running code vs. formal testing methods

[Reinhard Scholl <Reinhard.Scholl at etsi.fr>]

Maybe not that surprisingly, searching RFCs for SDL (Specification 
and Description Language), MSC (Message Sequence Charts) and 
TTCN (Tree and Tabular Combined Notation) doesn't give a single hit. 
Whereas the IETF/Internet is based on "running code", requiring 
different interworking implementations rather early in the standards 
process, the traditional telecoms standards process relies on formal 
testing methods which come rather late in the standards process 
(conformance testing; yes, there is also a lot of non-publicized 
testing among friendly competitors). Going a step further, in ITU-T's 
INAP CS-3/CS-4 the text has only informative status; it is the 
SDL-version that is normative.

If the above is correct, what is really the reason for the difference in 
testing philosophy? Is it historical? Cultural? Is it the business model 
(a few big players vs. everyone can set up shop in the Internet)? Is it 
openness in the Internet vs. a rather closed environment in the 
telecom world? Is it the complexity / robustness of telecom systems? 
Is it because telcos have very stringent requirements, with rather 
large penalties on equipment manufacturers if things don't work?

And will the testing philosophy change? Will bake-offs a la IETF find 
their way into the telco world, or/and will formalized testing methods 
find their way into the Internet world?

Trying to get a discussion going.


Reinhard Scholl
ETSI