On Wed, 30 Jul 2014 13:40:55 +0100 (BST)
(Ted Harding) <[email protected]> wrote:
> On 30-Jul-2014 09:23:54 Anton Shterenlikht wrote:
> > To me
> > .char \[-+] \f[S]\v'.05v'\z+\v'-.3v'\-\v'.3v'\v'-.05v'\f[]
> >
> > looks better vertically aligned, but it's just a bikeshed.
> >
> > Many thanks for all replies.
> > Anton
>
> Well, now that I can finally get round to it, Denis Wilson:
> .char \[-+] \f[S]\v'.1v'\z+\v'-.25v'\-\v'.25v'\v'-.1v'\f[]
> Mike Bianchi:
> .char \[-+] \f[S]\z+\v'-.35v'\-\v'.35v'\f[]
>
> have already come up with suggestions similar to what I was
> considering (as well as Anton's above).
>
> The main difference, which I strongly recommend, is to use
> 'm' rather than 'v' as the scale for the vertical motion.
> The reason is that 1v is one line-space, which can be set
> independently of the current point-size, while 'm' is, in
> effect, the point size, so that the result will scale as
> the point-size changes.
>
> What I had devised (and again some tweaking of the motions
> may be desirable -- it might look better to shift it down
> a bit) is:
>
> .char \[-+] \Z'\fS+\fP'\v'-0.370m'\fS\[mi]\f[]\v'0.370m'
>
> Then, in plain text,
>
> \[-+] \[+-]
>
> will display appropriately. And of course one can also define
> "-+" for the 'eqn' environment (where it will feel most at home
> anyway):
>
> .EQ
> define -+ "\[-+]"
> cos ( A +- B ) ~~=~~
> cos ( A ) cos ( B ) ~ -+ ~ sin ( A ) sin ( B )
> .EN
>
> and, to show how the scaling adapts:
>
> .EQ
> a ~+-~ b ~-+~ c ~~~~~ e sup{a~ +-~ b ~-+~ c}
> .EN
>
> This little thing has provoked quite some interest!
>
Indeed Ted! But I would make some small adjustments (a) make -+ the
same size as +- (b) lower it as you suggest; to my eye (both next to
+- and in an equation context) the following is nearly what we want:
.char \[-+] \v'.07m'\Z'\fS+\f[]'\v'-0.368m'\fS\[mi]\f[]\v'0.368m'\v'-.07m'
Following eqnrc I defined it as follows
sdefine -+ %{ type "binary" \[-+] }%
Denis
--
