Hi Rooters, I agree with Pasha's observations about CLHEP. Although Stefan Kluth's ROOT port of CLHEP is important for everybody wanting to use code based on "legacy" -:) CLHEP classes (G4), I think that not to slow down our work we (i.e. ROOT community) need to have full control over a set of well integrated physics vectors (3 vectors, Lorenz, etc). Recently we've been offered by Chris Jillings (chris@owl.phy.queensu.ca) a ROOTified 3-Vector class. The class seems to be quite complete and Chris is now finishing the full integration in ROOT and he expects to have things ready in a week or two. However, maybe Chris could already discuss some features of his class on roottalk so people can comment on them. Cheers, Fons. Pasha Murat wrote: > > > Hi Rooters, > > I guess this could be an intresting issue for public discussion: > > Yves Schutz writes: > > > > Does a class "Lorentz vector" also exist ? > > > > To the moment ROOT doesn't include classes for physics-oriented > (2D-, 3D- and Lorentz) vectors. I'd consider it as an important piece > which is still missing. One could think about several possibilities of > how to fill this "hole": > > - recently Stefan Kluth reported on his success in generating rootcint > dictionaries for CLHEP classes. (Great job, Stefan!) > In principle, one could take CLHEP classes as a starting point. The major > problem with CLHEP as I see it is that currently it doesn't seem to be > supported. > Moreover, physics vectors seem to be the least developed part of CLHEP, the > classes are missing quite a few very useful methods and it is not clear > what to do if we need these methods to be added: > - to ask somebody (whom?) to add missing methods (like pseudorapidity of > a 3-vector or delta(phi) between 2 2(3)-vectors) to CLHEP? > - to branch ROOT flavor of CLHEP ? > And definitely one'd like physics vectors to "know" about ROOT's TMatrix > class, whereas CLHEP vectors do not deal with ROOT TMatrices. So it seems > that even if we start from CLHEP, ROOT "flavor" of CLHEP will immediately > evolve into an independent package and it is necessary to understand all the > consequences of this step. > > - an alternative approach could be to start from another, different from CLHEP, > physics vectors package (preferably - non-templated) and iterate > on it to make it useful for everybody. This seems to be a relatively modest > job - for example, I'm using a simple package of myself which is pretty much > complete with the exception of several methods like Lorentz rotations / boosts > which is trivial to add. I implemented it for the only reason that CLHEP > classes didn't have methods I needed ... > I'd also expect that many other people have similar packages - it is much > more easy to implement them than to wait for somebody to do it... > > Anyway, as we already have a community of people actively using ROOT, I believe > that > > a) it is clear that we need physics vectors classes to be added to ROOT and > b) that it makes a lot of sense to discuss what exactly we want from these > classes before going into implementation - here we facing a situation where > it is more difficult to decide on what is the right thing to do than to > implement it. > Any opinions? - Pasha -- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland Phone: +41 22 7679248 E-Mail: Fons.Rademakers@cern.ch Fax: +41 22 7677910
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:34:35 MET