Dieter Wilhelm <die...@duenenhof-wilhelm.de> wrote: > Eric S Fraga <e.fr...@ucl.ac.uk> writes: > > > Bastien <b...@altern.org> writes: > > > >> Hi Xue, Eric and Dieter, > >> > >> die...@duenenhof-wilhelm.de (H. Dieter Wilhelm) writes: > >> > >>> (I would avoid the ambiguous expression "column two" since it is a > >>> relative specification) alternatively > >>> > >>> The TWO REFERENCES expand to a field range from the row above the > >>> current row, starting with two columns to the left up to the current > >>> column. > >> > >> Yes... but this is a bit long. > >> > >> I finally used this: > >> > >> @@-1$-2..@@-1 @r{in the first row up, 3 fields from 2 columns on the > >> left} > > > > Concise and correct! I'm happy with this. > > Sorry but I don't understand "in the first row up". Maybe better: The > (or a) row up, 3... > > Another grievance with such a terse description for me is although it > may describe the end result - the range - correctly but does not take > into account how the references at hand are working. > > But maybe I'm just picking nits here :-) >
No, I think it's unclear as well (I hadn't paid attention to the thread previously. Sorry for joining the party late). > What about such an approach: > > @@-1$-2..@@-1 @r{a range of 3 fields: a row up, from 2 fields on the left .. > a row up} > Perhaps factoring out the row part makes it clearer? Also, presenting it as a movement from the current cell might help - at least that's how I tend to read these specs: "a range of 3 fields: up one row, two columns over to the left .. the current column (implicitly specified)" Nick