Various questions

[Emmanuel Rialland]

> Yes, good question. I'd also prefer to do it in lisp. I think it would
> be a lot easier that way; not the least because of the interactive
> repl, which gives you a kind of FEP :-)
> Actually, I' ve done a bit of exploratory programming for examining
> the load-band and the microcode band in Common Lisp and it seems to go
> a lot faster. Since, at least at the moment, a lot of time an energy

Would you like to put that code in cvs (in everybody else agrees of course)?

> is still spent on figuring things out and not producing production
> quality code, I can imagin doing a dual way exploration. Why not write
> code in CL if it is easier and then later translating it into C++.
>
> As far as I can tell, the reason for doing it in C++ is, that it is
> presumably easier to get something running on the bare iron. Though I

GCL is able to produce exercutables. 

On my side, I understand the _pure emulation_ approach, trying to reproduce 
the exact feeling of a TI as it was. But I am far more interested in seing 
the virtual TI living on my desktop along with the rest. This means that:
- I am not that concerned about size
- I imagine that at some point someone (I have no idea how) will need to find 
a way to interface the e3 with the rest of the machine (corba/kpart/bonobo 
interface, drag/drop, files...) in a nice way. Creating a virtual nubus card 
+ driver could maybe do the trick?

As you can see, I am rich of ideas but poor of code. :-)

> would have done it in C then. Maybe OCaml would also be a good
> compromise for that.
>
> 'Andreas