Squeak Oversight Board minutes – 7/19/11

Attending: Chris Muller, Bert Freudenberg, Colin Putney, Randal Schwartz, Levente Uzonyi, Jecel Assumpacao Jr., Chris Cunnington; and Craig Latta

-Special guess speaker at this meeting Craig Latta discussing Spoon

-Spoon looks like this: You have an open 4.2 image with a browser called the Remote Browser. It looks into the miniature image that exists outside the 4.2 image. You can perform changes on the miniature image from within the 4.2 image.

-Adding code from the 4.2 image to the miniature image is called “imprinting”. Methods are sent over from the 4.2 ObjectMemory a method at a time. They are given an identifier before being sent to the remote image. Some additional information such as “Class” is added to the method before transport. A client in the 4.2 image sends the method to the server in the other image.

-The native state of a Spoon method can be passed between any dialect of Smalltalk as a pure piece of ObjectMemory called a module

-The inter-image communication is accomplished using a socket package called Flow, which does not seem to appear on SqueakSource, and yet is a project over a decade old

-Spoon “imprinting” is accomplished, because some changes are made to the virtual machine with the client.

-The process of images talking to each other is called remote messaging and it allows both images to know what methods are present in each image

-Spoon has modifications made in the garbage collector. Each method in the micro image has a bit that gets flipped if that method is actually used. The GC can then delete any method that was not used in an execution. This turns 99% of unwanted ObjectMemory to dust.

-The question of Spoon arose at the SOB last meeting when the issue of Cog becoming the official vm for Squeak was raised. This would be an important transition, as older images may not run on the new virtual machine. This issue attracted an older idea previously discussed of making a Squeak based on Spoon.

-It looks as though the work on Spoon could be adopted incrementally. Support for the possibilities raised by Spoon in the SOB is high.

-An Agenda Item For The Next Meeting: Software Freedom Conservancy; first impressions playing with the Spoon release

Squeak Oversight Board minutes – 7/05/11

Attending: Chris Muller, Bert Freudenberg, Colin Putney, Levente Uzonyi, Jecel Assumpacao Jr., Chris Cunnington

– SOB members going to ESUG next month: Bert Freudenberg; Chris Cunnington

– A review of how Squeak works with the Software Freedom Conservancy. Now that the issue of the new license is in the past, annual routines with the SFC ought to be outlined

– The SOB is enthusiastic about the work Casey Ransberger is leading to create a Hudson server for the trunk image

– The direction of virtual machine development was discussed. This encompassed several issues and will need to be explored in future meetings: when the standard release of Squeak will include the Cog VM; a single VM can now run a range of image types; the Linux distribution of the VM seems to be slightly more problematic than the other two standard operating systems for distribution

– An Agenda Item For The Next Meeting: Spoon.