rwithers: Do you have a link that I could read up on this? This sounds interesting, if it is language neutral.

tyler: The WOMP was expressly designed to be language neutral and to facilitate interoperation between capability systems.

rwithers: I seem to recall following some of the threads and remember this as an important design goal. Good.

rwithers: so WOS runs on top of WOMP which runs over HTTP?

tyler: WOMP runs over any protocol that can meet its requirements. Only an HTTP binding has been defined so far.

tyler: Implementing the WOMP in Squeak should not require much work, assuming Squeak has: HTTP server, SSL support, XML library,

rwithers: yes

rwithers: and my eventual sending framework, yes.

rwithers: I want to get into the event-loop and start implementing some of the special constructs like dead-man switch and so on.

tyler: If you implement the WOMP,

rwithers: Given the lack of SSL, I think that implementing CapTP is the shortest path to an operational system.

tyler: If you don't have SSL, then continuing with CapTP would be the most efficient path.

rwithers: I think you mean the VatTP protocol. Other than the state machine, you need to know msg content and all the handshaking and finally the encryption.

tyler: Thanks. I'll save it for a rainy day.

rwithers: I'll try to respond in more detail this evening before the Thanksgiving break.

tyler: I look forward to it.

rwithers: err....no,

tyler: ...

cg: Nope, no SSL. VisualWorks has an all-Smalltalk SSL implementation, but it took them quite some time to write.

rwithers12: With the x509 changeset, the crypto package from Luciano, and some tweaking of the startup connection code, we could probably build SSL.

rwithers: Tyler, I think the network protocol is in VatTP, like you use HTTPS.

tyler: Yes, the analogy between CapTP's use of VatTP and WOMP's use of HTTPS is right. I should have used the more specific name.