Improved hull walking algorithms for two-dimensional percolation are p
roposed. In these algorithms a walker explores the external perimeter
of percolation clusters. With our modifications very large systems (si
ze L) can be studied with finite and small memory requirement and in c
omputation time tau approximately L7/4. Applications in determining sp
anning probabilities, continuum percolation, and percolation with nonu
niform occupation probability arc pointed out.