Why Web Standards and Validation?

By Oli
At 4:50 PM · Tuesday, 21 October · 2003
To Coding · HTML

I recently got an email via the Japan Bloggers list asking Why bother with standards?. I wanted to post a reply, but it became a novella so I thought it’d be better to just reply with a link here ;-)

On 2003/10/19 2:02 AM, “gregg” wrote:
> What’s the point of validation?

I think you’re asking two related questions; why worry about W3 recommendations (the oft-mentioned standards), and why bother validating to check you’ve complied with these standards?

The main reasons:

Easier to find errors
If you make code that follows the recommendations you can use the validator to catch any small mistakes (things like incorrectly nesting tags or forgetting to close a tag [in XHTML]). If things aren’t working like you thought they would, validating can quickly tell you if that’s the potential problem. If not, you can look elsewhere. Many programs that generate HTML also make mistakes, so validating can help non-hand coders too.
Future compatibility reassurance (probably ;-)
W3 recommendations are a good thing to use because that’s what browser makers base their browsers on (theoretically ;-) “Version 6” browsers are built to follow them pretty closely, and any new browser from now will follow these recommendations strictly, so if your website does too it’ll be a _lot_ more likely to work, both now and in the future.
Compatibility
If you write to the recommendations, your pages should look the same on any browser that follows them. This means you won’t have to check on multiple browsers (/multiple platforms…) so much. While this isn’t completely true now (lots of non-compliant browsers out there), it’s getting truer.
Easier to get help
Using the recommendations means you’re following a published set of rules, rather than fudging it as you go along. If you validate you’ll find people a lot more willing to help you solve your problems, because they can understand what you’ve written more easily.
Displays faster
If your code is follows the recommendations it will be easier and faster for a browser to display. This is because the browser doesn’t have to guess what you meant. For example if you don’t include a charset tag, the browser has to fudge and guess what you intended. Generally it’ll work, but it takes a little extra time.
Easier for the web’s most important user
Google loves web pages that follow the recommendations. If your page validates it’s far more likely to rate better in ‘page rank’. I’ve only got about 8 sites that link to boblet.net (thanks Kurt! ;-) yet I’m 6th for oli in Google (out of ~216,000). If I was a SEO Specialist, that tip would have cost you a few hundred dollars ;-)
Cheaper and faster to make
This is more relevant to web design companies; designing to one recommendation is much cheaper and faster than designing for multiple browsers on multiple systems.

> What does one site do because it validates
> that some other site doesn’t?

It becomes easier to error-check/maintain for the creator, and easier/faster for a browser to display.

> So far validation sounds more like one big
> silly club and has no actual connection to reality.

I can’t understand how something provided for free that can help you solve problems (validation services) are silly or not connected to reality. Repeat your sentence with “spell-checking” instead of validation - sound strange? ;-) If your meaning is “following standards is silly”, I think you should read those reasons above again. While changing from old-style code to recommendations-compliant code can be time-consuming and painful, there are a lot of benefits.

If you’re still not convinced, check the Web Standards Project’s reasons for using the recommendations. Keep in mind they’re aimed at web developers not personal sites. It might not bother you if your personal site looks strange (or doesn’t display) on some browsers, but companies loose money and web designers loose customers for that.

> If a site displays what does it matter?

Just because it displays today as you expect on the browser(s) you tested on doesn’t mean it will tomorrow. An extreme example is JavaScript detection scripts - if they’re badly made a new browser won’t match anything, and won’t get served. For HTML you might have issues like the next version of IE using the correct box model, making all CSS-defined boxes suddenly bigger and maybe breaking layouts. For a personal site this is your choice though. For a company site support calls and lost business cost money, so the bar is higher.

> 98% of the net doesn’t validate.

It’s even worse than that. Zeldman’s famous quote is 99.9% of websites are obsolete.

> If anybody has some concrete examples i’d like to know of them.

If mine aren’t enough, or if you need a more authoritative source, check out Tristan Nitot at Netscape DevEdge on the Business Benefits Of Web Standards. Jeffrey Zeldman is also quite lucid on reasons to use standards and his book Designing With Web Standards is a great book about why and how to do so.

PS: Gregg — I think you might be confusing XHTML with ‘following standards’ because you mentioned fixing every <br> into <br /> as being a waste of time. HTML 4.01 is also a great recommendation to follow. If you’re following it, you can still validate: just declare your doctype correctly. Note that it’ll then flag the XHTML code you’re using as errors though ;-)

Discussion...

