Following a discussion during yesterday's team call, I emailed the developers to see if anybody could shed any light on the relative merits of the two solvers provided in Calc 6.2.
For information, I am forwarding the response from Kohei Yoshida which contains a useful link to a paper comparing various open source linear program solvers (including the two that we provide, lp_solve and CoinMP).
Regards,
Steve Fanning