Attendees - IG, JK, EB, PC, PM, RK, FC

We are beginning discussions that will become the final documents on all of the products. Irwin drafted summaries of the highlights of the ROOT and Histoscope discussions. Today's meeting was based on those summaries.

ROOT

Pasha brought up the fact that the interface to cint is abstract, and could be replaced by some different tool hooked to the interface

Rob says that the main point is that there is no perfect thing, that whatever we implement will involve a non-zero effort. ROOT is basically a hand-wired compiler, implemented as a set of functional units, with non-standard development techniques.

The discussion turned to the GUI. Pasha stated that Gordon's criticisms of the product were based on version 1.0, and that - v 2.X has nice vastly improved GUI. Rob said that the product is just interpreted cint, so that the GUI is just cint scripts.

As for the product's modularity, Rob stated that the product is modular, in the sense that the functional units are distinct, but that the modules are interdependent. Thus changing one module will probably effect others, with effects that are hard to predict. Furthermore, any external packages hooked up to the product must conform to the ROOT module. Pasha said that while this is true, those interfaces are well-defined.

Recent upgrades to the product are mostly additions. Releases have been occurring about once per month, and fixes to identified bugs are included

The question of a support model was raised. Irwin said that support structure will be collaboration CD/CDF/D0/Rene, and our recommendation will be spelled out in the document.

Nirvana

Nirvana possesses plug and play capability between distinct components, unlike ROOT, in which the packages are tightly coupled. Also, Python is better than cint, and adopting Python would allow us to leverage its user base.

The probable long time to bring Nirvana up to speed is an issue. The estimates of 6-12 months minimum are probably optimistic and longer than we can live with. Eileen raised a key question - What % of the required functionality is available now? That is, if the physicist sits down with the product right now, how much of his work can he get done?

This led us to one final point, that so far the testimony for ROOT has been from physicists, while that for Nirvana and for MATLAB has been from non-physicists. Pasha asked for a demo of MATLAB from Jeff.

Another meeting Friday after lunch.

Upcoming events:

Thursday, 9:00, WH11NW - IDL presentation

Friday 1:00, FCC2A, Meeting