OK. I can’t write rules I know but I’d like to learn. Could the above
rule (extract) be improved by using tagsets?
It could be written in a more compact way, but whether it’s really an
improvement needs to be tested, as this could lead to false alarms. Testing
can be done by the rule creator[1] in expert mode, and by locally checking
against Wikipedia, as described on our Wiki.
Also, is it an ‘error’ to include ‘a’ but not ‘an’ in the regular
expression above?