Non-destructive graph unification algorithm

[Fabien]

Dear comp.theorists,

I need to implement a - possibly efficient - feature-structure
unification algorithm in a pure functional settings.

The paper by David Wroblewski 'Non-destructive graph unification'
seems to be tailored-made for my problem, but unfortunately I am
unable to get my hands on an electronic of it.

Does anyone of you have some handy references or pointers relevant to
the topic ?

Thanks a lot, Fabien