[ reply-to set to port-hp300. ]

> > no special hooks, etc.  It would cost (on average) 4k extra disk space
>                                          ^^^^^^^^^^
> 					 Worst Case

No, on average.  "Half" of the time, you'd waste 4k padding the text
segment.  Half the time you'd waste 4k padding the data segment.

worst cast 8k, best case 0k, average (Assuming even distribution of
text and data sizes) 4k.

> Even though i never PLAN to have one, making the hp300 the same
> architecturally as the rest of the 68K system would be a big win in
> space savings in he long term.

That's unclear: yes, the hp300 port could share the other ports'
binaries.  however, converting it to 8k page size _isn't_ necessary
for this.  I do think that converting the hp300 to using an 8k page
size of user binaries would be a win, to avoid maintainence hassles,
though, and to simplify making binary distributions of software.

The point is, it'd be nice to avoid having to make a seperate m68k4k
set of m68k binaries, when building a snapshot or a release.  no, you
wouldn't _have_ to do so (because the hp300 can be made to run them),
but if you're not going to, you might as well make the 'standard'
hp300 binary format m68k8k.

> Since you would be able to use other people binaries, you wouldn't HAVE
> to keep source around for everything you want to load.  Not having to
> have a couple of Meg of source for a program that you just compile and
> go would be a thing of the past.

you don't have to keep sources around now.  ftp, compile, install,
delete.  works for me...


cgd