[LIBISIS-code-tickets] [LIBISIS] #64: Algebra for IXTrunfile

LIBISIS Library and Applications noreply at libisis.org
Fri Jan 26 13:52:50 GMT 2007


#64: Algebra for IXTrunfile
--------------------------+-------------------------------------------------
 Reporter:  Toby Perring  |        Owner:  Dickon Champion     
     Type:  enhancement   |       Status:  assigned            
 Priority:  major         |    Milestone:  Version 1.0 (launch)
Component:  Library       |   Resolution:                      
 Keywords:                |  
--------------------------+-------------------------------------------------
Old description:

> Need to extend the unary and binary operations performed on dataset 2d to
> IXTrunfile. Tentative scheme is:
> If runfile = (all other stuff) x (array of dataset2d)  x   i.e. r = a x
> d, then
>
> unary operations:  f(r) = a x f(d)
>
> binary operations: with scalar, vector, d1d, d2d:
>                       f(r,v) = a x f(d,v) ; validity determined by
> validity of f(d,v)
>                    between two dataset 2d:
>                       f(r1,r2) = a1 x f(d1,d2) ; valid if f(d1,d2) valid.
>
> This means that if divide r1/r2, picks up the sample, chopper etc of r1:
> that is what we'd like if normalising by vanadium run. If wanted the
> result to pick up r2, then we'd have to do (1/r2)xr1.
>
> This scheme is subject to TGP checking against the consistency with
> scalar, vector, d1d and d2d operations.

New description:

 Need to extend the unary and binary operations performed on dataset 2d to
 IXTrunfile. Tentative scheme is:
 If runfile = (all other stuff) x (array of dataset2d)  x   i.e. r = a x d,
 then

 unary operations:  f(r) = a x f(d)

 binary operations: with scalar, vector, d1d, d2d:
                       f(r,v) = a x f(d,v) ; validity determined by
 validity of f(d,v)
                    between two dataset 2d:
                       f(r1,r2) = a1 x f(d1,d2) ; valid if f(d1,d2) valid.

 This means that if divide r1/r2, picks up the sample, chopper etc of r1:
 that is what we'd like if normalising by vanadium run. If wanted the
 result to pick up r2, then we'd have to do (1/r2)xr1.

 This scheme is subject to TGP checking against the consistency with
 scalar, vector, d1d and d2d operations. also need to determine rules for
 operations wrt alternative populations of mon_dat and det_data

-- 
Ticket URL: <http://trac.libisis.org/code/ticket/64#comment:5>
LIBISIS Library and Applications <http://www.libisis.org/>
LIBISIS Library and Applications



More information about the LIBISIS-code-tickets mailing list