attached is what I found relevant and wasn't on the slides. I can also summarize the slides, but we have them so I don't see a need...

Cheers, Axel.

Rene:
* experiments will not take next production release unless it's a no-risk release.
* need to think 3-5 years ahead, also to give motivate for resources
* weigh new technology vs. risks
* next production release will be Rene's last, though Rene might of course stay longer.

Fons:
* discussion about JavaScript vs. native client/server GUI; needs separate brainstorming and discussion session to come up with a prototype

Philippe:
* times stated are relative times in semi-arbitrary scale

Lorenzo:
* VC package (vector with intrinsics) only supported on very limited number of compilers / platforms

Wouter:
* TTreeFormula interface possible

Joerg:
* multiclass + Bayesian NN will arrive by next week
* help for test suite integration into roottest
* need motivation to sync with trunk as often as possible

Gerri:
no comments

Axel:
* C++ precompiled header ready in June? Does progress depend on it?
* library size is crucial, don't define limit before we know what it is

Andrei:
* isn't GDML->CAD sufficient? Some Russians say it's not.
* request (for Matevz) to add time parameter to GL objects
* fluka license nightmare for fluka_vmc questions raison d'etre for vmc

Matevz:
* foresees manpower issue in 1/2 year
* transparency already a member of TColor::fAlpha; should be used in TPadGL etc