• Beware of XHTML – quite a few years ago, this article had appeared on this well-designed site that serves XHTML to browsers such as Firefox and HTML 4.01 Strict to Internet Explorer. This article on XHTML outlined the multiple problems with implementing XHTML, how its usefulness declines in real-world practice, and how the advantages are clearly outweighed by the disadvantages. This is a very good read!
• The article that started it all, Beware of XHTML covered topics on XHTML such as the complexities within Content-Type and Content Negotiation, the problems with Internet Explorer that likely will never be fixed with XHTML, poor performance of Firefox on pages with the content-type of application/xhtml+xml, and the fact that XHTML is not more valid or semantic than HTML 4.01 Strict.

XHTML 1.0 Strict is the standard I’ve used in web pages for a long, long time. However, I don’t do it the right way. In Firefox I could serve it up as application/xhtml+xml but this site along with the main site are test sites, and I haven’t advertised it much. In IE I could serve it up as text/html as I already do, but I’d have to use xml:output with XSLT to output the markup as standard HTML 4.01 Strict and at the moment, I don’t feel like it. The sites are new, don’t have enough members, and definitely not enough activity. It’s a wasteland in fact!

The fact is, why have to serve it as application/xhtml+xml? Most people do not want to do content negotiation, and why should you if all you’re making is a small web page, blog, or forum? With HTML 4.01 Strict, everything still works the same. Sure you can’t use the XML features, but most XML features such as MathML aren’t supported by most browsers in any useful way anyways! The jQuery, Mootools, whatever library you use for Javascript will still work the same, so why should it matter? Well, first we need to examine XHTML’s advantages.

Quoting the link again, the only-listed disadvantage of XHTML is its “limited practical use.” However, this is a BIG disadvantage. Internet Explorer didn’t support CSS 2 very well until IE8 came out – and this wasn’t long ago – in early 2009. But it still doesn’t support XHTML fully. The “application/xhtml+xml” content-type still breaks, thus if you serve XHTML with the right content-type, IE users will suffer. But if you serve XHTML with the “text/html” content-type, you’ll suffer because you aren’t using XHTML correctly. The standard is broken if you serve XHTML and plain HTML.

The lack of productivity I mentioned was that it takes time to get your markup to work properly and still be functional. Adding an XML prologue is necessary, yet once again, IE doesn’t accept anything before the document type declaration and thus will render the page in Quirks Mode, which means your page will look noticeably different.

“Strict” as compared to what? The document type that you use doesn’t come with the strictness guaranteed. You can still mess up by putting obsolete elements in your document. You can still mess up by mismatching tags. You can still mess up by not quoting attributes. You can still mess up through a simple and mundane typographical error. XHTML is not strict, not any more than HTML 2 was!!! The use of the standard is either strict or not strict. That’s up to the developer or designer. HTML 2 had rules. HTML 3 had rules. HTML 4.01 Strict had rules. So did XHTML, and XHTML 1.1. So does XML and CSS. So does Javascript. Need I go on? Just because a particular specification implies more strict rules doesn’t mean that this is what you’ll end up with! I’ve seen MANY XHTML 1.0 Transitional pages that wouldn’t even make valid HTML 4.01 Transitional documents!

“Strict” and “Compliant” are just two more of these “keywords” designed to make it sound like you know what you’re talking about. Like “pro-active”, “be positive”, “deprecated”, “obsolescence”, “quantum singularity”, “warp core breach”, and “1337″. Knowing how to spell and pronounce the word doesn’t mean that you understand it. It just means that the easily-impressed will think that you do without questioning your logic, which is very common in nations like…mmmm…politics…never mind.

