|
[Rivet] Delayed minutes!Andy Buckley andy.buckley at durham.ac.ukWed Feb 4 15:55:04 GMT 2009
Hi all, Here are some slightly late minutes from last week's phone meeting: please at least check the actions to see if I (or you!) forgot something ;) RIVET GENSER INSTALL Lars is no longer MCnet, so it's best if someone else takes over the responsibility of getting Rivet installed in the Genser collection. What is the status of Rivet in the Atlas kits? ACTIONS: * Jon will install Rivet 1.1.2 on the AFS Genser area for now, and talk to Alberto about getting write permissions in that part of Genser for James &| Emily. EVENT ALIGNMENT CONVENTION James has been re-working the DIS projections, and wanted to know what to do about getting the electron and proton z-directions predictable. We diecided to require conventional input events. Jon happened to mention some HERA knowledge about HERWIG not being able to generate photoproduction/DIS events in the HERA ep direction convention, and hence needing to rotate/flip the event post-production. This fits with Dan Traynor's emails to Andy about Herwig in AGILe needing negative beam momenta to make HERA-convention events: presumably he is used to getting ready-flipped events in H1 MC samples and HZTool analyses. We want to add similar auto-flipping to Rivet: as well as for HERA events, which are not our major analysis target right now, this convention affects Tevatron ppbar events, where the effect will be more subtle, but important. B-factory events and even symmetric e+e- presumably also have conventions that we need to respect, perhaps on an analysis-by-analysis/expt-by-expt basis. The beam ID checker may need to be made ordering-specific to deal with this. ACTIONS: * Andy to mail Dan Traynor (done), plan auto-rotation in Rivet 1.2.0 (ticketed) DESY MEETING / HERA ANALYSES Andy and James were guest stars at the DESY "Pheno Week". The intention was that we would spend most of our time working with students/postdocs and getting new HERA analyses into Rivet / converting the most important HZTool analyses. Instead, we realised that the DESY MC Analysis group's interest was in getting Rivet & Professor to work for Atlas, and to bring HZTool up to date. Accordingly, we gave some talks on Rivet and Professor and have agreed that for now most HERA analyses will probably go into HZTool unless the author decides to use Rivet. While not ideal, this will allow us to use HZTool to make data for validation and tuning more easily than transcribing all the HZTool code by ourselves. ACTIONS: * Andy to make HZTool accept HepMC input (for C++ gens) and to make Cascade write HepMC output. * James to look at converting HZTool output from HBOOK -> ROOT -> AIDA/YODA * Andy to disable current two known-bad HERA analyses in Rivet; Jon to fix them when he gets a chance (hmmm ;). CDF PHOTON CLUSTERING We are pretty sure that our CDF Drell-Yan MC analyses, particularly the Z and W pT spectra, are not clustering the photon emission from the FS electrons properly. Accordingly, the Z will not typically be hard enough in the resulting comparisons to the expt measurement. Needs to be fixed before we do the Atlas MC tunes with Professor. ACTIONS: * Emily to look into a good definition of photon clustering (cone size, wide angle ME-correction photons...) * Hendrik to forward our email exchanges with Peter Skands and Torbjorn Sjostrand to Emily. * Emily/Andy/Hendrik to fix the affected analyses. Definitely a blocker for 1.1.3. METADATA AND DOCUMENTATION Andy mentioned plans to add more metadata to the Rivet analysis classes. This metadata can then be queried from the Python interface and used to build synchronised LaTeX entries for the analyses in the manual and on the Web, as well as the command line: this will be much more maintainable than having to maintain a manual entry, large Doxygen comments, wiki pages *and* coded metadata. The intention is to have analysis descriptions at least as complete as in the HZTool manual. Suggested info was: analysis author(s), long description, whether the analysis is validated/trusted, run conditions (not gen-specific, but info about {energ{y,ies}, process type(s), min pT in plots, whether kinematic efficiency cuts are needed (and suggested values), lifetime-based particle stability}, whether the reference data is detector-corrected... Unfortunately, since we're naturally lazy and abhor writing documentation, adding extra metadata methods will have to block compilation if anyone is actually going to add anything. Andy will deal with this. Andy has also updated the Web and bundled documentation, and will continue to update the manual for 1.1.3. The manual should b complete and ~stable by release 1.2.0. ACTIONS: * Andy to plan new metadata methods, send proposal around for comments. Everyone to reply ;) * Andy to add new virtual methods to Analysis interface, blocking compilation. Provide a good example analysis to be used as a template. Send round an email associating authors to analyses which need metadata. VALIDATOR Frank thinks the validator is pretty much done, except for making the submission backend use Ganga (enables Grid submission). Andy has had some trouble making the analysis configuration structure work well with Pythia and the analyses used by Professor tunings, since typically several analyses group together and most gens need to split their runs into different kinematic regions: he will look into addressing this structurally. ACTIONS: * Frank: get the validator plots on the Rivet website in an initial form. Improve the HTML a required. We really want a couple of gens, but just Sherpa vs. ref data will do fine for now. * Andy: try out a more flexible way of grouping analyses privately before bothering Frank with silly ideas ;) MORE ANALYSES Next focus should be on extending the hadron collider analyses to lower energies. We have reference data for UA5 and similar in the Rivet tarball: these analyses are typically chgd multiplicities and pseudorapidities, and are so very easy to implement. We also want to get some RHIC, SppS, ISR analyses in. ACTIONS: * Hendrik: Will start to implement the UA5 etc. analyses. Hendrik will be touring the eastern US in April to talk with Rick Field and the RHIC people that Peter Skands has been talking to. * Jon: find out what's happening with the RHIC/Alice student that Frank Krauss is trying to get on a MCnet short-term studentship and make sure Hendrik is aware of the status of that before talking to the RHIC guys and BNL/FNAL. * Frank: continue adding more Tevatron analyses from the Sherpa analysis collection. Phew: long meeting, long minutes! Best wishes, Andy -- Dr Andy Buckley Institute for Particle Physics Phenomenology Durham University 0191 3343798 | 0191 3732613 | www.insectnation.org
More information about the Rivet mailing list |