Python as a Testing Language - TTCN-3 Comparison

Bernard Stepien bernard.stepien at
Tue Aug 12 16:04:00 CEST 2008

Hello all,


            It looks like this comparison I did between TTCN-3 and Python
sparked some interesting reactions. I thank you all for that, this is very


There are two things that are missing:


1.       this comparison you downloaded are slides from a presentation at a
TTCN-3 user conference. Obviously, the lyrics are missing. We are working on
a paper on the same subject that will have considerably more explanations.

2.       the use of python dictionaries was in deed looked at and we decided
that it was even more hopeless than objects, thus we decided, mostly because
of time limitations for the presentation (still was 1.5 hours) to drop that
subject altogether. It however will show up in the paper.


Finally, no matter what you are doing (python test framework, etc.) there
are two important things with TTCN-3 that you don't have with Python:


1.       TTCN-3 is an international standard that comes among other things
with very precise semantics, thus everyone in the world using it will talk
exactly the same language. This also reduces considerably the amount of
documentation you need for the next of kin after the developer that has
developed a test suites or tool moves on.

2.       TTCN-3 is strongly typed, Python is not. No matter how good a
programmer you are, one bad day, the lack of strong typing will be extremely
costly. I think pointing out these differences is far from unfair.


Testing is a very precise activity. Python, due to its lack of typing and
semantics is by definition not precise. Thus, in that domain it is better to
use TTCN-3. However, this doesn't exclude Python. Python can be used very
efficiently for the codecs as another paper by Industry showed at that same


Bernard Stepien

University of Ottawa

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Python-list mailing list