[metapost] The strangest issues?

[Giuseppe Bilotta]

Friday, March 18, 2005 Daniel H. Luecking wrote:

> Since the two equations to be solved for t and \tau are both cubic
> in one variable and quintic in the other, if they could be solved
> for t and \tau in terms of the lambdas they would be multiple-valued.
> There would therefore be the possibility (even likelyhood) of "branch
> points" where these different values come together and differential
> analysis is impossible.

Well, I have personally not met such situations in my
experiment, and I suspect that such situation cannot arise
(except maybe for cupsidal curves).

> And here we have rapidly converged to the limits of my knowledge of
> these things.

That's ok. We even went way off topic with the discussion, I
suspect :)

-- 
Giuseppe "Oblomov" Bilotta