Why is validation so neglected?

accessibility and usability

[eboda]

If web developers would take their roll more seriously and validate their coding, many of the usability issues would be reduced for both Mac and PC users. So many of us expect the browser to resolve bugs in dirty coded websites - Actually we should push a lot of the blame over to the web developers.

To support my point… Would you buy or live in a house in which the foundation did not meet building codes? Of course not… it puts the contents of that residence at risk.

If the foundation of a website is weak, content is at risk. Accessibility and usability are limited and that could result in a loss of memberships, sales, and users.

I love it when people say: “Mac users only make up 5% of our customers, why should we design for them?” Can you imagine a 5 percent increase in users and or sales just by providing them access to a website!

This is almost not worthy of a discussion as validation should speak for itself. Someone (I think it was a gal named Chio? Here at WebmasterWorld) said: “It is a sign of a good programmer.” Those words echo in my mind with everything I do now… I am also working on some research regarding validation. I am sold on validation for months and is the core of my work ethic. I cringe at my old work – non-validated code! Blegh! Everything I do from here on out has been more productive and enjoyable for my clients, their site users, and my reduction of headaches.

So why is Validation shelved? It should play a big part in what we do… Not to mention saving our clients money in the long run.

[digitalghost]

Another problem comes into play when you do take the time to create valid code and the latest browsers can't render it properly. This increases the frustration level quite a bit.

I'd also like to add that creating valid code is just the first step regarding accessibility. Valid code in no way means that a site is 100% accessible.

[papabaer]

Actually, there is a bit of a groundswell supporting VALID CODING among many young webmasters... validation is now 'cool.'

XHTML1.1 is bangin'!

It's nice when fads have some real added benefits....

Wha?[sic] You ain't writin' straight-up X? you whack!

[g1smd]

Valid code stands more chance of working than non-valid code, so I do the former. It doesn't take long, and is less likely to cause funky unwanted effects on the site, trip up spiders, or confuse users.

Best check that HTML code: [validator.w3.org...]
and the CSS: [jigsaw.w3.org...]

There is a whole load of broken CSS on the web; I think that will cause a lot more problems than broken HTML ever did.

[papabaer]

What has, and still, amazes me is that many whose first priority is SEO, fail to realize the benefits of validation. One might agonize over the exact placement of keyword phrases throughout a page, but overlook sloppy coding that will stop a spider cold...

Amazing....

[eboda]

I agree validated websites do have a better chance at usability over junky coded websites.

Honestly, I can't imagine what programmers have to face when developing or producing a browser! There are soooo many browsers on the market: Opera, IE, Netscape, Safari, OMNI and on and on. I am sure the effects of that have a heavy influence on platform development?

Since the majority of sites are created on WYSIWYG programs, browser engineers have to probably consider commonly used non-validated code.

Could part of the blame go to the WYSIWYG software engineers? Speaking in regards to Adobe and Macromedia these editors insert junk everywhere? but allow the ability export most of the junk out after site completion (still a pain). Where do the wysiwyg software developers see themselves in relation to validation?

Agreed papabaer!

What has, and still, amazes me is that many whose first priority is SEO, fail to realize the benefits of validation.

I don?t think many of us think validation makes much of a difference. If we all validated our code, and valued clean code? I think we would progress faster? as we should.

[dkubb]

I agree with the need to validate. As an application developer, I've come to understand the need to write simple, clean and easy to maintain code in applications. I carry over this obsession with clean code into any markup languages I work with, like HTML and now primarily XHTML.

Just today I was looking for better validation tools. I use the W3C HTML validator religiously, but I'm looking for something a little more in-depth. Not only do I want to know if the code is XHTML 1.1 strict compliant; but I also want to make sure that I don't do anything that breaks popular browsers. On top of all that I'd like something similar to Bobby that makes sure I comply with all major accessibility guidelines.

I saw a program called the CSE HTML Validator (hope I can say the name here.) and I was wondering if this is any good?

Is there any tool or set of tools you use to validate your code that would meet my above criteria? Is there anything you do over and above what I've already mentioned while validating your code?

I'm really interested to hear what other people do with their "best practices" in this area.

[papabaer]

One other thing "amazes" me... writing valid code is NOT rocket science.

It's far easier to follow a set of logical rules, structured accordingly, than to hope for the best and throw "code-caution" to the wind.

[tedster]

I think the reason why valid code has been neglected for so long is historical.

Today, the W3C is starting to call the shots. But in the early days, the W3C's version of HTML lagged far behind the features that Netscape and Microsoft were developing in their browsers.

