Re: core addon packages
"Rob Withers" <[email protected]>
| Newsgroups | gmane.comp.lang.smalltalk.squeak.foundation |
|---|---|
| Message-ID | <012601c2a04b$7e139880$6401a8c0@bacchus> |
From: "Stephen Pair" <[email protected]> > You have more faith than I in the consensus building process...but sure, > the fittest solutions in each of those areas will emerge over time and > different implementations will converge on a single solution. Stephen, I do have a lot of faith in squeak and its community. The entire focus of squeak releases has now fundamentally shifted with the change in ownership of release managment to SqF and the Guides. Now I don't want to jump into the infeasibility in my excitement, but I finally see concrete evidence that the way forward for squeak lies in a phase of rigourous deconstruction of the monolithic image to result in a reconstructable base. The only point I'm attempting to make is based on the fact that there are areas of core functionality in any modern language that propably wont be in the teaniest image, but which ought to be standardized and well documented and a commitment made to preserve that API. Areas where weakness in the current squeak model exists are those areas where we see alternatives discussed or start to appear. Right off the bat, I think of the sockets api and the FileDirectory api. NB. (what does NB mean? :) I finally realized the practical implications of the flow model when I quickly looked for a way to play a streaming mp3 from the web. In the base image, i couldn't find a way to play from somafm.com nor any way to quickly hack it together. :( cheers, rob