-
but I didn’t know it was so bad for browsers
It isn’t. That’s the whole point; the HTML parser has been deliberately designed to not care about syntactic variations. Serving things as
application/xhtml+xmlisn’t the ‘answer’ either. Mozilla specifically recommends against it. And so do various other authors, while simultaneously pointing out the serious pitfalls of even thinking of switching.To be honest with you, I don’t really give a crap about web standards myself; I think they’re a mess and until (hah) they get cleaned up I don’t think they’re really worth adopting as anything other than a curio/challenge. The fact of the matter is there’s nothing XHTML does that strict HTML can’t. Real HTML has the same validation (it’s not browser-enforced, but neither is most people’s ‘XHTML’) and the same separation of function and form. These are the two things most of the XHTML bleaters get excited about, and I just find it hilariously amusing that the whole time they’re going on and on and on about standards (and, let’s face it, often attacking other websites who aren’t ‘compliant’) their own pages are horrible, HTML-invalid tag soup messes. Because, when it comes down to it, most people who pretend they’re experts in this sort of thing don’t know jack. I didn’t; and now I do I’m not sure what to do with it. (I’m honest-to-gods considering re-writing the whole backend of sk.log to detect whether a user’s browser has an XHTML parser or not, and serving either XHTML or HTML content as appropriate. The one problem being that I’m now not 100% sure of the wisdom of using XHTML at all.)
To me it seems like XHTML was never quite finished
From a design specification point of view, XHTML is fine. Where it falls down is in implimentation, since browser support is either substandard (Mozilla) or non-existent (IE). XHTML 2.0 will be no different; like I said, the design specifications for the technologies aren’t the problem. But the W3C doesn’t make browsers, and it can’t force either browser-makers or users to adopt its specifications. The same thing, incidentally, goes on elsewhere. The whole XHTML/HTML thing reminds me very much of IPv6/IPv4.
I’m not sure if you know the history, but about ten years ago people started panicking that we would run out of IP addresses. This mostly had to do with the fact that there were suddenly exponentially more devices attempting to connect to networks than the designers had imagined, and the fact that large swathes of addresses had been assigned to single entities (e.g. Xerox). Now, according to IPv4 router specifications, every machine on a network should have a unique address. But there weren’t, and aren’t, enough addresses (4,294,967,296 to be exact) and so ‘interim’ measures such as NAT and dynamic IPs were invented. NAT is the system pretty much everyone in the universe now uses where you have one ‘outward’ facing machine (the gateway) which has your internet IP, and an internal network using one of the reserved private networks (usually either 192.168.xxx.xxx or 10.18.xxx.xxx). Now, this technically breaks IPv4 specifications but it works, and everyone uses it.
Meanwhile, some busy bees were working on IPv6, which uses more bytes and therefore has more potential addresses. The problem with IPv6 is that it is not compatible with IPv4. At all. People started to realise that to implement it, you’d need to have two separate internets; the IPv4 internet and the IPv6 internet. The Catch-22 was the no-one would shift address spaces unless there was content in the new internet, and there’d be no content in the new internet unless people switched address spaces. So various tunnelling technologies were invented (passing IPv6 packets through IPv4 networks and vice-versa), which bought a whole plethora of other issues. Most hardware nowadays supports both types of packets, but IPv6 is still not widely adopted. The biggest push, incidentally, is coming from China with the US government (specifically defence-related areas) a little bit further back.
Meanwhile, everyone else was used to NATing and dynamic IPs. Some people are now starting to question whether they’d even want every machine on their local network to have a universally unique IP. Local networks give LAN machines a measure of anonymity, and most network security in this area specifically relies on gateways breaking public and private networks. Not to mention what’s the point of having unique IPs on, say, an air-gapped network?1
The point being, it’s more or less the same argument as is currently going on about HTML/XHTML. Sure, standards are great and whatever, but how much point is there if they don’t actually work in the Really Real World?
… plus I also just like kicking iconoclasts in the shins. It’s a bad habit of mine.
- A network that is not connected to anything other than itself. If you connect two GameBoys together via a cable in order to share Pokemon, this is an air-gapped network. (Okay, actually it’s not, but it’s a good enough example) ^