Subject: Re: [linux-audio-dev] SuperClonider
From: John Lazzaro (lazzaro_AT_CS.Berkeley.EDU)
Date: Sun Apr 28 2002 - 21:54:17 EEST
> Robert Melby writes
>
> first declare yourself a working group (let's say XJ5.34) of IMG
> (international made-up group). then declare version x.xx of the
> compiler/interpreter to define the canonical syntax and behaviour of
> the language spec version x.xx .
[...]
And from my experience as an early SAOL implementor and a contributor
to the Corrigenda process, the act of writing the semantics of the
language down, and in the process differentiating the semantics of
the language from the behavior of one implementation, qualitatively
changes the nature of the language and its community.
The document enables compatible implementations that are grounded in
the certainty of documentation designed to define what the language is,
not how to program in it. And it places all implementations on the same
footing -- the goal is for all implementations to match the document,
and in the process of reaching that goal, the document is refined to
the point where it says everything it needs to say to define the
language.
This is why writing a C compiler that effectively interoperates with
real code is doable, but writing a tool set to parse the Microsoft
Office binary file suite (.doc, .ppt, .xls) has proven to be quite
difficult -- the playing field isn't level because there is no document,
there's only a set of file filters Microsoft can change at will.
Note in my original message, I explicitly said that a "do it yourself
document" could meet the goals just as well as a document through an
organization -- and cited corporate examples of language documents that
took that route. I've found it beneficial to work through the IETF, and
unofficially through MPEG, because there are many smart people in those
working groups who understand how to make a standard work -- becoming
part of the process has its advantages.
> wouldn't be hard to put up a yacc grammar
Grammar is easy -- its the definition of the semantics of the language
that enables interoperability.
-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------
This archive was generated by hypermail 2b28 : Sun Apr 28 2002 - 21:55:01 EEST