'll go along that an error might be an error. I'd add however, there are all kinds of errors and some of them aren't going to be worth hunting down and eradicating if their existence (or non-existence) has absolutely no impact on anything other than a validation scan.

Originally Posted by Easywebdev

But until the browsers start to drop support for deprecated tags or display warnings/errors then the "good enough" brigade will be able to excuse their sloppiness.

First, any browser dropping support for deprecated tags is a doomed product. For better or worse, things are what they are. Nobody is going to upgrade to anything that works 'less' than what they have (some Windows versions excepted, of course).

Second, here we are calling folks sloppy and lazy again. I think this is the fundamental basis of my beef with the validation camp. The VAST majority of sites (large and small) are going to fail validation for some reason or another. Is it fair or reasonable to assume that these folks are all just sloppy and lazy?

I don't think so.

Lot of them are using Dreamweaver or some other wysiwyg editor (that they may have dropped a nice chunk of change on), checking their end product out in a few browsers and making sure the search engines are crawling/indexing them.

That may all sound real easy and all but I can guarantee you a good deal of thought work and effort is going into the equation - maybe not in the areas of validation, but in layout, design, functionality and things like that? - absolutely.

The lazy thing just doesn't always hold water. Furthermore, validation proponents don't do their cause a whole lot of favors throwing 'lazy' around. For one you are trying to convince someone to not use or make difficult config changes to their wysiwyg editor, because it isn't validating - which doesn't mean it won't render in the browsers fine or be a problem indexing, it just doesn't validate. They must be lazy. That aint exactly the ideal rationale to win the hearts 'n minds if you catch my drift.

If you're going to try convince someone that they need to rework the site they've already made a big investment in time and money in - not because it displays badly, not because search engines don't like it, and not because it won't do what they designed it to do - but for the sake of validation, I would start somewhere other than 'you're lazy'.

Originally Posted by Easywebdev

People are lazy, why read up on "how to" when you can buy an editor that spits out "good enough" code?

Well for 2 reasons, first being that they can buy an editor and do it themselves and second because there is absolutley no penalty whatsoever in any measurable capacity for NOT doing it this way.

Originally Posted by Easywebdev

Whether it matters or not, it is here and it will only grow.

I dunno... if it doesn't matter, why bother? Because somebody's gonna chide me and call me lazy if I use a height attribute? I dunno... I think I can live with that.

Originally Posted by Easywebdev

There is no valid argument for not using valid markup.

Sure there is. The search engines don't care, the browsers don't care and the people viewing your site don't care. As such, if you can get Dreamweaver (or whatever) and roll your own page, I say go for it. That, in fact, is what there is no valid argument against.

Again, I don't want to be an anti-validation guy. I'm not at all. I think there needs to be some guidelines and we need to be on the same page etc etc. The W3C is a good thing. Really. I'm not even suggesting otherwise.

All I'm saying is that good enough, for us unwashed masses, is indeed good enough by any practical application in this debate. I start to get irritated with all of the name calling and value judgements that seem to spring from the validation camp.

You do it your way, I'll do it my way - they both work. When they don't both work anymore, we'll take it from there or upgrade our Dreamweaver. In the meantime, any discussion of how or why you should worry about validating your pages should avoid words like; lazy, dumb, careless, etc etc. That's not gonna win anybody over and you actually risk sounding elitist and nitpicky.

[Jabber_uk]

I'm with Mike on this issue as I am sure you have guessed ;-) Much of what some people might say about validation being the 'only way' or the 'better way' or 'whatever' is the same as saying that when people have problems with security issues in Windows, some people's answers will be 'get Linux!' or 'use Firefox!'. That just sux if you ask me.

Originally Posted by easywebdev

People are lazy, why read up on "how to" when you can buy an editor that spits out "good enough" code?

Which seems to suggest that non validating code is somehow 'less' or 'not reaching a required standard'.

Validation is good but let's not get carried away with it as there could be downsides also. Downsides? A bit like using Firefox is supposed to be the better browser, but I have had more issues with Firefox than any other browser! On my PC it takes up more memory than IE for a start (a LOT more). Some say that Firefox is more secure in its basic form - thats because it doesnt do much in its basic form!

Lol sorry I don't want to turn this into an anti Firefox thread cos it isn't. All I am saying is that let's not get on our high horses and look down on those who don't produce 'perfect' code. The area of creating websites is huge and validation is just a small part - pretty important I guess but not the be all and end all.

Oh and a question:

Is HTML coding actually coding as in the programming sense of the word? If not, then we cannot apply the same high standards required or demanded from producing programmed applications.

If you're a programmer then getting around the errors is what makes you produce perfect code not pro