[RsR] Package title

[Martin Maechler]

>>>>> "FrL" == Friedrich Leisch <Friedrich.Leisch using tuwien.ac.at>
>>>>>     on Fri, 9 Dec 2005 09:13:57 +0100 writes:

>>>>> On Thu, 08 Dec 2005 11:26:11 +0100,
>>>>> Peter Filzmoser (PF) wrote:


    >> How about "robust-base"? I rather prefer longer abbreviations
    >> than shorter if confusion can be avoided. The advantage is that
    >> we can then also have packages "robust-mva", "robust-glm" etc.,
    >> and it is easy to find all "robust" packages since they are
    >> in alphabetical order.

    FrL> The dash is not a legal character for package names, because, e.g.,

yes, thank you Fritz;
and thanks to the other proposals and thoughts.

     .....

    FrL> We could do robust.base, roubust.mva, ... though.

As long as packages cannot export their C API, and also for
other reasons, I don't think I'd aim for more than one
"basic robustness" package.
One of the major points has been that the package becomes so
important (and "well done") that it will become a 'Recommended'
one.
For that very reason, it should not try to do "everything"
(robust) but the *important* things, maybe in the sense that we
("workgroup regression" and others, at Treviso) had outlined
earlier.

Other people would build specialized packages that have 
'Depends: BasicRobustStatistics' in their DESCRIPTION setup,

where, given the comments so far (and my taste..)
I'd still vote for
    'robusta'  (from Valentin)
or  'robustats'

I'd definitely like something easy to spell (correctly) and to
pronounce.  The 2nd proposal (above) is already at the border of
being easily understood in oral communications.

Martin