ietf
[Top] [All Lists]

Re: I-D ACTION:draft-etal-ietf-analysis-00.txt

2002-03-29 07:00:03
On Mar 28, Ian Cooper <ian(_at_)THE-COOPERS(_dot_)ORG> wrote:

True, though I thought LOC counting was done as an initial metric until 
(much) better things were found.

Actually it's what poor management did until (much) better management were
found to replace them. LOC has no value to anybody other than people who
want to fire you for replacing a 2000 line C++ program with one line of
shell script.
 
I found the parts of the document that would enable more subjective 
measurements (like documenting the progress of documents within the group) 
more interesting than the actual "counting".

The question is, why measure? Management in large corps measure things
because they have nothing else to do and would rather look at a series of
pie charts than actually go and talk to guys on the shop floor. What is the
objective perceived in measuring anything around the IETF? If you read the
draft, pretty early on we encounter the following:

   "Historically, the IETF has an excellent track record; however, as it has 
   grown, there often is concern that IETF efforts are becoming less 
   efficient and perhaps less effective.  To date, the community has relied 
   on its subjective sense of this change.  Evaluation techniques are needed 
   that are both objective and useful."

At no point is there an explanation as to why these evaluation techniques
are required - to say there "is often concern" is a bit of a cop out. Now,
I'm not saying this shouldn't have been written and the authors have wasted
their time, but am I the only one who thinks this smells a little of an
attempt at the over-engineering of a voluntary group? Not a criticism, just
an observation, and I'd be grateful for input on why measuring performance
is something that should be applied to the IETF. Oh, actually, I think it's 
a great idea if we're going to link it all to performance related pay... :-)

-- 
Paul Robinson