relative roots---one other thing

[Conor McDonough Quinn]

Dia dhoibh ari/st, a chairde!

One other thing.  We're discussing this issue, it seems, because it leads
up to making a policy decision about how to word glosses for the whole
dictionary.  And we don't want to have to go over the whole dictionary
again, should this approach fail for the practical users.

But the relative roots are, for the most part, easily segregated in their
contribution to the glosses, which means, with they exception of some
cases that might get stretched out into discontinuous forms (and these
could always be flagged as they come up), all we really need to do is give
the relative root portions of the glosses some kind of unique identifier
pattern, such that we can easily globally convert them (with the few
flagged ones searched for as such and dealt with by hand) from one form to
another.  So even if "Xmanner" doesn't work out as a final presentation
form, it might be a decent core database form, that can be output in any
number of presentational glosses.  Also makes it easier for us
Algonquianists to search each other's dictionaries for all the forms with
relative roots, just straight from the glosses (i.e. without the
dictionary-makers having to redundantly tag the entries as such, though
that could be done, and done by machine!).

Till later, keep safe and sane.

Sla/n,
bhur gcara