This brings up the point that both Lisp and Smalltalk are sort of mini-databases for their respective objects. This is something I’ve always known in the back of my head, especially reading about debates on how to hash this object or that object, or how to avoid too many collisions in a very large image. But this sort of brings it out, unpacking the concept, making it more explicit. I do remember something in Beck’s Guide or one of his posts in the Portland Pattern Repository (the original WikiWiki) that brought this point up. Let the image be your database until it becomes necessary to move it to a persistence layer: you ain’t gonna need it?
9:18:27 PM  #  comment []

Jay Han usually tosses things into his Scrapbook without comment, so I sort of skimmed over this really cool sounding thing, only noticing it when I went through the monthly archive:

The Prevayler persistence layer, where all objects reside in RAM. No databases! There’s support for Squeak, yay! How do they handle all that ACID stuff, snapshots, and replication? They have an FAQ for skeptics. The site is a Wiki, so you can explore and debate the concept in the rather sloppy fashion that Wikis tend towards. Looks like it evolved from a Kent Beck Smalltalk spike, though this is just a guess. Also the Sourceforge page.
8:45:38 PM  #  comment []