• From: rjelliffe@a...
• To: "Richard Salz" <rsalz@u...>
• Date: Fri, 27 Mar 2009 04:29:17 +1100 (EST)

> Any comments on this blog entry?
>         http://blog.jclark.com/2009/03/getting-involved-with-m.html

It reminds me of a more declarative version of OmniMark cross-translation.

Omnimark had two coroutines which could communicate, the first being a
regex-based parser which parsed text in and generated tagged data out.
This was then the input to the second stage, which parsed the input as
SGML (unminimizing, validating etc) and then generated text out.

Think of it as like an Awk (or Perl) feeding and XSLT 2 processor, where
full SGML was the input and where the context of the XSLT 2 processor was
available to the Awk front end. (I.e. smarter than just one-way pipes.)
Add that it was all streaming, and fast, and cross-platform, and it was a
great product/language.

Microsoft used OmniMark for CD-ROM products such as Cinemania, so perhaps
M is the memory of it.

OmniMark was very useful, particularly because it altered the barriers
between what information could be parsed and how to handle massaging and
bringing strange-notation data into the markup realm.

Another thing M reminds me of is Jeni Tennison's DTLL, which is part of
ISO DSDL. It is concerned with parsing data content.

Cheers
Rick Jelliffe

• References:
  • M as the next XML?
    • From: Richard Salz <rsalz@u...>