| The promise of interoperability would be a lot more convincing if we could get the Soapmill implemented in other languages. See Implementing Soapmill for details. | |
| C and C++ | This is a good start, especially for natural language applications (because speech recognisers are often written in C). |
| Perl | SOAP implementations like SOAP::Lite could make this really easy. |
| Exotic Languages | Standards can give relatively exotic languages like Eiffel and Haskell room to flourish. These would benefit from a Soapmill implementation. |
| Benchmark tool | We need some means of determining if our Soapmill implementations are fast enough. |
| Soaphub | A simple means of broadcasting Soapmill messages. |
| SoapmillUnit | A functional tester for Soapmill agents. Something which automatically sends test data and summarises their results. |
| Soapswitch | Once Soaphub is complete, we should improve over it by only sending messages to interested parties. |
| Features which ought to be added to the Java (at least) and other implementations. See Implementing Soapmill for details. | |
| SOAP Performance | The Java implementation needs improved SOAP sending/receiving performance. |
| Unique Agent Names | Soapmill messages should have unique names for each agent and message. |
| Subscription Lists | Soapmill implementations should have some means of announcing what message types their agents are interested in receiving. |