This is a list of knwon bugs in the DYqT numerical code v1.0
(some of them do not affect the numerical results, the others can affect the numerical results at permille level).

Thanks to S.Camarda (CERN) for reporting bugs from 1 to 6.

1) In the file functions.f, in Cqg3(sh,th,uh,q2,what) at line 435:
     /     (1-uh/th)/(4*la**2)+(uh**2-th**2)*((q2-s2**2)/(4*sh)+
should be instead:
     /     (1-uh/th)/(4*la**2)+(uh**2-th**2)*((q2-s2)/(4*sh)+

(notice s2**2 -> s2)

The same bug is also present in Cgq3(sh,th,uh,q2,what) at line 521:
     /     (1-th/uh)/(4*la**2)+(th**2-uh**2)*((q2-s2**2)/(4*sh)+
should be instead:     
     /     (1-th/uh)/(4*la**2)+(th**2-uh**2)*((q2-s2)/(4*sh)+

See Eq. A12 of http://journals.aps.org/prd/pdf/10.1103/PhysRevD.40.2245

2) In the file functions.f in Ddd(sh,th,uh,q2), at line 1177:
      Ddd=temp*xlumqqbdcc
should be instead:      
      Ddd=temp*xlumqqbddd
where xlumqqbddd has to be defined in flavour.f and it is obtained from xlumqqbdcc
starting at line 313 of flavour.f by changing fh1 <-> fh2

see Formula 2.18 of http://journals.aps.org/prd/pdf/10.1103/PhysRevD.40.2245

3) In the file flavour.f, lines 23-27 the sumckm are attributed incorrectly
They should exchange u <-> d, c <-> s, and t <-> b
The lines:
      sumckm(1)=ckm(1,2)**2+ckm(4,2)**2
      sumckm(2)=ckm(1,2)**2+ckm(1,3)**2+ckm(1,5)**2
      sumckm(3)=ckm(4,2)**2+ckm(4,3)**2+ckm(4,5)**2
      sumckm(4)=ckm(1,3)**2+ckm(4,3)**2
      sumckm(5)=ckm(6,2)**2+ckm(6,3)**2+ckm(6,5)**2
should be instead:
      sumckm(2)=ckm(1,2)**2+ckm(4,2)**2              !sum xd -> apply to d (Vud^2 + Vcd^2)        
      sumckm(1)=ckm(1,2)**2+ckm(1,3)**2+ckm(1,5)**2  !sum ux -> apply to u (Vud^2 + Vus^2 + Vub^2)
      sumckm(4)=ckm(4,2)**2+ckm(4,3)**2+ckm(4,5)**2  !sum cx -> apply to c (Vcd^2 + Vcs^2 + Vcb^2)
      sumckm(3)=ckm(1,3)**2+ckm(4,3)**2		     !sum xs -> apply to s (Vus^2 + Vcs^2)	     
      sumckm(5)=ckm(5,1)**2+ckm(5,4)**2		     !sum xb -> apply to b (Vub^2 + Vcb^2)	     

This affects xlumqg,xlumgq,xlumqqbdbc,xlumqqbdcc,xlumqqbddd,xlumqqead in W production
This bug makes a difference only when switching off some of the ckm elements, i.e. for a non-unitary CKM matrix.

5) The files fodyqt.f and functions.f the partonic Mandelstam invariants th and uh should be interchanged: th <-> uh
In the file fodyqt.f changes are requires at lines 312, 313:
      uh=q2-ss*x2*tm*dexp(yv)
      th=q2-ss*x1*tm*dexp(-yv)
lines 437, 438:
      uh=-tm**2/z1*(1-z1)*(1+lb)
      th=q2-tm**2/z2*(1+lb)      
lines 541,542:
      uh=-tm**2/z1*(1-z1)*(1+lb)
      th=q2-tm**2*(1+lb) 

and in the file functions.f at line 1678:
      a=-uh

This bug does not affect the numerical predictions. For y-integrated (or y-symmetric) cross sections, as used in DYqT, the result is not affected as y -> -y is equivalent to uh <-> th.

6) In the file evnew.f at line 1191:
       QQBN1 = QQBN1 +  FX1(si)*FX2(-sj)*sigmaij(si,sj)*Hqq        
should be instead:
       QQBN1 = QQBN1 + (FX1(si)*FX2(-sj)+FX1(-si)*FX2(sj))*sigmaij(si,sj)*Hqq

7)

In dyqt.f line 372

replace
         sigw=q2*brw*ww*q2/(pi*Mw)/((q2-Mw**2)**2+Mw**2*zw**2)

with

         sigw=q2*brw*ww*q2/(pi*Mw)/((q2-Mw**2)**2+Mw**2*ww**2)

(notice zw->ww)