Hi Göran--
> Craig Latta wants ids and perhaps they would improve things...
Just to clarify, I want to use object references within a single object
memory, and UUIDs to refer to objects in other object memories. In
Spoon, as far as the system is concerned (virtual machine, compiler,
etc.), each class can have any name, so there is no need for namespaces
per se.
-C
--
Craig Latta
improvisational musical informaticist
craig(a)netjam.org
www.netjam.org
[|] Proceed for Truth!
Hi--
With help from John McIntosh (thanks!), I built a new MacOS (Carbon)
Spoon VM. Stéphane had pointed out that the previous one had very poor
graphics performance (when running a snapshot with graphics support :).
This one feels "normal" to me. There are no new Spoon VM changes here,
so I'm not sure why the last one was so slow.
http://www.netjam.org/spoon/releases/current/bits/MacOSSpoonVM.zip
thanks,
-C
--
Craig Latta
improvisational musical informaticist
craig(a)netjam.org
www.netjam.org
[|] Proceed for Truth!
--
Craig Latta
improvisational musical informaticist
craig(a)netjam.org
www.netjam.org
[|] Proceed for Truth!
Hi--
After a hiatus to work on a natural-language interpreter ("Quoth",
premiered earlier this month at Transmediale in Berlin), I got back to
the task of making the first Spoon modules.
The very first one is a sort of "hello world" for graphics. It does this:
***
Display
fill: Display boundingBox
rule: Form over
fillColor: Color red
***
Since my target, before loading the module, is headless without any
graphics support at all (not even BitBLT or DisplayScreen, see
http://www.netjam.org/spoon/notes/survivors for a list of the surviving
classes), getting there is a bit of a wild ride. My plan is as follows:
---
1. implement bi-directional message sending
Up until now, I'd just been sending remote messages *to* the headless
system from a headful one (mostly system browsers ripping things out of
the headless system). Now I want to send messages in both directions, so
that I can...
2. make remote debugging work
I could sorta limp along trapping messages not understood in the C
debugger, but that has gotten tedious. :) Also, I'm pretty sure I'll
want to be able to use a normal Smalltalk debugger to debug processes in
the headless system from the headful one, to debug the inevitable
initial glitches with...
3. imprinting onto the headless target
So far my imprinting experiments have been between two headful systems
(so that I could debug comfortably). Once this works with a headless
system, I can...
4. imprint the "hello world for graphics" expression onto the headless
system (creating a module for all the required behavior as a side-effect)
Then I can...
5. run the "hello world for graphics" expression in the headless system
(and transfer its module to other headless systems)
---
Today I made part one work. I successfully sent a message from a
headful system to a headless system, which answered the result of a
message sent from the headless system to the headful system. It's an
exciting step!
thanks,
-C
--
Craig Latta
improvisational musical informaticist
craig(a)netjam.org
www.netjam.org
[|] Proceed for Truth!
Hi--
I've named Spoon's module system: "Naiad", which stands for "name and
identity are distinct". Should make for more fun logos, as well as being
another sorta-water-related name (joining Flow and Spoon).
-C
--
Craig Latta
improvisational musical informaticist
craig(a)netjam.org
www.netjam.org
[|] Proceed for Truth!