PS Yes, I am aware of Bill Moseley's Text-Aspell-0.03 on CPAN. However, I am

• looking for a pure perl layer over the aspell binary, and
• looking to learn about pipes.

Hence reinventing the wheel.

PPS In the Text-Aspell docs, it says that implementation "should provide good performance compared to forking the aspell program for every word".

In the implementation I posted above, I keep open one aspell process, yes, so I would not suffer the poor performance Bill mentions? I don't really care about speed here, I am just curious. rkg