Kalusa: Replies to several questions raised
|From:||Gary Shannon <fiziwig@...>|
|Date:||Thursday, May 25, 2006, 1:06|
> I voted twice positively for "Tapu kisa da zhura. =This table is
> blue." That boosted it from 100 to 150 to 167. So ifthere's a
> sentence I really love or really hate, I cansingle-handedly give it a
> very high or very low score. I'm just sayin'.
> Larry Sulky
I've been wrestling with this problem. It's easy
enough to limit people to one vote per person when
it's just some single thing like a poll question, but
when hundreds, and potentially thousands of sentences
are involed it becomes more difficult to work out a
good way to keep track of which sentences a person has
or has not voted upon. I'm still trying to solve it
though. Any solution would require that people login
to the site to establish their "identity", although
the login process could be made anonymous and secure.
> Wow, as I've typed this up, the corpus is now up to301 sentences!
> Wild... Anyway, not to bug, Gary, but didn't yousay you were
> going to link to the files you used to create this?Seeing how
> many sentences I've added to Kalusa, I'd love to beable to do
> the same for my own languages to build up a corpus!
I was going to upload the source to the site, but then
I thought maybe I should fix a couple of bugs first,
and that lead to a few more changes, etc. etc. I'll
upload it this evening and put a link on the main
page. But be warned, I'll be adding new features, so
newer versions of the software will be available
> All the "vito" sentences with a ranking of >=100
> have "SUBJ vito es OBJ". "vito" seems to want itsobject
> marked with "es".
> All the handful of "a" sentences currently have CQrankings
> below 50.
> Jim Henry
In the beginning, "es" was required under all
circumstances to mark the direct object clause. Most
sentences use the direct object marker "es" at all
times, but some others seem to imply that the direct
object marker on a clause is optional when the clause
order is SVO, but required when the clause order is
other than SVO. So all three of these would be
Ma vito es palu. (SVO marked) I see cat.
Ma vito palu. (SVO unmarked) I see cat.
Es palu ma vito. (OSV marked) Cat I see,
But this would be incorrect:
*Palu ma vito. (OSV unmarked)
That's just my take on what I've seen so far.
Personally, I've kept my input to a minimum since I
had very definite ideas on the direction I wanted it
to evolve, and it's not evolving "my way" at all! :-)
So be it!
> One thing that makes the ranking a bit interestingis that there
> is no way to object to part of the sentence. So,for example, say
> someone has written a long sentence with words thateveryone
> agrees on, but glosses it in the past tense andfails to use "dun".
> The result is ungrammatical (at least, according topopular
> demand), but only because of the omission of "dun";not because
> of anything else in the sentence.
Good point. I'm not sure how to fix this, except maybe
to have two ranking numbers for each sentence; a
grammar rank and vocabulary rank. Perhaps when I get
the dictionary pages up a running with their own ranks
for each word then specific words can be voted on in
the dictionary and the sentence ranks can all be
assumed to involve grammatical issues.