[math-font-discuss] Ugly drawn rules in Adobe Reader

[Karl Berry]

Hi Taco,

    Specification would be like \mathchardef, say:

\rulechar"0130 or \rulechardef\foo="0130, then, maybe?
And maybe it should be \mathrulechar?

Is there a character in an existing font that can be used, or does a new
font have to be created?

And how is it going to be used?  Sorry, I'm still not clear on this.
\radical and such are primitives, their functionality cannot be
duplicated in macros, can it?

    See   http://tex.aanhet.net/temp/mathtest.pdf
    Especially look at the  top left of the sqrt signs. 

Meanwhile, in the alternative ...  I do not get such bad results.  For
me, the biggest sqrt is not aligned correctly with xpdf 3.00.  With raw
gs 8.15, it is fine.  With gv 3.6.1, the beginning of the horizontal
rule is too thick but alignment is ok.  With Acrobat 5.0 (came with RHWS
4), it is fine.

I'm puzzled as to why my results are so different from yours.  Nothing I
see seems nearly bad enough to be worth the enormous and permanent
hassle of defining a new primitive.

BTW, can you send me the TeX source (I know it's trivial, but ...).

    and that makes it hard to align the two objects correctly.

I have seen the misaligned sqrt's before, but I guess I always assumed
that something in pdftex was outputting the horizontal part in the wrong
place, and/or xpdf was rendering wrong.  From the results above, I would
have thought it's an xpdf problem, but obviously you are seeing
something different.

    Adobe will not listen to bug reports if you are not a multi-million
    client

Well, Wendy, Ross, Hans, and Nelson, made a special trip to Adobe last
year to tell them about various bugs, and they seemed at least somewhat
receptive.  And I believe they finally fixed the bitmap rendering
problems because of the TeX world's complaints.

Cheers,
karl