[offtopic] marshalling

Al Snell alaric@alaric-snell.com
Thu, 4 Jan 2001 20:23:20 +0000 (GMT)


On Thu, 4 Jan 2001 linas@linas.org wrote:

> do so?  This seems highly unlikely to me.   First, there are social
> issues (try convincing a BSD user to use linux: the kernels are
> *almost* the same, and the libraries & apps really are the same).

It's the broken filesystem layout, generally shoddy kernel, and poor
system management tools that turn me off of Linux - but I won't go into
that here :-)

[OpenGL+phigs example]
> function in hardware.   I can't imagine that the issues confronting
> programming languages are easier than 3D.

Quite. The problem of writing a glue layer between two things is far too
easy to underestimate!

However, most programming languages deal with two basic kinds of
abstraction: objects (which have identity) with methods and attributes, OR
just a list of procedures, backed up by a library of types along the lines
of:

 - Various sizes of signed or unsigned twos complement integers
 - a few sizes of floating point number
 - fixed-size arrays of N dimensions of any other type
 - records (with a list of field names and types)
 - unions (likewise)

There is some variation - eg, are unions tagged? - but not too
much. Usually.

However, there are languages with very different data models - Prolog, for
example.

Some languages deal with the concept of time differently - most languages
you guys have used will probably have a central clock with mutable store,
eg a single stream of "instructions" that can change memory. Other
approaches are monads, linear Worlds, etc.

ABS

-- 
                               Alaric B. Snell
 http://www.alaric-snell.com/  http://RFC.net/  http://www.warhead.org.uk/
   Any sufficiently advanced technology can be emulated in software