Beginning to think about nstcl 2.0

Inspired by Snit reaching v1.0, I've started to think about what could go into a version 2.0 of nstcl.

nstcl 1.2 has proven itself fairly stable. It replicates a lot of AOLserver and OpenACS APIs faithfully. Literally faithfully. The database APIs are typically the main reason people use nstcl, in particular the db_* commands, since you don't have to mess around with database handles or result sets, etc.

But under the hood, in version 1.2, the db_* commands are all built on the same ns_db and ns_set primitives from AOLserver that the original db_* commands in OpenACS use.

It works, but given that the ns_db commands merely wrap Tcl extensions for each particular database, I suspect there is effeciencies that would be achieved by writing the db_* commands directly in terms of the C-based database extensions, instead of in terms of ns_db, which in turn wraps the C-based database extensions.

Aside from performance enhancements, I don't have a lot else in mind—it's a fairly mature and feature complete package. However, I'd love to hear any suggestions anyone might have for additional enhancements beyond speed improvements for the database API.


—Michael A. Cleverly

Permanent URL for this post: http://blog.cleverly.com/permalinks/140.html