My “why move away from SGML?” reason is the way that every time I have to explain to someone that their Mozilla bug in invalid because HTML is actually an SGML application [...] I finish up by saying “if you want to see the actual spec that I’ve been told says that, you can buy a copy for 230 Swiss francs.”
Sign in with 
OpenID
One of the things that is weird about having an ISO Standard is the fact that ISO depends on the revenue from the sale of standards. In the US, ANSI is the same way. This doesn't work with the current structure of the information-systems world and the need for many individuals to be able to have access to the standards that bear on open systems work.
Fortunately, in some cases the submitter organizations manage to keep their documents in sync and you can obtain free copies from those organizations, including ECMA and OASIS.
My next question is, can the SGML specification be obtained through OASIS or is it now completely a captive of ISO/IEC JTC1? I'm not sure, but I wouldn't be surprised that ISO is now guardian of the whole ball of wax.
If we are counting on ISO standardization as some sort of legitimization of open standards for document formats, we need to deal with the fact that ISO is not exactly open and public with accessible results.
This reminds me of the Prolog programming language, another closed standard that ISO is selling. From a public point of view, the standard seems to be totally opaque to external corrections, proposals for extensions, and other typical processes that would be quite visible and accessible in the case of open standards.
The funniest part is that many of the Prolog implementations out there claim to follow ISO Standard, probably counting on the fact that, since you have to pay to obtain documentation on the standard, no-one will really double check on that. But not even the most important commercial implementation is 100% compliant.
I'd be interested to know if ISO standards are more implementable than free ones. From what I understand (admittedly a shallow well to be drinking from), a focus of the WHAT WG's efforts was writing a spec that could actually be implemented consistently, as the W3C's specs had fallen short of this.
Basically: is writing good specs pretty hard, and is it too hard to do on a non-commercial basis?
(P.S. - I know the WHAT WG is a volunteer group too. I'm guessing our evidence of how implementable their spec is/will be is limited, as no-one's tried a full implementation yet.)