Advice on selling Ada to a C shop
in [ADA]
|
From: usenet on 19 Jun 2010 19:01 I may have the opportunity to pitch Ada for use in some new projects (industrial equipment) at work. The place has been using C by default, not by any conscious choice, and of course they've been paying a typical price for that default. The head of engineering (who is not a software guy) MAY be open to trying Ada if presented with a compelling case. The programmers are another question - they may resist or they may be intrigued by some of Ada's features e.g. concurrency. They may also be open to improving the very ad-hoc software culture now in place, and the idea of having fewer bugs, and those caught sooner and less painfully. So, what points would you emphasize to each (manager, programmers) to make the Ada sale? I've read a lot of the advocacy and I think I have a fair handle on it, but I'd still like other viewpoints, and perhaps even examples of such a successful switch. In the end, it seems, it all comes down to money. Money saved in greater productivity and less debugging, money saved in lower liability and maintenance, money gained from satisfied customers who get a robust product, and maybe even money saved in having programmers who have a more powerful and safer tool at their disposal. I know that language choice is not the be-all-and-end-all of better software, but I'm convinced it's a key component, and I'm also convinced that the sooner we leave C and its baggage behind the better we can become.
From: Jeffrey R. Carter on 20 Jun 2010 02:17 usenet(a)scriptoriumdesigns.com wrote: > > In the end, it seems, it all comes down to money. Money saved in > greater productivity and less debugging, money saved in lower > liability and maintenance, money gained from satisfied customers who > get a robust product, and maybe even money saved in having programmers > who have a more powerful and safer tool at their disposal. Where we have hard data, they show that Ada reaches deployment at half the cost of C, with a quarter as many post-deployment errors which cost a tenth as much to fix (factor of 40 reduction in post-deployment error fix cost). The hard data include the Rational assessment for compiler development: http://adaic.org/whyada/ada-vs-c/cada_art.html and a study by Pratt and Whitney reported here on jet-engine control SW (I'm not aware of any link to this). There's also McCormick's academic results, in which no students familiar with C were able to complete a real-time project in C controlling a model railway, even when given 60% of the solution, while 75% of students familiar with C but not Ada were able to complete the same project in Ada when given < 20%: http://www.stsc.hill.af.mil/crosstalk/2000/08/mccormick.html http://adaic.org/whyada/ada-vs-c.html is a good overall reference. -- Jeff Carter "Perfidious English mouse-dropping hoarders." Monty Python & the Holy Grail 10
From: Jeffrey R. Carter on 20 Jun 2010 02:24 usenet(a)scriptoriumdesigns.com wrote: > > So, what points would you emphasize to each (manager, programmers) to > make the Ada sale? I've read a lot of the advocacy and I think I have > a fair handle on it, but I'd still like other viewpoints, and perhaps > even examples of such a successful switch. I think there is no way to sell Ada to developers. Ada is a SW-engineering language; coders don't like it. In my experience, 98% of developers are coders. SW engineers generally like Ada once they become aware of it. So if your people are SW engineers, Ada will sell itself; if they're coders, there's no way to sell it to them. Proposing a switch to Ada may however, be a mechanism to tell which of your people are coders and which are SW engineers. That may be useful in deciding whom to retain and whom to let go. -- Jeff Carter "Perfidious English mouse-dropping hoarders." Monty Python & the Holy Grail 10
From: J-P. Rosen on 20 Jun 2010 03:00 usenet(a)scriptoriumdesigns.com a �crit : > So, what points would you emphasize to each (manager, programmers) to > make the Ada sale? I've read a lot of the advocacy and I think I have > a fair handle on it, but I'd still like other viewpoints, and perhaps > even examples of such a successful switch. > Theory will never impress a manager. Examples will. For example, it has recently been quoted here the case of Meteor (parisian metro), in use since 1998, and whose software is still in version 1.0: no problem discovered in 12 years of usage. I also remember an example from a SIGAda conference (no reference at hand, does someone remember?) about a system in a Boeing airplane, which was doubly written in C and Ada for reasons of diversification. The C system had several major upgrades, while the Ada one worked flawlessly. -- --------------------------------------------------------- J-P. Rosen (rosen(a)adalog.fr) Visit Adalog's web site at http://www.adalog.fr
From: Nasser M. Abbasi on 20 Jun 2010 03:40
On 6/20/2010 12:00 AM, J-P. Rosen wrote: > usenet(a)scriptoriumdesigns.com a �crit : >> So, what points would you emphasize to each (manager, programmers) to >> make the Ada sale? I've read a lot of the advocacy and I think I have >> a fair handle on it, but I'd still like other viewpoints, and perhaps >> even examples of such a successful switch. >> > Theory will never impress a manager. Examples will. For example, it has > recently been quoted here the case of Meteor (parisian metro), in use > since 1998, and whose software is still in version 1.0: no problem > discovered in 12 years of usage. > > I also remember an example from a SIGAda conference (no reference at > hand, does someone remember?) about a system in a Boeing airplane, which > was doubly written in C and Ada for reasons of diversification. The C > system had several major upgrades, while the Ada one worked flawlessly. Showing examples are one excellent way to illustrate Ada advantage over C. But I think what is needed are small coding examples, where each case shows clearly an advantage of Ada over C. Coders like to see code. Seeing small actual code examples, side-by-side, showing clearly that the Ada way is better than the C way, and why that is, will let the code speak for itself. But these examples have to be written by someone who is both an expert in Ada and C to be fair. There is another thing to overcome: many of the Ada advantages starts to show up when one is doing programming in-the-large, and this is hard to sell to someone who mostly thinks at the function level (the so called coders here?) instead of at the system level as well. --Nasser |