Validating the experts

Come to think of it, these people are probably using transitional doctypes just so they can pass the validator, rather than using a strict doctype and having a few errors. Obsessed with validation? Maybe. My opinion is that it's better to use a strict doctype with a few errors than a transitional one, even if that means you'll get a green bar. The reason is because transitional will let you get away with stuff that strict will catch.

My suspicion is that some people probably validate to strict but change the doctype to transitional when they find errors they can't fix.

So, for the next part of the experiment, I will validate pages with transitional doctypes as xHTML strict.

A List Apart (not previously mentioned) - comes up with 6 errors, all improperly nested form elements. All easily fixable but still - what's the big deal?
Zeldman.com - 9 errors, 2 are the aforementioned improper nesting; others are improperly nested input elements, a border attribute on an image, and language= attribute in a javascript link. Again, easily fixable but not a big deal either)
Cameron Moll - 14 errrors, only 5 more than what was found on the transitional doctype. Big deal? Nope. Easily fixable? Yes.
Molly - one error, an improperly nested link.

That's all I'm going to do. The W3C spec states that "Authors should use the Strict DTD when possible, but may use the Transitional DTD when support for presentation attribute and elements is required." (as cited here) None of these sites are using presenational code. An example of a proper use of the transitional doctype under this assumption would be this forum - the BB code tags write font tags, so the transitional doctype is appropriate until we can fix it.

(note that all of these test are only done on the home page - other pages may have other problems that legitimately require a transitional doctype)

The DOC tags do have different results. I had a site I did as transitional, thought why not just go strict, so I did and fixed the errors until it validated - then viewed the site. The site was a mess (layout wise), put the transitional tag back in and the site went back to how it should of been.
End result; validated as strict but used the transitional tag.

For me, strict isn't worth the effort (xhtml, not html). On a basic layout it's fine but more complex you have to design to the strict tag as well as the browsers which can be a back and forward nightmare.
Using html strict is easy to do and doesn't have the layout issues.

Yep, i experienced that too, when validating for transitional to strict, it kind of messes the design... also I noticed that when making a layout and you forgot to add the doc type and your finished with the layout and the codes are in there and everything is dandy, once you add the doc type everything quirks out...

Sometimes I just put a quirks mode comment line just before the doc type on top of the html... that fixes it...

I try to validate all my pages as possible as I can, even the contact page etc for my new project... Ill post it here once im done with it....

Well I think that since Zeldman pretty much led the standards march and he's always talking about doing things up good and proper, then I would expect perfection from his code. That said, I think he uses his zeldman.com site for experimenting too.

Cameron Moll....yeah he can code well, but that man is more about design....and man can he ever design!

I've never had a problem with layouts chanigng (a lot) when switching to a strict doctype. Not when everything is coded properly, anyway. That's another reason why I think that using strict is better than transitional even if there are a few errors. With strict you always get standards mode (not quirks or Almost standards). Strict means it's rendering correctly.

The other interseting thing here is that most of them don't seem to be bothered by the whole problem with serving xHTML as HTML.

Anyway, I somehow doubt that any of those people are using transitional doctypes because of display problems. I'll have to test that next

Alright, lets compare standards and strict mode. I can't be bothered to do screenshots right now so you'll have to trust me or try it yourself.

Zeldman.com - minor pixel adjustment when moved to strict (would not be noticeable if you weren't comapring two tabs directly)
Molly.com - same (the text content moves up by 1-2 pixels)
Cameron Moll - same (seems to be spacing around headers that's different)
Simplebits - no noticeable problems (can't seem to get one tab from cache and the other from the live site anymore )

So, I'm pretty sure display problems aren't the issue. I know ALA posted something at one time about why they're using a transitional doctype but I can't find it. I'd assume Zeldman is doing it for the same reason.

btw Meg, which browser did you try it on? try Ie... the quirks are sometimes directed for Ie...
I hope you can find the ALA reason... Im really interested with these...

You're right, I only tried in in Opera (but I don't know of a way to reload from cache in IE???). I'd have to save the whole thing. I still don't think that's the reason though.

I tried ALA and molly.com - both look identical in IE with strict and transitional doctypes. So that's not it. If you're talking about display problems, it could be something like IE 5/mac that they wanted to accommodate. However, from what I was able to find out, the big display differences between transitional and strict are the handling of table cells and image spacing in mozilla browsers. But, none of these sites rely on tables or hard coded images for display anyway. So that's not it.

From what I remember about the ALA explaination it had something to do with passing the validator.

Finally found the ALA explaination. They used to have this listed on their Q&A page but for some reason took out most of the actual explaination. The current page simply says "It’s our way of sticking it to The Man." Gee, that's helpful.

So I had to go to archive.org to find this:

Quote: Just kidding. Fair question! Nearly all our content could now be delivered as XHTML 1.0 Strict, which goes even further than XHTML Transitional in separating underlying structure and semantics from visual presentation. But a few of our older articles require attributes that are illegal under XHTML Strict. Rather than switch DOCTYPEs between articles, we chose the DTD that would best serve all our content, old and new.

Um, okay, what-ever. I think that's being a little overly fussy about validation actually. I think I would probably put in the strict anyway, if it was only a few problems on a few pages.

So what about all those other sites? Could be the same problem. I might have to email some of those people and ask (and while I'm at it, ask about the xHTML as HTML thing)

Thanks for the links... I think im quite addicted now with validation... although I now like validating to xhtml1.1 and validate the css... What I notice though as you get accustomed to fixing things for validation, the next time you code by hand, you automatically correct the errors, and finally finish with 6 to 25 errors you can fix in 10 minutes...

But its really puzzling why experts tend to ignore the standards they impose, but anyways, I also found this site made from blogger engine that says "I design and develop professional, valid sites for a living. This is my personal attempt to save the internet one webmaster at a time. I know this site doesn't validate. It's Blogger. Get over yourself." . Just show how serious designers are now with validating...

I even recieve dozens of e-mail for putting "we pride ourselves with validation etc.." when our main engine was only html 4.01 validated and the main page shows an xhtml1.0 strict badge... where the main page is actually xhtml1.0 validated. Decided to correct everything.

So anyway, there are actually people who get ticked when your first three pages are validated as xhtml strict and the others only at html4.01 ... this pretty much drives me a little over the edge sometimes...

You should actually read the whole "serving xHTML as HTML is harmful" stuff. That's why #1 on the list up there has got HTML 4 strict. Nothing wrong with that if you understand and believe that side of the debate.