Kevin,
>The issue is a conclusion that can't be fully made after 3 months.

That can be a mighty encompassing statement depending on the interpretation/target of "fully". Not that I disagree that most people working in dotnet today needed a larger time span to feel "really" at home. But first consider possible mitigating factors:

- amount of previous experience with statically typed languages, perhaps even java: granted, won't give you the knowledge of the fwk, where most of learning pain is, but it might lessen fear of new systems some programmers exhibit. Even the OP mentioned it

- the calibre of the team. There ARE programmers able to produce bad code in any language and from there starts a wide bell curve.

- the time spent in the three months. I bet we both have some war stories of gigs with too much talking rounds that left us feeling too bushed to program more than a few lines after midnight - after typing more of those !%$§$§"%%&§ documents for "managment"<g>. On the other hand three months can be a lot of time, is expensive and the OP mentioned that it was all research.

- I am sure we both would bill significantly less H then spent learning, if we get such a chance - the same mechanism works with some employees (outclocking but returning to their desk), but some would still go home after the 7.33 H specified in their contract and be exact with the second decimal even when given such a chance of learning.

- How much they prepered themselves by reading some books on the topic so they can just bulldoze through the first walls they meet.

- AFAIR he went on testing with a development fwk - at least it will give you a boost at start, even if you redesign parts later on <g>.

Secondly view it from the other side: If after 3 months I am not able to give ballpark estimates good enough to base business decisions on the productivity reached then and the probable productivity gains during the next 6 months, the company I work for is either in trouble, has BIG pockets or is the government<g>.

I'ld go to mangagment, tell them about the need to test python for example, to get a dynamic contrapoint to c#/java (java would have to "tested" as well, since the java framework IS different, even if the language feels similar). Then there are dozens of people working on AJAX based solutions, why not eval javascript as global GUI development language ? There were some dotcom's founded on LISP-writing genius' - need to eval that as well. Oh, there are OCaml and Haskell - and D gained momentum as well, especially reccomended by brainy people. Would be like Disneyland for me<bg>.

MORE interesting would have been the writeup (probably 2 days of the 3 months) giving *detailed* info on the specific points leading to THIS decision - and yes, I also think that the sentence "My conclusion is that for most developers, .Net is not ready for prime time." is too encompassing a statement by far. YMMV and HMMV <bg>.

regards

thomas