The Pharo and Squeak communities, two major open source Smalltalk communities united technically by a common heritage and a shared VM, and socially by many members participating in both communities, have reached an interesting stage. There are several interesting and successful applications that have been implemented effectively by start-ups, by members of the communities. But these have sprouted from within the communities.
Pharo and Squeak have been steadily improving their quality and performance, and also, particularly in the Pharo community, supporting exploratory programming. These improvements make both interesting to industrial Smalltalk users who up until now have implemented their systems in commercial dialects.
The potential from porting to Squeak and Pharo by existing commercial users, or new users from outside the communities, is an exciting one. It could help both the users, who stand to enjoy more rapidly evolving and higher-performance Smalltalk, and the communities, who stand to receive significant revenue from this external use.
But completing this marriage requires two key problems to be solved. On the one hand, commercial users need an organization upon which they can rely for support and to which they can pay for services. On the other hand, the communities need fair distribution of that income if the communities are to grow in harmony and avoid the discord that will result from unfair distribution.
Organizations like the Pharo Consortium and PharoPro are first steps, but they do not properly address the need for fair distribution of income.
This talk reviews the state of the Squeak and Pharo communities, and the efforts so far to support commercial Smalltalk users. It then looks forward to some new economic thinking in how to distribute fairly revenue within open source communities.Download presentation
He is a Smalltalk VM implementor and systems programmer having written his first VM in 1983. He worked on Peter Deutsch’s HPS VM for VisualWorks throughout the 90’s and early 2000’s and essentially double its speed (mostly by adding polymorphic inline caches and rearchitecting its mapping of contexts to stacks). He was technical lead for VisualWorks from vw 3.0 through vw 7.4.1, leaving at the and of 2006. Amongst other things he invented method pragmas/method tags as seen in VisualWorks and Squeak. He spent an all too brief but stimulating time at Cadence in Gilad Bracha’s Newspeak team, and then went to Qwaq (now 3DICC) where he implemented another fast VM for Croquet, a 3d immersive collaboration architecture built above Squeak Smalltalk, but this time (unlike VisualWorks) the VM is open source. He returned to Cadence early in 2011 where he is working half-time for Yaron Kashai on system-on-a-chip design support in Newspeak running above the Cog VM. He is also collaborating with Tudor Girba in helping companies exploit Pharo and Cog.