Lists Home |
Date Index |
Claude L Bullard wrote:
> I remember some discussion about ASN.1 and SGML
>'getting together'. It died in the rush to HTML.
The ASN.1 and SGML debate raged *long* before HTML was even an
idea... This stuff went on in the early to mid 80's. By the early
90's, when HTML was created, and OSI was breathing its last few
breaths, the sides had been so firmly drawn that the issue was dead...
Personally, I think the key strategic problem was that ASN.1
defined a text-based "value notation." This meant that whenever I
argued for TER or "Text Encoding Rules" as a way to make peace with
the IBM guys who were pushing SGML, the response was always something
along the lines of "We already have TER -- it is value notation. We
don't need two text based encodings." or "Value Notation is 'better,'
since it is more expressive, or more readable, more <something>..."
If ASN.1 hadn't had value notation, it would have been
massively easier to argue for TER and it would have been easy to just
say: "Let them use angle brackets rather than curly-braces..."
Let this be a lesson to people -- if nothing else, that a tiny
little thing can result in decades of useless debate. SGML should have
been ASN.1's TER (textual encoding rules) and we should have specified
TER in the first releases of ASN.1. But, I think the combination of
people's pride plus value notation made a mess of things.