substring match vs. insignificant space removal

[Donn Cave <donn@u.washington.edu>]

Working on a white pages search application, I ran into
a problem with substring searches:  if you search, for
example, for 'cn=jo* b* lo*', you would (naively) be
surprised to get, for example, 'Joe Blow' or 'Joby Lowe'
as a result - the spaces before 'b' and 'l' were supposed
to filter for separate words starting with those characters.

I'm guessing this goes back to the U Mich implementation.
It isn't inconsistent with the RFCs I could find, though even
recent working drafts don't specifically raise any question
as to whether insignificant space removal should apply in
exactly the same way to parts of a substring match, as to
a regular string match.

Is there any way to work around this?  Quoting the internal
spaces doesn't seem to help.

Thanks,
	Donn Cave, donn@u.washington.edu