Comments on: Python vs TTCN-3

By: kay

kay — Mon, 27 Apr 2009 14:31:50 +0000

Pekka, the predecessor of TTCN-3 can be tracked back to the late 80s. TTCN-3 can be considered as a pre-OO and pre-UML attempt to define something like “model based testing” and it is still widely used for conformance tests in some industries. It is obvious that the authors are biased and try to promote TTCN-3 and there wouldn’t be anything wrong about it if their arguments weren’t just flawed.

By: Joseph Lisee

Joseph Lisee — Mon, 27 Apr 2009 01:35:39 +0000

They also said that python has no Eclipse plugin which is just not true. Pydev is a great eclipse plugin for Python and I use it all the time.

By: Ionel Maries

Ionel Maries — Sun, 26 Apr 2009 15:36:17 +0000

I would dismiss the comparison on the spot. It’s full of misconceptions about python and typing theory. Take for example this outrageous statement in the pdf slides: “Known limitations of Python: OO language but not strongly typed” - python IS strongly typed AND duck typed. I wonder if the authors took any classes on typing theory.

By: Pekka Klärck

Pekka Klärck — Sun, 26 Apr 2009 12:48:15 +0000

As a tester/developer I’ve always wondered why people want to separate testing and development in general, but having a separate testing language is just insane. A general purpose language with testing related modules/tools is so much better idea.

About the comparison: The example on page 21 with trying to use two __init__s and complaining that only the latter is used makes it obvious that these guys don’t know any Python. Default values for methods are explained pretty early in any tutorial — and they are so much easier to use and clearer than having multiple methods with different signatures.

A simple solution for ‘‘ and ‘?’ wildcards in Python is using fnmatch module (with a caveat that ‘‘ doesn’t match newlines).