Since distmesh has no "good" convergence criteria (although differencing some median mesh quality measurement has worked for me in the past), wouldn't it make more sense to let the program complete a set of fixed iterations and then return the connectivity and points instead of asserting and haulting the program?
Generally, 50 to 100 works then I typically do some standard mesh improvement stuff.
Hey again,
Since distmesh has no "good" convergence criteria (although differencing some median mesh quality measurement has worked for me in the past), wouldn't it make more sense to let the program complete a set of fixed iterations and then return the connectivity and points instead of asserting and haulting the program?
Generally, 50 to 100 works then I typically do some standard mesh improvement stuff.
just a suggestion,