groff / eqn / ps2pdf / square root problem

[Greg Shenaut]

I really don't know where to direct this problem. I have a manuscript =20=

that I format with groff -e -ms that has an inline equation using the =20=

eqn "sqrt" operator. When I format the manuscript to postscript, and =20
look at it (with gv) or print it on a postscript printer, the square-=20
root radical comes out perfectly. However, when I use ps2pdf to =20
"distill" the postscript to pdf, the equation is a complete mess: it =20
starts OK, prints the =C3 ok, but then backspaces about an inch, and =20
prints the rest semi-randomly on top of the first part. When I use =20
Adobe or pstopdf, I get slightly different errors, but neither is =20
correct. It appears to be an interaction between how groff generates =20
the postscript code and how the pdf distillers interpret it.

I am using OSX 10.4.10, groff 1.19.2_1, and ghostscript 8.54_0.

For the moment, I've used a little algebra and rewritten the relevant =20=

section of my manuscript in such a way as to avoid the square root, =20
but it would be really great if someone could take a look at this for =20=

future purposes.

Cheers,

Greg Shenaut=