>     Why do you use (obviously flawed) research methods?
    > 
    > My method is to ask other people to do it for me.  I use that method
    > because it is efficient.  Its results are accurate, too.
    > 
    > However, when a person tells me his OS is free, I have not always
    > checked.  Sometimes I just took his word for it.  The problems that
    > have been reported here in various free systems (and, mostly,
    > corrected) show I need to discuss the criteria more carefully with
    > them.

    You contradict yourself.  You say it's efficient and accurate and then
    point out its inefficiency inaccuracy.  I find it stunning that you can
    reconcile this.

There is nothing to reconcile -- you have combined two statements
about two different things, so the resulting contradiction didn't come
from me.

When I want research, I ask people to do it.  That is efficient, and
we have not seen any errors in it.

In the case of AROS, it's possible I did not ask anyone to do
research.  I might have just taken the developers' word that the
system is free.  It was years ago and I do not know what happened.

However, most of these problems had nothing to do with quality of
research, because they did not arise until after I had decided to
endorse a program.  Research can only check the present, not the
future.  For instance, the reference to unrar on BLAG's site was in a
wiki; it was posted by a user in the recent past.  (It is possible
that this happened with AROS too.)  Likewise for the GNU/Darwin
problem.  I think this occurred in several others too.

My conclusion is that I should do more detailed discussions with the
developers of the FSF-endorsed systems about these specific possible
problems and how to avoid them.