On Mon, 2012-07-30 at 02:33 -0700, Dale Wijnand wrote:
[…]
> There is a research that seems to prove that "unit testing isn't enough,
> you need static typing", which is very interesting. Here:
> http://evanfarrer.blogspot.it/2012/06/unit-testing-isnt-enough-you-need.html

Thanks for this pointer, I shall be reading the full paper later (mainly
because it seems interesting but also because he used Python), for now I
have just read the blog post.

In the context that this is a masters project not a doctoral project or
a big funded experiment, indications are it is a nice piece of work.
However a number of observations:

— the original 3 points would never be held by anyone who was being
sensible.  1 should always read "detecting all bugs" and 2 and 3 are
just wrong, even from the perspective of a proponent of dynamically
typed languages. So given these as starting focus of the hypothesis, the
outcome of the project was never in doubt!

— the experiment is only one way dynamic → static, to come up with a
strong conclusion there needs to be static → dynamic as well. Also some
more controls are needed to turn a good personal masters thesis into
strong evidence to drive an industry.

— are type errors actually important in the codes under examination?

— the conclusion is phrased well, and is entirely agreeable in the
context of this thesis.

— I have no intention of giving up using both dynamically typed and
statically type languages.
-- 
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[email protected]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [email protected]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to