From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news1.google.com!news3.google.com!feeder1.cambriumusenet.nl!feed.tweaknews.nl!194.134.4.91.MISMATCH!news2.euro.net!feeder.news-service.com!tudelft.nl!txtfeed1.tudelft.nl!multikabel.net!newsfeed20.multikabel.net!newsfeed.freenet.de!news.tu-darmstadt.de!news.belwue.de!LF.net!news.enyo.de!not-for-mail
From: Florian Weimer <fw@deneb.enyo.de>
Newsgroups: comp.lang.ada
Subject: Re: Software Quality in Science
Date: Wed, 10 Feb 2010 14:13:38 +0100
Message-ID: <87fx59p725.fsf@mid.deneb.enyo.de>
References: 
 <1198a288-b013-45a8-907f-7fe227e6294e@m27g2000prl.googlegroups.com>
	<04185bf3-f83a-4fbe-b380-c6d8aa4105e6@w27g2000pre.googlegroups.com>
	<wcc6366x9em.fsf@shell01.TheWorld.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: idssi.enyo.de 1265807618 13790 212.9.189.177 (10 Feb 2010 13:13:38
 GMT)
X-Complaints-To: news@enyo.de
Cancel-Lock: sha1:AxhKE0ogZK23qX7jHe8RXFIO1Ss=
Xref: g2news1.google.com comp.lang.ada:9078
Date: 2010-02-10T14:13:38+01:00
List-Id: <comp.lang.ada>

* Robert A. Duff:

>>...The net result of changing languages appears
>> to be that the overall defect density appears to be about the same,
>> (Hatton 1997). In other words, when a language corrects one
>> deficiency, it appears to add one of its own."
>
> That assertion requires evidence, and I don't see it here!

(Hatton 1997) is here: <http://www.leshatton.org/IEEE_Soft_97b.html>
For Ada, it cites Compton & Withrow, "Improving Productivity: Using
Metrics to Predict and Control Defects in Ada Software", in "Second
Annual Oregon Workshop on Software Metrics", Oregon 1990.  It then
concludes that language choice had no impact on fault density
patterns.  Of course, the evidence is only anecdotal.

But the actual cross-language studies I've seen show that most metrics
one can conceive (such as defect rate, performance, development time,
even lines of code) vary as much among programmers as among
programming languages.  This is a bit sad because it means that
language design does not really matter as far as actual results are
concerned.

Ada subsets may be helpful if your goal is to avoid the last
(relevant) bug.  But that requires matching development practices,
which are in place for only very few code bases.  Certainly, these
practices aren't magically introduced if you just substitute Ada for C
or Fortran.