[Cuis-dev] Morphic's design

rabbit rabbit at callistohouse.org
Sat Aug 26 08:44:04 PDT 2023


Thanks for the PDF, Ken. It is printing for later careful reading.

I would note that I was thinking there would be only 1 UIKitty for a single-threaded UI, although there would be concurrent eventual sends in. Reactive eventful and all. So the UI would be eventual.

Please run the code below to load Crypto and to see the BoxOfKitties model, without any Spectre class/method generation.

Here is the latest Crypto comment:

> massive refactoring. 1 tests FAILS: ELinda is currently broken: transient Outs. I tried to stabilize BoxOfKitties for demo. Not usre where it is regarding the InMemoryListener used in BoxOfKitties class>>#exampleBoxOfKittiesEventual.
>
> YMMV

It is broken, as it is only displaying 1 litter of 4 Kitties, not the full 16 and the boxOfKitties ivar is not being set. Not am I demonstrating Neighbor sends. WIP. YMMV.

> Installer ss project: 'Cryptography'; install: 'Crypto'.
> ((Smalltalk at: #BoxOfKitties)
> perform: #new:identity:
> with: 16
> with: ((Smalltalk at: #SessionIdentity) perform: #sampleVatId1)) explore.

I tried in the following code, but it is also broken. I'll get back to you...

> BoxOfKitties exampleBoxOfKittiesEventual explore

• rabbit ❤️‍🔥🐰

On 8/26/23 09:55, ken.dickey at whidbey.com wrote:

> Another solution strategy is "take it out to the next level and solve
> the problem in the wider context"
>
> Microkernel OSs "send messages" to implement device drivers in user
> mode.  So a crashing driver (think debug!) does not trash the kernel.
>
> Work has been done to allow _multiple OSs_ to share a window system.
> https://genode-labs.com/publications/nitpicker-secure-gui-2005.pdf
> With the proper support, having the rendering "driver" as a separate
> process which gets drawing messages from Morphic objects could be an
> interesting project.
>
> Making it "fast enough" is something the microkernel folks have been
> working on for some time..
>
> $0.02,
> -KenD

--
••• rabbit ❤️‍🔥🐰
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20230826/a5f34fb1/attachment-0001.htm>


More information about the Cuis-dev mailing list