Comments (8) · TrackBacks (0)  to  http://www.boblet.net/cgi-bin/mttb-external.cgi/28
1. Comment by Daphne  · 21 Oct, 2003 · 8:31 PM

Hello, Oli—

I’m glad to meet you. Thank you for your very helpful information. Nobody else had mentioned why standards was important except you.

I’m not familiar with the area where you live, but if ever you’re in Tokyo, do come to the meet-up so we can meet you personally. Everybody talks shop there—all I do is listen…

Daphne :-)

2. Comment by Rudolf  · 22 Oct, 2003 · 3:52 PM

Hi Oli — neat write-up!

Your work on the Center for International Students looks quite nice, too. I teach at Mie University, and I’ve bugged them before, to no avail, about the bad English and less-than-impressive design throughout the rest of the site. Maybe i should try again.

3. Comment by GMan  · 22 Oct, 2003 · 6:42 PM

I’m sorry this is going to sound argumentative but you asked for a reply so.

> Easier to find errors

First all, let’s get one thing straight. There are no *standards complient browsers* and if even someday there is one, as long as a large percentage of users are using one that is not standards complient, validation is not going to assure that your page will display correctly.

> Future compatibility reassurance (probably ;-)

You already gave my reply there. You put “(probably ;-)”

> Compatibility

See the first answer.

> Easier to get help

that may be true. Sounds kind of snobby to refuse help based on if something validates or not

> Displays faster

Sorry, this one doesn’t fly. Pull up some proof that it’s actually true and that it matters. If it is true for any particular browser that would suggest those programmers don’t know what they are doing and even if it was true because of poor programming the difference would have to be like 3 milliseconds, hardly worth worrying about.

> Easier for the web’s most important user

I call BS. Prove it

> Cheaper and faster to make

See answer #1. Validation will not tell you if your pages work on multiple browsers. You still have to check all of them

Validators are a tool that can help you find errors. I agree. But the ferver with which people ask you “is your site validating” pushes a button with me. It’s like asking some book author “did you use a #2 pencil”. Read the book or page, ignore how they were made.

BTW: I don’t know if it’s intentional but on page for your article there is no “Name”, “Email Address” and “URL” fields. Clicking “Post” gave me a “Comment Submission Error” and a new page which those fields. Is that a spam blocking thing?

4. Comment by oli  · 23 Oct, 2003 · 3:06 PM

Gregg:
Thanks for your comments - not argumentative at all. OK here goes:

finding errors
Validation isn’t meant to assure that your page will display correctly. Of course validation can’t do that. It can however help find problems that prevent your page from displaying as expected. Only testing and fixing the problems can assure correct display. However if the page doesn’t display as expected and you don’t know why, the validator is your best buddy.
future compatibility
Of course I put probably — if I had a crystal ball I’d be too busy getting rich to reply to emails! There’s no guarantee that Microsoft will follow the recommendations it helped create when it releases it’s next browser with Longhorn (2 years away?). Their comments on email re: new versions of Office sound typical of their past evil practices. However in the last 6 years the browser market has gone from ‘proprietary tags first’ and partial support of recommendations, to browsers supporting recommendations more completely with each release. The most modern three (Mozilla/Opera/Safari) are advertising their support of recommendations when promoting their browsers (!). I think that coding to the recommendations is both prudent and by far the least work in the long run.
compatibility
If you coded to the recommendations last year, your website would automatically have worked in Safari this year. By using CSS for your layout your page is accessable to pretty much anything, because the HTML file isn’t munged by presentational fluff. In the bad old days compatibility meant code forking and browser detection scripts = much more work. Following the recommendations is an easier road to general compatibility.
Easier to get help
Sounds kind of snobby to refuse help based on if something validates. I agree. However the problem could just be a missing </table> tag, and why waste your time when the validator can tell them just as easily? I’d rather help someone who was trying, rather than someone who just expects you to do everything for them. Generally sites that haven’t tried validating will have a raft of other problems to solve first - it’s just too much effort.
displays faster
It’s common sense that if you give a browser garbage, it has to work much harder to make something out of it. Surely you’ve heard of GIGO before? While it’s hard to provide solid numbers without having a ‘before and after’ page to test, when reading Wired’s Redesign Report the figure of 50% smaller pages sounds pretty good. This is mainly because they began using CSS instead of tables though, and your beef is more with validating — changing from tables to CSS for presentation is a whole different discussion.
Easier for Google
I call BS. Prove it . I like it how you completely ignored the example I did give ;-) How would you explain my ranking? I certainly don’t know why my website ranks so well - it sure can’t be all the people linking to me. I agree that Google’s boffins have written some excellent code to tease the meaning from the garbage pages on the web. However faulty code will likely not help them to read your page. The consensus on all the search engine news sites I’ve read is that good content and valid code are the best two things for Google. Without consulting a SEO pro or Google I can’t give a more definitive answer, but that’s the word on the street.
Cheaper and faster
Validation will not tell you if your pages work on multiple browsers. I don’t recall ever saying it would. Stay on target now Gregg ;-) Validators are a tool that can help you find errors. I agree. Thank you! That’s one of your two arguments put to bed (finally ;-) the ferver with which people ask you “is your site validating” pushes a button with me. Aaah, now we’re getting to the root of the problem. It’s not that you don’t like the recommendations, or that you think validating is a waste of time, it’s more that questions about these things are annoying you. I’ll take a wild guess that it’s because your own home page doesn’t validate, because it doesn’t follow the standards. Would you consider this a possibility? If so I’m happy to help… (we could even get some ‘before and after’ examples to speed-test out of it ;-)