In practice, the game for browser developers was to innovate with new features that the W3C might adopt, rather than worrying about support for the standards that they had already created. It was the browser wars, and the winner was going to "own the web".

So if a develper wanted to have a "cool" website (and that was really important in those early days) they needed to write outside the standards -- and that usually meant ignoring the standards completely. The pressure of getting a site online meant making it look good, no matter whether the code was valid or well-formed. Standards just were't in the practical equation back then.

The shift to valid code that has begun in recent times (and it has just barely begun) is really welcome. But old habits die hard, and there's little motivation for browser makers to REQUIRE standard code...who would use their browser?

...but overlook sloppy coding that will stop a spider cold.

Validation is my first port of call whenever I wonder why "this page isn't in the search engines". And often the error is not some minor detail, but a brazen muck-up. Like omitting the ">" on a tag!

I'm working with one development team right now - minimum 3 year experience for all members - and no one there had ever heard of validation. Their TEMPLATES don't even validate. And this is still the rule, in my experience, nowhere near the exception.

There's plenty to do, isn't there?

[edited by: tedster at 6:31 pm (utc) on May 16, 2003]

[Edouard_H]

Glad you mentioned that Tedster - a major benefit of valid code, and particularly XHTML, is the ease with which it can be edited in a collaborative environment. Improperly nested tags, unquoted values, kludges, and other such vagaries and individual coding quirks don't survive validation. Any capable coder can pick up and run with it.

Any firm or organization where multiple developers work on a project should insist on validation; saves many work hours over the long haul.

[g1smd]

>> So if a develper wanted to have a "cool" website (and that was really important in those early days) they needed to write outside the standards -- and that usually meant ignoring the standards completely. The pressure of getting a site online meant making it look good, no matter whether the code was valid or well-formed. <<

I still don't think there has ever been any excuse to ignore the rules for basic HTML structure, the rules for nesting of elements, and such like. I don't mind the odd bit of browser specific code, but totally ignoring basic rules is a no no.

When you write a web site, you use HTML. HTML is a language, just like a computer program is. The language has some rules. If you don't follow the rules then the "program" may not work how you wanted it to work. With a computer program you test it. With an HTML page, you have no chance to test it using every one of the browsers out there today, so instead you get something to check whether you followed the rules. Hence you validate your code.

I simply use the W3 Validator:

* to get rid of typos in tags, like IMMG= and COLOUR= ,
* for sorting out missing > like <H1 or </A ,
* to investigate incorrect nesting like <B><I></B></I>,
* to make sure that I have no Block elements inside Inline elements.
* I also add ALT="" attributes on all IMG tags,
* and add TITLE="" attributes on all <A HREF=...> tags
* make sure that all attributes are quoted, so WIDTH=85% becomes WIDTH="85%" ,
* I turn all unescaped ampersands like & into &amp;
* and correctly mark out JavaScript code with <!-- starting and // --> ending tags,
* noting that comment text, and style codes, need only the <!-- and --> tags,
* adding a TYPE="TEXT/JAVASCRIPT" attribute on all SCRIPT tags,
* and a TYPE="TEXT/CSS" attribute on all STYLE tags.

However, I do leave in some of my pages a few tags that are Netscape or IE extensions, so some of my pages don't completely validate to the full standard. I am confident that there are no logic problems with the HTML structure, nor any incomplete tags, or ones with typos in. I think this methodology is the minimum that webmasters should adopt. However I note that far too many people don't check their code for these sort of errors at all.

For those that do not check their code in a validator, I ask them one simple question:

