risc or cisc

[Ian Clatworthy]

Robert Collins wrote:

> I think a mini language is definitely more complex *today* but we're
> adding more and more methods - and the complexity of creating good
> abstractions is getting higher and higher.

I'm a fan of and have implemented the mini-language approach several
times in the past. The application spaces were different (SCADA and
enterprise OLTP) but the root issues were the same:

* latency is a performance killer

* result integrity (i.e. using the same read lock so that the
  information returned from different 'RPCs' is consistent).

I think the language can be very simple. There are 2 key features
required to minimise round-trips in my experience:

1. parameter chaining - passing the result of one RPC into a latter
   RPC in the block

2. conditional execution.

That's not to rule out looping and other magic but they aren't
typically needed in the initial implementation. And if you need
*that* complexity, sometimes it's worth just adopting some existing
template technology from the web space (and paying the xml parsing
cost) than rolling your own.

Hope that helps,
Ian C.