Re: [edict-jmdict] adj-i/adj-ix issues

[clindsay@*********]

> I'm not sure you would really need special casing for just いい entries either, to put いい first. You could maybe have a general rule telling your program to check if there's a more common "noinfl"-tagged reading and if there is, default to showing that, if it's compatible with the same kanji. Either way, it really doesn't seem like a major issue to me.

I could. My reluctance is just that there's nothing in the meaning of the "noinfl" tag that implies that it should be the default reading to display for text that deconjugates to a less common different reading. That's something specific to いい／よい (mostly because it's such a common case, and displaying よい would be confusing to beginners). If the "noinfl" tag were added to other words in the future, I might not want that same behavior. 

All that said, I could work around that. And as I read this conversation and think on it more, I'm growing more comfortable with the idea of a "restricted inflection" tag. I don't think it would require as much added complexity in my deconjugation logic as I originally imagined. And my sense is that the subtleties of which inflections/derivations are restricted could be handled by application developers, to the degree that's appropriate for their specific application.

Chris