Re: [Groff] RE: eqn matrix?

[Tadziu Hoffmann]

> Thirdly, using  an "eqn matrix" is often not the best
> way to set up an array of equations. I know it is the
> "classical" way, as described in the original "eqn"
> manual, but in fact if you place your equations in
> a table you get better control over the layout.

Interesting idea, but it occurs to me that there's
no straightforward way to number such an "equation".
(Incidentally, what is the recommended way to number each
line in an "equation matrix"?  I guess normally one would
use separate EQ/EN pairs and mark/lineup, but that only
works for a single alignment point.)

By the way, changing the definitions to

  define cot %{type "operator" roman "cot"}%

etc. (analogous to the definitions of sin, cos, tan)
gives you the correct spacing without the use of
explicit "^" formatting instructions.