The goal would be to have discoverable thrift servers that can compute things, specifically I would like to know what services they have and what are the ttypes for the parameters those services need. Ideally all I would need to know is the IP and port of the server and be able to get back all this information before I use it.
Let's say the server has the following service:
service Calculator {
double add(1:double num1, 2:double num2),
double subtract(1: double num1, 2:double num2),
double sqrt(1:double num1)
}
By just having the IP and port, is it possible to know this server has these three methods, and how many variables each one takes and of what type?
Thrift does not provide an service location framework infrastructure (think WADL), simply because this is (at least currently) out of scope for Thrift. Despite that, one has a number of options indeed.
Publishing the Thrift IDL file(s) as HTTP-downloadable items, like Evernote does.
Using thrift -gen html yourfile.thrift to generate a nice HTML page containing a fully fledged description of everything what's in the IDl file, including any /** docstyle comments */ you made. The result looks very similar to the Evernote API Reference, except that the layout has improved somewhat in the meantime.