The LHE file format has been around since 2006 and has been very successful. At the Les Houches workshop in 2009 an update was suggested, mainly to improve the handling of NLO and matching/merging. The proposal was published in the proceedings (arXiv:1003.1643), but have never really been properly used. Nevertheless, some shortcomings has been noticed and some superflous features have been identified.

In the discussions at this years workshop some ideas on how to better specify weights were suggested, and also some other new features were suggested and old features were suggested to be dropped. Here is a summary of what was discussed.

Here is a list of XML-tags that were introduced in 2009 with comments on what should be with them, and also some suggested new tags.

• <generator> There should be more than one generator tag with a version attribute allowed in the <init> section.
• <xsecinfo> Unchanged
• <cutsinfo> Unchanged
• <procinfo> Unchanged
• <mergeinfo> Should be removed
• <weightinfo> New tag added to define which weights are given the <weights> tag (see below) in each event. The attributes are (also other attributes are allowed, but they do not have a standardized meaning):
  • name Should follow the convention proposed for HepMC
  • muf The factor multiplying the nominal factorization scale for the event for the given weight
  • mur The factor multiplying the nominal renormalization scale for the event for the given weight
  • pdf The LHAPDF code corresponding to the given weight
  • pdf2 The LHAPDF code corresponding to the given weight for the second beam if different from pdf
• <weightgroup> Can be used to group together <weightinfo> tags. The only attribute is
  • name A string which will be combined with the name attribute of the included <weightinfo> tags to give the HepMC weight names.

• <weight> Removed and replaced by <weights>
• <clustering> Removed
• <pdfinfo> Removed
• <weights> contains a list of weights for the events given in the order they are defined by the <weightinfo> tags in the <init> block. If for some obscure reason a weight is absent, it should be anyway be included as a zero weight. Note that the weight for the nominal event is still given by XWGTUP.
• <scales> contains information about different scales used by the MEG in the given event. The scales are given as attributes, and any attribute name is allowed. However, only the following attributes have a pre-defined meaning and if they are missing the value in SCALUP is assumed:
  • muf The factorization scale (in GeV)
  • mur The renormalization scale (in GeV)
  • mups Suggested shower starting scale (in GeV)

In http://home.thep.lu.se/~leif/LHEF there is a small example class which is able to read and write the suggested new format.

There are a couple of things that still need to be discussed:

• Are the weights all absolute, or are the later ones multipliers on the first? Both schemes have been suggested, but we should decide on one and specify it in the definition, rather than letting it be secretly generator-dependent.
  • Leif: I have always thought of this as being the absolute weights. I suggest we go for that.
• Should this suggestion be version 2.1 or 3.0?
• Can the tags suggested to be removed above, simply be removed or do they need to be left for backward compatibility? Since they probably have never really been used, it should be possible to just throw them away.