Embedded s/w engg. tools

Whats wrong with UML? Its not that bad once you get use to it.

--

Wing Wong. Webpage:

FAQs about me: Are you and undergrad? Yes Are you looking for work and do you want to work for us? Yes, but only if it pays. Are you insane? No, not at the moment.

The two most useful tools in the engineering of embedded software (as with any other software) are a decent set of requirements and a project plan. Both of these can be created with a word processor of which there are many free ones out there.

After that, I'd go for (in no particular order) Fagan inspections (not free, I'm afraid, as peers' time is precious, but very cost effective, a good configuration management system (not just a program though it can utilise one - Perforce is free for single use, else use CVS or RCS) and a test plan.

This is what you asked for (software engineering tools), but not what you wanted. As far as documentation tools go, I've found that RFFlow, while not free is quite cheap and integrates nicely with word. It's *just* a diagramming tool, though and doesn't perform any static checking on your models or code generation as some UML tools do. But it can do SDL, flowcharts, data flow diagrams, entity relationship diagrams, state charts and that sort of thing. Visio can do the same sort of things.

-- Alf Katz snipped-for-privacy@remove.the.obvious.ieee.org

I find that a whiteboard and some coloured marker pens are useful, as is a large supply of coffee and a product manager who's willing to listen.

pete

Hi Thanks for those wonderful and ingenious answers. Actually I heard about one tool called "Rhapsody" and found it's features to be interesting, although I didn't try that. It's not that I am against UML but I'd be glad if I can find a similar tool that isn't UML-based. Thanks again.

That is a VERY good tool but it depends on your target.

For small systems using C I think there are better (lighter) tools available. C can be used as OB or Object Based and i-Logix can support this as opposed to full OO. However, call me a ludite, I still prefer for use a structured methodology and modular techniques for C with small MCU

/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ \/\/\/\/\ Chris Hills Staffs England /\/\/\/\/\ /\/\/ snipped-for-privacy@phaedsys.org

Although I'm *not* a big fan of graphical s/w engineering tools, I have to ask you *why* the aversion to UML?

IMHO, one of the problems with UML is that the vendors have tried to specify it in such a way that it can be compiled. I also believe that the UML meta-model is such that a vendor can use their own visual representations and still claim to be UML compliant. The tool formerly known as Object-Time and (last I heard/cared) now known as Rational Real-Time (or something along those lines) is a good case-in-point.

Object-Time was based on the ROOM (I forget what that stood/stands for) Method. I was impressed with the methodology that I read the ROOM book.

However, Object-Time was way beyond my means (and the means of my employer at the time.)

After the UML conference several years ago, I developed the following motto:

"Tools are for fools."

Which, although a *bit* overly general, neatly expresses my sentiment wrt the graphical s/w engineering tools of the time.

These days, I use UML, and some of my own personal graphical conventions for conveying analysis and design ideas graphically.

Such models will never compile, but they work great for capturing and describing both architectural issues (encountered during analysis), and both behavioral and structural ideas during design phases.

My favorite graphics are the UML state machine descriptions, and of course class diagrams.

I also use circles, arrows, and a few other graphics, (both solid and dashed line variants), to convey ojects such as threads, signals, interrupts, and inter-thread communications objects.

All of that long winded non-sense to say that I agree the best tools include the whiteboard and markers ;-)

And don't forget a camera or a white board scanner to capture it!

The camera has never worked very well for me.

I usually sit down with the laptop and capture to Visio, or more recently Open Office Draw.

I've always lusted after one of those huge white boards with the vector drawing capture. I have no experience with their *actual* usability, however.

Not sure about the purely electronic ones, but the electro-photographic (?) ones that can print your doodlings are very unreliable. Every time I've been in a conference room that has one, I've been told it's out of order. And now I work for a company that can afford such toys, I find that there are two in my building, on my floor, and neither of them work either.

We have two and both work fine (different brands). We use them all the time to remind people of the outcomes of meetings (diagrams, todo lists, designs, etc). One is fourteen years old and needs the "darker" button pressed. The other has a computer interface to send the scan to your PC, neat but not often used.

Clifford Heath.

There is a low cost ($80 or so) package from Pixid called Whiteboard Photo that does a great job of turning a digital camera image of a white board or a sketched sheet into a clean image. Even corrects perspective error on the original image.

w..

I've used both - one a 4'x8' with printer attached and one with wireless pens with a reciever suction cupped to the side. The printer one was the best. The wireless one ... the rule was write slow and big for it to capture with any legibility

Jim