[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Sather up here update!
Sather up here update!
Sun, 19 Nov 2000 12:35:21 +1300
Pluto/2.02b (RISC-OS/4.02) POPstar/2.02
Nobbi (and y'all),
I have been feeling a bit guilty this last couple of weeks for not
having surfaced to let you know what is going on up here- so here is brief
a. The Run-time engine is in a (final?) re-write for
efficiency/size/speed. The chap writing it leaves here in a month or
so, so it should be ready for beta testing soon. I have had an offer
to test it on a 100 CPUs - initially just with a test program before
we start mating with the compiler to generate direct calls.
b. Malome and I have produced a paper on software design consistency
- which grew out of my reslisation that there are many (small) holes
in the library (old or new!) arising from inconsistencies in
c. I have been over the whole specification again and re-written
fairly large slices. I am diving into several class groups of
specification, getting the formal bits done and the holes found. I
haven't yet got too many examples written, but I intend to update
the documentation as another intermediate issue at the end of this
d. I have re-written the cultural compiler and have wodges of new
stuff (except translations as yet) in the internationalisation arena -
some changes in resource structure (relating to the cultural specs in
text form). I want to get this out in early December if possible.
e. I am starting on a 'text-book' tutorial on Sather for my third
year system programming course next year - based on notes I issued
this year which were too brief and not sufficiently structured!! This
is likely to take most of my efforts until the middle of next year -
so any help on other things which need doing is most welcome - like
porting the really interesting SQL interface to use the new library -
and, again, looking at the inconsistencies which may be there too.
I shall be away from late December to early February in China, trying
to get my own feel for the culture in small towns and villages throughout
the South East.
I hope to have a student start completely re-thinking windowing and
window manager interaction,etc. All that others seem to have done so far
is add layer upon layer of 'interface' to get some sort of common look and
functionality. We think that it is time to start from scratch again,
looking at everything from OpenLook, X11, Tcl/Tk, Win32, RISCOS, MacOS,
SunOS, AIX, HP-UX,etc to see where the basic commonality lies and devising
a way to produce desired behaviour in a lean, mean, green way on all of
them (except, of course, Tcl/Tk which is bloatware iteself!!!).
Oh! By the way we define 'lean, mean and green' to be 'in
one-and-a-half CPU instructions - or less' - REALLY thin!
Now down to drawing another hundred or so inheritance diagrams!
[PGP key available if desired]