Hard processes in massless perturbative QCD (to be updated)
All coefficient-function formulae linked below are
given in the MS scheme.
- The Fortran routines xclns3p.f and
    xclsg3p.f for the parametrized third-order 
    coefficient functions for the longitudinal structure function, as 
    presented and briefly discussed in  
    hep-ph/0411112 (Phys. Lett. B606 (2005) 123-129).
    
 
- Their exact counterparts xclns3e.f, (non-singlet)
    xclps3e.f (pure-singlet) and 
    xclgl3e.f (gluon) have been published, with the results for 
    F2, in  hep-ph/0504242 (Nucl. Phys. 
    B724 (2005) 3-182 [!]) -- here is a version 
    without the appendices 
 
- The corresponding routines for F2 are 
    xc2ns3p.f, xc2sg3p.f, 
    xc2ns3e.f, xc2ps3e.f 
    and xc2gl3e.f
 
- (Unoptimised) N-space subroutines for the parametrized non-singlet and 
    singlet expressions are provided by c2lmom.f
 
- Finally a FORM file is available with the complete 
    integer-N and x-space results for the above coefficient functions
 
The exact x-space numerics requires the weight-5 extension of the 
Gehrmann/Remiddi package for the harmonic polylogarithms (HPLs) published in 
hep-ph/0107173 = Comput. Phys. Commun. 141 (2001) 296. 
NOTE: The sign of the ζ5 term in the first line of 
Eq. (4.19) is mistyped in the archived version of the second paper. 
The third-order coefficient function for the odd-moment F3
- The Fortran routine xc3ns3p.f for the parametrized 
    third-order coefficient function for neutrino-antineutrino sum 
    F3 as given in  hep-ph/0608307
    (Nucl. Phys. B (Proc. Suppl.) 160 (2006) 44-50)
    
 
These parametrizations have been given in  
hep-ph/9907472 (Nucl. Phys. B568 (2000) 263-286) and
 hep-ph/0006154 (Nucl. Phys. B588 (2000) 345-373). 
While being accurate enough for normal applications, they are 
considerably faster (especially for F2 and F3), than 
the exact expressions of Sanchez Guillen et al. and Zijlstra / van Neerven.
NOTE: misprints in Eqs. (A.3) and (A.4) of hep-ph/0006154 have been corrected
in the ps-file linked above.
  - The Fortran routines xc2ph2p.f and
      xclph2p.f for the photonic coefficient functions
      for F2 and FL published in 
      hep-ph/0110331 (Nucl. Phys. B621 (2002) 413-458).
  
 
  - The Fortran routine xk3cnvp.f for the
      parametrized convolutions contributing at NNLO for
      F2,NS and
      F3 
  .
 
  - The corresponding convolutions entering at N3LO are given
      in xk24cnvp.f for F
      2,NS and in xk34cnvp.f for 
      F3,NS
  
 
These results have been presented in  hep-ph/0103123 
(Nucl. Phys. B603 (2001) 42-68). The approximations for the third-order
coefficient functions in that paper have been superseded by the above exact
results.
  - The Fortran routines xcttns2e.f for the
      transverse, xcltns2e.f for the longitudinal
      and xcatns2e.f for the asymmetric coefficient
      functions at the second order in aS, recalculated in
       hep-ph/0604053 (Phys. Lett. B638 (2006) 61-67)
  
 - The FORM file including (besides the 
      splitting-function differences) these and the lower-order results
  
 
These results confirm the calculations of Rijken and van Neerven in 1996/97
(there are some misprints in their papers). The numerical routines require
the HPL package by Gehrmann and Remiddi mendtioned above.