Feature request: singularLocus efficiency

Nobody has claimed this yet.

The method `singularLocus` in M2 is not very efficient. I recently looked at the file [`bir.m2`](http://goo.gl/eT4rCR) written by Giovanni Staglianò for his [article](http://dx.doi.org/10.1016/j.jsc.2015.11.004), where one gets the option to use `Singular` to carry out this computation. For example, here is a test case.
```macaulay2
R = ZZ/11[x_0..x_3];
I = ideal apply(10, i->random_3 R);
elapsedTime isSmooth(I, Use=>Singular); -- 0.0754029 seconds elapsed
elapsedTime isSmooth(I, Use=>M2); -- 3.26353 seconds elapsed
```
Is it possible to make this an official feature, which will make `singularLocus` much more usable?
**Update**: The above example is a bit degenerate. Here is a better one.
```macaulay2
I = Grassmannian(1,4,CoefficientRing=>ZZ/101);
R = ring I;
J = I + ideal(random_1 R, random_2 R);
elapsedTime isSmooth(J, Use=>Singular); -- 2.72101 seconds elapsed
elapsedTime isSmooth(J, Use=>M2); -- 18.1278 seconds elapsed
```

Start with M2's singularLocus implementation and the referenced bir.m2 implementation, then trace how isSmooth handles Use=>M2 and Use=>Singular. Done means Singular-backed computation is available as an official option for singularLocus and the two supplied examples produce the expected behavior and timing comparison.

Written by the indexing model from the issue text.