It seems like there are a bunch of people trying to generate/simulate their own MC from scratch and running into a lot of problems.  This isn’t that surprising as it’s complicated and it’s really meant to be “expert only”.  This is why we have a production manager who’s job it is to produce MC (and data, when we get it).   

There are a few good reasons for someone to generate their own events (see below), most MC should be produced and stored locally at JLAB by the PM.  If the benefits of this aren’t obvious, I’ll spell them out:

• events are available to everyone
• (via the PM) the correct versions of each programs are used (e.g. SLIC, geometry, etc)
• events are correctly documented, from MadGraph run & param files, to version numbers & dates in the files
• the user doesn’t have to figure out all the steps it takes to generate this stuff

We have a bunch of trident MC (with & w/o truth) on tape at JLAB right now (recon lcio & DST) and will soon start with the “A’ signal” at various masses.  If you want a specific sample generated, just send an email to the PM (Sho, for now) and me (Matt) with this information: 

• sample you want generated
  • something standard, like “35 MeV A’s with 2.2 GeV beam energy” that’s easy
  • if you have a MadGraph file (.lhe) of events, that's ok too but you should also send the code that generated them so it’s documented
    • alternatively, we can make the MadGraph events for you 
  • SLIC macro that generates the events is fine to (e.g. particle gun of pi0’s or something) 
• number of events you want
• specify whether you want them with or without pileup?
• detector version (“latest" is ok)

We may have some kindly suggestions based on your request but we’ll try to accommodate.  

Thanks, Matt

****   Reasons you may want to generate your own MC include:  playing around with detector geometries (e.g. alignment studies); testing out new sources of events (e.g. cosmic ray generator) prior to it being submitted for production, etc.