Ada UK Conference 2006 (slides & videos)

Good stuff here, especially if you want to learn more about Ada 2005.

Robert Dewar's talk on Ada 2005 & high integrity systems should be of general interest. The main argument is that quality depends at least as much, if not more, on the language culture than it does on specific language features. Dewar presents some of the elements that contribute to Ada's success in the mission critical world which demands high quality software. Among them: emphasis on readability, static typing, the package construct and the ability to subset the language (e.g., SPARK and RAVENSCAR). The talk isn't about the technical aspects of these features, but on how they influence the way programmers approach their tasks.

Alan Burns discusses the Real time issues in Ada 2005. This revision of the language adds many important features (such as CPU monitoring and accounting, budgeting for the execution time of groups of tasks, and new scheduling polices including non-premeption, round robin, and Earliest Deadline First) to an already rich concurrency model. The non surprising conclusion is that Ada is the best language for RT programming not only on earth, but in the entire universe...

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

A few comments on Dewar's talk

"The Ada compiler is a pain, it’s really hard work to get a program past it, but I have to admit that when I finally succeed, my program usually works first time," is something that is frequently said about ML-related languages. I've even said it myself. But hearing it about Ada is a little creepy.

That the culture around a language has as much to do with the successful use of the language than any set of features is a very valid point, but doesn't it conflict with the use of subsets of the language?

Not really

doesn't it conflict with the use of subsets of the language?

Not really. The core elements which define the cultures are things that cannot be subsetted out, like the type system or the package structure.

There's a lot to be said for this

I think there's a lot to be said for otherwise minor features of the language like style of syntax, and naming conventions in standard libraries etc., having a surprisingly significant influence. There are a lot of subtle things that languages can do to encourage practices in culture that forms around the language. I think the arguments made in favour of Ada here are good ones. Consider other languages that are regarded as a good choice for high quality software like Eiffel: again you have a syntax that focuses on readability, and a strong static type system. Where Ada has the separation of package spec and pacage body to encourage a certain amount of problem analysis and specification first Eiffel has Design by Contract which, while different, encourages similar practices.

Good

The non surprising conclusion is that Ada is the best language for RT programming not only on earth, but in the entire universe...

I am happy to note that this claim has remained uncontested...