On Tue, Dec 20, 2005 at 04:20:11AM -0500, Mathieu Bouchard wrote:
On Thu, 15 Dec 2005, Chris McCormick wrote:
- what's the link between LISP and abandoning vonNeumann arch ?
That architecture separates storage and processor. LISP blurs the line between the data and the algorithm.
The first sentence of the Wikipedia article on vonNeumann arch reads:
« The term von Neumann architecture refers to a computer design model that uses a single storage structure to hold both instructions and data. » -- http://en.wikipedia.org/wiki/Von_Neumann_architecture
So?
IMHO (and not in that webpage), the blurring of the line between data and algorithm dates back to the 1830's, when Charles Babbage and Ada Byron were busy inventing that concept that we now call software, under the name "the analytical machine". The revolutionary idea was that programs needed not be made of hardware (cogwheels or relays) anymore -- they could be made of pure information, which they called "machine language".
Byron also came with that idea that programs would be easier to reuse and had that literary nature so that collections of reusable procedures would be called "libraries" (that was her word).
Thanks for the history lesson. :)
- what's the link between what you are replying to and abandoning
vonNeumann arch ?
You indicate that what you want is more 'meta' and less specific.
People can be meta all they want in a vonNeumann context. It's just that most of today's programming languages have been strongly influenced by the concept of adding a strong program-vs-data distinction back into the game. (for better or worse; and yea, I don't approve)
True.
Using a small number of heavy duty processors to emulate real world extreme computational parallelism is not scalable in my opinion.
Best,
Chris.
chris@mccormick.cx http://mccormick.cx