Re: [Groff] eqn: sqrt problems

[Deri James]

On Friday 02 Sep 2011 12:22:46 Denis M. Wilson wrote:
> Hello all,
> 
> There have been various postings on this list about the wrong
> positioning of the radical extender. I found that it depended on which
> font family was being used.  It turned out that some text faces have a
> square root sign which was therefore being used in place of that from
> the symbol font, but the extender was still from symbol.
> 
> On removing them from the text fonts, the problem was cured.
> 
> So (a) when using a different family use a matching symbol face, or (b)
> maths symbols should be omitted from text faces, or (c) eqn should
> ensure that maths symbols are from the symbol fonts, or ...
> 
> What do people think?
> 
> Denis

Denis,

If the chosen font (which contains a radical) also contains a matching 
radicalex, the problem should not exist, since groff would choose the pair from 
the same font. The problem exists if only one of the pair of glyphs is 
available in the chosen font, the missing glyph will then be found in the 
Symbol font.

The example below is using a font called U-TI which contains a radical glyph 
but no radicalex. The .char request specifies to groff to use the radical glyph 
from the Symbol font always .

.char \[sqrt] \f[S]\[sqrt]\fP
.EQ
gfont U-TI
1 over sqrt {ax sup 2 +bx+c}
.EN

Someone else may have a more elegant solution.

Cheers 

Deri