XHTML does have firmer rules, though. It does promote compliance and better standards, but someone who isn’t willing to validate their website to stick to the standard should just stick with HTML 4.01 Transitional or Strict. XHTML served with the content-type of “application/xhtml+xml” is not to be written by the amateur web developer or author or designer. It really is hard-core. If you misquote just one attribute, the page will not display and instead be replaced by the XML well-formedness error. And even if you do get your website all shiny and coated with XML’s juice of epic win, it will just prove that you have too much time on your hands, or it will just mean that Internet Explorer won’t be able to use it. If you use content negotiation, and do so the wrong way, you’re putting a lot of overhead on the server-side processing and slowing down the page on the client side as well. Say your site has a forum on it, and you serve everything as application/xhtml+xml except in the case of Internet Explorer. And say someone makes a forum post that somehow compromises the XML well-formedness of that topic’s page. You’re screwed! But those viewing it with Internet Explorer will have no problem at all!

If everything isn’t right, and I mean everything, XHTML 1.0 Strict or 1.1 plus the content-type of “application/xhtml+xml” will be a big all-you-can-eat buffet of fail.

The XHTML family of hypertext markup has been quite successful for many pages out there, including CSS Zen Garden, which itself serves XHTML 1.0 Strict as text/html. There are many ‘valid’ XHTML sites out there, yet a small fraction of a percent are actually being served as XHTML – not XHTML tag soup served in text/html. “Stricter” rules still don’t mean that abuse won’t occur. HTML 5 is the newest flavor of HTML – the first we’ve seen in a long time since HTML 4.01. It comes in its own two flavors – one called HTML 5, logically-so, and another called XHTML5 (which I call HTML5+XML as the former name is just misleading and confusing). Even HTML 5 introduces some idiosyncrasies such as having a header tag instead of just using h like XHTML 2.0 was going to do. If they need to add a new element, why not add one that uses one character, like the paragraph (p) tag does? Not only that, but numbered headers can be nested within the ‘header’ element, meaning that you get two elements for the price of one – just great, more DOM processing.

The W3C along with all organizations initially have their good intentions, that as they grow from there, end up getting skewed into things they should not have become. Initially, SVG and MathML sounded like great ideas. But the bottom line is that there’s hardly anything that supports them. XHTML 2.0 seemed to be a logical evolution because it was a new specification that, through its lack of backward-compatibility, would begin a new line of web authoring based on better semantics and more sensible elements and attributes. HTML 5 added its own array of elements, sharing one in particular with XHTML 2.0 – the ‘section’ element. However, at this time, only HTML 5 is supported, XHTML 2.0 is dead, the working groups have merged, and there are still few browsers that support HTML 5. Internet Explorer doesn’t support some of the new sectional tags that were introduced in HTML 5, and this comes as no surprise – not because Internet Explorer is SO inferior, but because Microsoft only accepts a standard into implementation if 1. it’s their own standard or 2. the standard is mature. The former reason is not so respectable, but the latter reason seems logical. CSS3 is constantly changing its specifications because it’s hardly a working draft, only parts of it working in a scattered way from browser to browser.

And the future? Who knows? Someone at sometime will have to accept that they will be losing a big part of their audience if they decide to use all these XML modules, CSS3, and so on. Somebody else is going to remain behind, using HTML 4, with a much greater audience, yet still behind the times. This cycle will likely repeat for many many years. One day there probably will be an XHTML 2, or perhaps a markup type like CHTML (C++ Hypertext Markup Language) which uses some C++-like rules in some strange way. Maybe the web will “asplode” and end this whole paradigm called the “Internets”. Or maybe, just maybe, everyone sooner or later is just going to decide to stick with HTML 4.01 Strict forever and ever, until the end of time. There will be no content-negotiation, Internet Explorer will still not support XHTML anyway, and nobody will bother with it any longer. However it goes, it’s not such a serious issue as many used to think it was. It’s just the internet, and when society finds some other medium to use to live out their nerd-filled passions of arguing Star Wars and Star Trek, the internet will probably start to lose ground.

I say let the W3C do what they will, let the browser vendors do what they will, and let’s just go do our own thing for once, like complaining…about things like this on the…internet…..yikes. I guess the internet really will be around forever then…