Hi Daphne, I hope it was useful. Thanks for the invite!

Hey Rudolf — how did you know it was me who made the CIS site?! Wow! Yeah, it’s quite plain, but the requirement was “must look identical in NN 4” (!). Oh, the pain! I think it turned out pretty well though. I’ve yet to get any feedback about it, which is a pity. I just noticed that they finally updated the link from the University’s English homepage ;-) If they’d like to improve the rest of the website, I’m sure the company that hired me would be very willing (I sure would be — I’ll soon have a G5 to pay for).

Keep the replies coming…

RE: the comment form being screwy, sorry! You definitely should be seeing the usual fields — maybe you can help me on this in exchange Gregg?

5. Comment by GMan  · 29 Oct, 2003 · 12:59 AM

I’ll just point out a few things.

displays faster:

If you go back and read your original message you’ll see your original “displays faster” is about the browser drawing the page, not downloading the page. To give you a simple example, if I open a command prompt or shell and type “more mypage.html”. It doesn’t matter what’s in the file, it’s going to display at the same speed.

In order to display any page they have to check for errors, it’s that checking for errors that slows the code down, not handling them. But, they have to check for errors otherwise the browser would not display pages with errors.

You’ll just have to trust me, there is no speed difference in rendering a page with errors vs one without.

> Easier for Google

Did you notice that Google itself DOES NOT VALIDATE. Neither does Yahoo or Altavista or MSN or CNN or Metafilter any of the other major sites.

There is no reason that google would be slower to search a validating site nor is there any reason the pagerank would be different. As someone that has written search engines I can tell you it’s pretty simple to search a page for content whether or not it validates.

In fact it’s about 3 lines of perl to find all the content words on a page whether or not the page validates.

the ferver with which people ask you “is your site validating” pushes a button with me

This is it, I really don’t get why you or anybody else should care that anybody else’s site validates or not.

No, I don’t care if my site does or doesn’t validate. But I just don’t get the fire. It’s like on a scale of 1 to 10 of importance, getting a site to validate is 1, getting information to people is a 10, getting new content to attact people is a 10. Getting the content to be pretty might be a 7 or 8. But getting it to validate is irrelavent.

For proof I can just go back to point out that the worlds most popular sites don’t validate. They aren’t serving less people or messing up the web or doing anything that justifies or warrents the kind of fanatical “let’s all get our sites to validate” movement I see on blogs all the time.

6. Comment by oli  · 30 Oct, 2003 · 12:55 AM

This might make you laugh:

Step right up, folks! Witness amazing feats of bravura and derring-do! Watch as bunk markup vanishes before your very eyes!

That’s the opening quote from the hilariously illustrated promo for structural HTML and CSS. While there’s certainly taking a moral stance on it, there are lots of good links and it’s funny!

Hehehe

7. Comment by GMan  · 31 Oct, 2003 · 6:41 PM

The best thing about that article is it’s by two people:

One from Scott Design, Inc which uses tables and flash for layout at http://www.hotdesign.com and including all of the samples in his Portfolio->Web section.

The other a guy from Adobe Systems and if you go to http://www.adobe.com you’ll see they use tons of nested tables for all their layout.

On top of which the site it’s hosted on http://www.seybold365.com/ is all table based layout.

Kind of hard to take people seriously that can’t follow their own recommendations although I would argue that CSS just isn’t ready for anything more than simple designs ;-)

8. Comment by loufoque  · 2 Feb, 2004 · 1:49 AM

CSS just isn’t ready for anything more than simple designs

Just look at CssZenGarden