Any experience with Rhapsody in embedded systems ?

I have been using Rhapsody on ARM9 processors. There where no (hard) real-time requirements so far. But this will change now. The processing power and available memory may not be the problem.

Coming from a more desktop-like environment 10ms now seem to be like eternity. As for now one simple could do a tm(1000) [aka 1sec timeout] in a statechart. It wasn't really important if the timer expired 1000ms plus/minus something.

But now there will be really tough timing constraints. I wonder if Rhapsody still is an option (given a decent RTOS underneath it)?

Reply to
int21h
Loading thread data ...

Op Wed, 25 Jan 2012 16:45:37 +0100 schreef int21h :

Rhapsody is successfully being used on 8- and 16-bit processors.

The precision with which timeouts are handled depends on the Operating System Adapter Layer being used.

--
Gemaakt met Opera's revolutionaire e-mailprogramma:  
http://www.opera.com/mail/
 Click to see the full signature
Reply to
Boudewijn Dijkstra

You mean IBM Rational Rhapsody? I know of folks who have used it for embedded development in an environment where deadlines were at least somewhat important.

I suspect that, like any other high-level language, the key lies in either (a), testing the snot out of it, (b), taking IBM's word for it that it'll work well, or (c), carefully inspecting the code that it produces to assure yourself that it doesn't do anything really stupid.

Or use it for all the soft real time stuff, and code the important stuff by hand. Few of the real-time systems that I've worked on have needed "real timeliness" for every task, or even the majority, yet I've seen many where tasks with less of a need for meeting deadlines were allowed to seriously interfere with the higher-needs tasks (mostly as a function of confused thinking, which is most trying when developers confuse the assigned priority of 'their' tasks with the size of their gonads).

--
My liberal friends think I'm a conservative kook.
My conservative friends think I'm a liberal kook.
 Click to see the full signature
Reply to
Tim Wescott

It is. And here's Grace Hopper with a nanosecond...

yeah, it is always an option. If all else fails, pay Bruce* his rate to consult. But for things like ObjecTime or Rhapsody, one must merely make the external ... environment match what's expected by the tool to map to something. One Rhapsody claim is that you can run it sans an O/S at all.... as I recall....

*or whoever he is these days...

I think the libraries upon which it depends are in source code form ( or can be ) so....

formatting link

And another option is to have the hard-realtime stuff be external to Rhapsody. Won't be easy, but nothing worthwhile ever is.

-- Les Cargill

Reply to
Les Cargill

It's not bad. I get the feeling it'd be easier to just build something in 'C', given message sequence charts to start with. But if you need the provenance* for some reason, you get those for free out of the running system.

*It makes real purty pictchers n' stuff! He'pful when you are afflicted with "systems engineers".

The only real weakness is UML itself, which can be odd. Well, the usual C++ hair-pulling....

LOLz.

-- Les Cargill

Reply to
Les Cargill

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.