Menu
▾
▴
#124 Collateral: Groups cannot be finished
After more testing, I decided that the topic I submitted
to the developer's forum ("Groups cannot be finished"
mafaehnd 0 2003-12-20 14:01) is a bug, since it keeps
me from finishing the whole entry.
Here is the problem:
I have the following data in Annotate (threshold = 50%):
a ba tuđan
a tur toz..
b ya kuoppas
c ya biːl
c ya buor
d ba saŋ
When I go to Tabulate, I am asked to chose one of the
different data for group "c". Once I do that, I am stuck.
I cannot tabulate group "a".
The only way to work around it is to go back to
Tabulate and assign different group tags to the two
data for Yakut (ya). However, if I think that these two
are cognates (doublet), then that would not really
reflect my analysis of the data.
Logged In: YES
user_id=618259
[Maria submitted this to Developers Forum]
I am not sure whether this is a bug or a feature request, or
something else. Here is what happens:
I have one entry with three varieties. One variety has two
data, but noth of these data have the same group tag. The
data in the other varieties have different group tags:
a ba ata/
a ba ataj
b ya aga
c tur baba
When I go to Tabulate, I am asked to choose one of the
multiple datums defined for group "a". Once I do that, nothing
further happens. The entry cannot be finished and stays in
Tabulate.
I have a threshold of 50%.
[Joe] Part of the problem is visual: the dialog box that asks
you to choose hides the other data against which you have
to make the choice. Pragmatically, you want to choose the
one that is most likely a true cognate with the others (the
ones you can't see), and include a remark saying that there's
probably some kind of borrowing analogy going on in the other
one, especially if you have an idea what it's an analogy with.
Romance has a bunch of doublets of this kind, but one of
them is usually a learned borrowing from Latin, not a product
of regular sounds change (except within Latin - it's usually
borrowed from late Medieval Latin, not from Julius Caesar.
-----
Let's see if moving the dialog box to a neutral corner helps.
Logged In: YES
user_id=630520
I have attached a demonstration collection for this bug. To
see what the problem is, do the following:
> go to View "Turkic"
> go to Annotate and select the entry "dust"
> go to Tabulate in entry "dust" (you may have to click
Refresh if all the group tags are "?")
> you will be asked to "select one of multiple datums for
group tag c and variety ya.
> select one, it doesn't matter which one.
> you will now be stuck on this entry, unable to tabulate
group a.
Logged In: YES
user_id=630520
Another problem associated with this bug: sometimes the
group which contains a double entry itself cannot be
tabulated after one of the double entries has been chose. For
example:
Tag Abbr Aligned
a ba bar
a ya bar
b ba kajt
b ba k/it
b tur g/it
c ba kil
When I go to Tabulate, I am prompted to choose one of the
datums defined for group b and Variety ba. This makes sense,
since there are two entries for group b and variety ba.
However, once I have chosen one of the datums, I get "No
Groups Available", although there are two datums in this
group: the one I have chose for ba, and the one for the
Variety tur. I cannot tabulate group a either.
Logged In: YES
user_id=630520
There is another problem with this, involving grapheme
cluster: when the user chooses one datum of a variety with
more than one datum with the same group tag, and that
datum contains a grapheme cluster, Tabulate displays only
the correspondence set containing that grapheme cluster.
To duplicate this bug:
> import the attached collection (bug 863745 graph cluster)
> go to Annotate
> go to Tabulate
> when prompted to choose one datum, choose the one
containing the grapheme cluster