>> Oh yeah, you do test your site in all of these browsers, don't you? <<

Act, Activator, ActivatorDesk, Amaya, AMosaic, AOL Browser, Arachne, Arena, Armadillo, Gzilla, AvantGo, AWeb, Beonex, Communicator, Bezilla, BrowseX, CAB, Cello, Charlotte, Chimera, Cineast, Clue, CyberDog, DocZilla, Emacs/W3, Espial Escape, Fizzilla, Fresco ANT, Galeon, Grail, HotJava, IBrowse, iCab, Internet Explorer, Jazilla, Ua ho'oku'u 'ia Ka Ho'olele!, K-Meleon, Konqueror, LibWWW Line Mode Browser, Lite, LunaSuite, Lynx, MacLynx, MacWeb, MacWWW, Minuet, MMM, NCSA Mosaic, Mozilla, MultiWeb, Netscape Navigator/Communicator, Netcomber, Netcomber netomat, Netsurfer Control Panel, Net-Tamer, Newt's Cape, Off By One, Omniweb, 1X, Opera, Palmscape, ProxiWeb, ReqWireless WebViewer, UdiWWW, Voyager, WannaBe, WebBoy, Web Prowler, Web Stalker, Wen.Suite, WWW/LX, Zen, Air Mosaic, aMozillaX, ArcWeb, Ariadna, Browse-it, eWorld Chameleon WebSurfer, embedix UI, Lineo Embrowser, Emissary, eWorld, I-Squire, Jazilla JoZilla, the Light of Adamas, Lynx/2, NetPositive, Netshark, Pathworks Mosaic, Procomm Plus, SlipKnot, Spiderwoman, Spry Mosaic, Spyglass Mosaic, StarOffice, Tango, Mac Tiber, Win Tiber, VOL Tiber, WebExplorer, WebNav, Wingman, Winweb, ALynx, Arena i18n, Avanti, Bobcat, BrailleSURF, BrookesTalk, Browse, Chai Farer, Charon, Closure, DOSzilla, DPWeb, EnterWEB, ESEWWW, Express, Galeon, Go.Web, IBM Web Browser, ICE Browser, Internet Adventurer, iPhone, JustView, Links, Lotus Notes, Mathbrowser, MidasWWW, mMosaic, Mnemonic gtkBrowser, mothra, Netgem netbox, NetFront, NEWSie, Smoke Zone Phoenix Pro, Planetweb, Private-Eye, pwWebSpeak, Q.BATi, QtMozilla, QWeb, Rozilla, squeak scamper, NewDeal WebSuite Skipper Pro, SPIN, Termite, TkWWW, QNX Voyager, w3m, Warpzilla, the Wave, Wbrowser, Web-O-Matic, World Finder.

No? Then validate your code instead!

[jim_w]

Geeze, I didn’t realize there were so many browsers. (checking .htaccess to see how many were banned by accident)

[Wertigon]

I believe the biggest reason is the fact that hardly anyone knows about validation today. And also the fact that people are lazy and you need to practicly give them a "validate now" button in the editor they use. ;)

[excell]

:) using the validator on a site that really needs a re-design is a good way to show the owner why, along with some screenshots of how it messes up in different browsers, it works like a gem.

missing from list I think.. webTV(does anyone still use it) the new MSN browser MSN Explorer

and of course if they have trouble displaying in IE there is really something wrong, but it does happen.

[buckworks]

<<I don't think many of us think validation makes much of a difference.>>

We're lulled into thinking this way because browsers are so forgiving.

If something will make a positive difference, even a tiny, tiny one, take it seriously. We tend to look for one big thing that will make a 100% difference. Sometimes we can make the same progress by improving a hundred things 1% each.

Code validation: won't hurt, might help.

[eboda]

tedster
I'm working with one development team right now - minimum 3 year experience for all members - and no one there had never heard of validation. Their TEMPLATES don't even validate. And this is still the rule, in my experience, nowhere near the exception.

This is tough even when you are educating a programmer who just walked out on the streets with a degree in computer science. It really amazes me how far behind our most our Universities are on programming. I could put them in a category of spamming the industry with junk code.

When did we start accepting sloppy code?
Was the dotcom crazy? The faster we can crank them out, and the more the better. Yeah perhaps...

It is really funny how I have come full circle.
- I started learning legal HTML coding from books, W3C and such.
- WYSIWYG released (learning to produce at high speed)
- Fell in love with junk code dropped into WYSIWYG
- Started hand coding again for more control
- Discover the importance of validation
- Back to legal coding and W3C

There is no excuse really why (at least myself) I drifted away from legal code. I guess I just got lazy ha!

[contracosta]

Remember a couple of things from the early days. I started doing sites in the Internet backwoods... rural Ohio in 1995. 28.8 modems were the rule and 14.4 wasn't unheard-of. The ISPs were rotten and generally gave 14.4 throughput no matter what modem you really had.

I prided myself at that time on an absolutely minimal HTML style, because every byte counted. And (embarassingly) I stuck with that code style until 1999 when I got a dotcom job and the other developers started teasing me about it.

Validation is something we can afford to do now, with connection speeds commonly higher. But I used to be proud of the 1 or 2K I could shave off of an html to file make it load faster by deleting everything that wouldn't break the page -- and validation be danged.

Today it's more important that pages validate. Hypocritically, I now look down on those who don't do so!

[senox]

For those who use Mozilla-based browsers there's a useful plugin called Checky [checky.mozdev.org]. It's an interface to online validation and analysis services on the Web. Once configured, you just need to hit one key to run one or more selected services at the same time on the page you see in your browser window.

I've been using it for quite a while now and find it very helpful when it comes to validating pages.

