On Tuesday 24 June 2008 12:28:40 Ruben Smits wrote:
> I was looking at the omg's latest specifications and bumped onto the
> recently (april 2008) released Robotic Technology Component specification:
>
>
>
> I was just wondering how Orocos(::RTT) fits into this specification?

I know this spec in some detail, there is a bug report refering to it as well

I consider it a good effort in the right direction, and allthough it got some
things better than Orocos (like the data flow calculation execute/update
semantics), it is also lacking essential features, like a 'configured' state
and there is no notion of 'buffered' ports, they only have 'DataPorts'. The
Modes chapter is imho overly complex and should be left out in favour of
classical state machines.

From a spec-technical point of view, the spec is gravely underspecified and
contains a lot of ambiguities and even unimplementable features. So far no
different than any other OMG spec...

I had sent a lengthy email to the maintainer of this spec listing all my
criticism, but I never got a response, other than it was received. The spec
was already in the finalisation stage, that's probably why the drastic
changes I proposed were not taken into account (and that my employer is not a
member of OMG as well)...

To answer your question: RTT shares some ideas with the RTC but is/will
nowhere be compatible with this version.

Peter