LLL

[Mike Prince]

On Thu, 22 Dec 1994, Dr. J. Van Sckalkwyk (external) wrote:

> IMAO (in my arrogant opinion), there seems to
> be a potential design problem. Mike seems happy
> to use C for the initial coding for the
> "common language" that will run on all machines
> where we implement "Joy". I am _not_ happy
> about this decision mainly because:

>     (1) I am not sure that C is up to it!;
>     (2) I believe that this will bias our
>     design and lead us to ignore viable options;
>     (3) I think it violates KISS.

I hope I'm making myself clear.  I'm using C with assembly patches 
because it's faster to write things than in pure assembly, and stands the 
chance of having some portability.  C is not the common language.  I want 
to use C to create a demonstration kernel, that we can use test some of 
our ideas and refine our specifications.  Later we can write the kernel 
in whatever we want.

My focus is the design of a LLL (this won't be tied to C at all), the 
specifications for a transport protocol between kernels, and ties to an 
HLL we will write our code in.

I'm just for C because it will get us off the ground quickly.  I want to 
keep the kernel really small (mine is 20K+40K of borland libraries, which 
includes a debugging interface, simple simple compiler, and a file system 
interface).  If you went for "just" a kernel you could do it in under 20K 
easy.

Because of the ease in rewriting the kernel, I don't think it's that big 
of a deal right now what we program it in.

I hope I'm not reiterating stuff which you already understood.

Please let me know if there's still a problem,

Mike