Re: [oc] think different (was: Beyond Transmeta...)

[Suboner@aol.com]

> I fear your ideas may be wasted if you insist on running X86 binaries
> or even recompiling conventional languages such as C.

Well I'm not necesary insisting on running X86 binaries, or compiling C 
programs, so to speak, but its nice to have all those applications available, 
with most of the work on the converters. One of the important aspects that I 
think is vital to this, is to be able to take a binary for another system or 
source code in a standard form, and to be able to create a network or 
reconfigure it to run optimally on another system. This may end up being a 
difficult thing to do so it should not necesarily be its defining thing, I 
think what is defining is its ability to network bits and to reconfigure them 
for optimal usage, and it should be able to at least be possible to create a 
network that will allow for X86 programs to run, but not necesarily that its 
purpose is to do that.

> You should consider different architectures (like Dataflow, which was a
> popular research topic in the 1980s) and different languages (like
> SISAL or similar functional languages).

I think those things can and should be considered. One thing I want to move 
away from eventually would be language, I think eventually programming could 
either move away from language in the relms of using visual means to program, 
or allow for normal english language (or other human language), so that 
someone can program simply by telling a computer what they want, or maybe 
even a combination of the 2 (some kind of star trek kind of thing).

> Take a look at the ASPLOS Wild and Crazy Idea Session '98 Submissions:
> 
>   http://www.cs.berkeley.edu/~kubitron/asplos98/complete.html
> 
> These might give you some ideas of what is possible.

Thank you. There seems to be a number of interesting topics on those pages.

Leyland Needham