Screenshots of what I think is increased capitalization problems in word bank questions
I've submitted these as bug reports as well.
They’re from the iOS app.
Edit to add: actually, one of them (the one with the SeS problem) is from the web version, accessed from an iOS phone.
I just went back and tried the music skill via my phone’s web browser, and the increased capitalization problem seems to remain.
It seems about the same to me on the iOS app and the web-based site accessed from an iOS phone -- some of them have a few words properly capitalized, and some of them appear to be entirely in lowercase. For example in the sentence My flute has twelve holes, Dov'aghwIj is correct, but DISuS is not. The first screenshot below is the web version, the second is the app.
It seems to be a problem in handling '. Words that begin with ' mess up the tiles. I've actually seen contributors for other languages complaining about the problems that ' causes, but somehow I've never come across it for our Klingon course. It's going to be a problem for the upcoming Gaelic course, too.
Huh. I’m pretty sure I’d noticed a surfeit of ’s in multi-word tiles, but I definitely hadn’t picked up on the fact that it happens when words start with that letter. That’s an interesting consequence of programming for a smaller set of languages than the software ultimately supports.
For what it’s worth, I think they’re working on the iOS capitalization problems. Right now it’s technically worse (almost every sentence includes entirely lower case tiles), but I’ve actually found that less distracting than having some words partially correct.
It seems improved, but not entirely solved. When I noticed the change, I saw an uptick in both the quantity of capitalization errors and the diversity of categories those errors fell into. For example, the letter I (“i”) was suddenly a problem, and it was a problem in nearly every exercise, which was significantly more frequent than the errors (e.g. the letters Q and q) I was used to seeing.
As of the last day or so, the overall frequency of errors appears to have improved. It’s not perfect, but seems to be roughly the average I was used to seeing.
The increased diversity of error types remains, however. I’m still seeing errors with the letters I and S, for example. But those happen fairly infrequently, which is an improvement over the last two weeks or so.