[lorax]

I think valid code - as the defacto standard - will happen in the near future. It's a progression of knowledge among website owners. Much like there didn't use to be building codes until the public finally figured out that there ought to be. I think the same will happen to our industry. Sooner or later I think we will see enough momentum behind one set of standards or another (W3C?) for the different flavors of markup and data and then acceptance/compliance will begin to gain steam from there. While I have yet to have a client ask me if I produce valid code - I expect they will someday.

[eboda]

A good example of why to validate and not to validate…
WebmasterWorld does not validate.
Why?
DOES WebmasterWorld need to validate?

WebmasterWorld set itself up to be a standard for online communities - these communities contain industry professionals in the area of Web Development, Marketing and so on... So... In my opinion yes WebmasterWorld should perhaps validate their coding. Clearly this site presents some authority in being a place for developers to come and communicate industry related topics. WebmasterWorld should be an example of standards we discuss and further more setting a standard, which earns some level of respect.

Now I am not trying to hammer WebmasterWorld for not validating. Don’t take it down the wrong road. I am looking at why educated professionals choose not to validate? Perhaps there is something to learn in this…

[Wertigon]

Probably because it's a *mayor* pain in the bum as soon as PHP/Perl/Whatever gets involved. For the regular HTML documents though, there are IMO no excuses.

[DrDoc]

*mayor* pain in the bum as soon as PHP/Perl/Whatever gets involved.

Just open up the page in Opera and validate from there...

[deft_spyder]

dont forget that ina production environment that doesnt have time to go work on each little nuance of netscape 4.7, or macs browser, etc... it may just not be worth it. I know of one instance where if it worked for IE, which was going to be 96% of the user base anyways, going back and fixing things wasnt practical so people can use obscure browsers.

I dont say that to be flamed by opera lovers, linux etc, but thats just a fact. If the majority of your audience can see it, other factors may say 'thats fine', lets move on.

(FYI, i just spent some time making some java work with mozilla, so I do both :) )

[thaedge]

Hmmm well as a developer/programmer I will defend myself, somewhat. As far as my own sites I verify changes and such in the major browsers IE 5+, Opera and NS 6.2+ and at least make sure NS 4.7 is decent. The problem when at work is that some things they want look great in IE with some work, but move over to NS and its soso and NS4.7 it wont work yet they dont care.

I agree and validate my code as much as possible, but at same time when you have 2-3 projects and one day or management says u have to have it done, there is a trade off between "good clean code" and "hey it works".

my 2 cents

[eboda]

I don't think we can forget alternate browsers for disabled people.

Validation extends further than just HTML, and CSS but providing accessibility to content. I think largely we forget about another audience. I already can see even more why validation and usability are hand in hand.

According to National Federation of the Blind www.nfb.org:
- 1.1 million Blind people total in the US
- Every year 75,000 Americans become blind

By ensuring that the website uses correct code, one can prevent glitches or errors which may arise during this process.

I don't think I have to post a detailed example - However, if tomorrow one lost their vision and could no longer read through a favorite website. I wonder how quick one would be shouting for validation because the text reading browser could not read the content.

[nathank000]

[ducking] WTF is validation and why is it so important?

[bcolflesh]

Check it out:

validator.w3.org/

Regards,
Brent

[Oaf357]

I have been preaching validating code for YEARS.

But the problem I'm running into with XHTML 1.0 is the whole & thing. Is 1.1 any better?

[waldemar]

(Hello!) Yes I believe XHTML 1.1 is better, because it's strict only.
As someone else pointed out here, there is the problem, that today's "webdesigners" convert any layout in the "standards" of html32+netscapeprop.+ieprop. With XHTML all the style-stuff is banned from html, now we have what HTML was originally ment to be - a page description language. I agree with papa "many whose first priority is SEO, fail to realize the benefits of validation", code validation (of any kind) should either be built into a webdesigners/programmer's brain or be accessed through helping appliations like validators. Anything else if spaghetti code that last no longer than 1 year.

[percentages]

The validation answer is simple.....97%+ of users actually use IE.

Are the other 3% desirable? Maybe yes. But, if it takes 2 weeks (often more) to make one single site complient with all browsers why bother?

In two weeks you can SEO your way to a 10% to 20% increase in visitors....so who really cares about the 3% or 4% of none IE users?....I want the extra 20% of IE users achieved in the same time period!

Another angle...eventually all none IE or IE compatible users will get so fed up with pages they can't view they will switch to IE compatible. Netscape has gone from majority shareholder to a voice in the corner...let it die and move on.

Got to run....MS shares are up 14% in the last 2 months needs to buy some more;)