On Jan 21, 2:53=A0pm, ol...{at}web.de (Oliver Schmidt) wrote:
> >That still seems to be a bit excessive to
> >type in. =A0The Apple /// bootstrapping code is 77 bytes long, and I
> >still feel guilty about that. =A0I wonder how much we could strip out...
>
> As you can imagine this is already stripped very far (i.e. without
> ICMP / ping). There are only few bytes left to gain. You could however
> change the way you get the bytes into RAM. Maybe on the Apple2 the
> monitor would be preferable to BASIC. On the other hand you'd have to
> abandon the checksum checks then - which were quite useful to me.
> Although I did it very carefully, I had three typos...
I agree, it's a lot of typing to get right on the first try.  There's
not a lot of margin for error.  Checksums would be valuable.

> Anyway I personally think it's worth being ported - and integrated
> into ADTPro, doesn't it?

It would be a simple addition on the server end.  Once the comms
mechanism is set up, we just pump the bits out on the wire, just like
with the other bootstrap communications methods.

It might be tricky locating the code on the Apple side.  It would have
to hide in a place where incoming code and ProDOS booting wouldn't
touch it.  There's two chunks to download... the ProDOS MLI and then
the ADTPro executable.  In between the two, we have to re-invoke the
minislave so it can pull in the second part.

Anyway, yes, interesting enough to pursue. :-)  (Sorry, BLuRry, for
hijacking your thread...)
--- SBBSecho 2.12-Win32