Re: Re: A proposal:xsl:result-document asynchronous at

People are easy: sort alpha by forename then surname. That's usually pretty random, then pick every nth entry.

You're joking! Ordering by name will give you a list that is strongly
clustered by age and ethnic origin, and therefore a high chance that
certain groups will be completely absent from the sample.

I think we're talking about two different things here: a
representative sample as distinct from a random sample.
I'm also assuming that N and n have been sized suitably.