Frequently Asked Questions
Q: What are the current limitations of ichi.moe?
- There's a limit on the length of query. You can't copypaste entire novels into it.
- Personal names, place names and so on are not recognized. They will be most likely segmented wrong.
- Some grammatical patterns are not yet supported.
- Some informal patterns of writing are not yet supported.
- Words (either rare or neologisms) that are not in JMDict dictionary are not supported.
Q: Where do the definitions come from? What to do about incorrect definitions?
The definitions are taken from JMdict project. You can try to submit/edit definitions at JMdictDB, maybe eventually they'll be accepted and imported back into ichi.moe.
Q: This segmentation makes no sense, how can I improve it?
You can add spaces into the sentence to aid the segmentation. You can also select one of the 5 best segmentations that ichi.moe came up with by clicking on the parts of the original query.
Q: How can I report incorrect segmentation?
You can send an email to admin or raise an issue on Github.
Q: What is the algorithm behind ichi.moe?
Ichi.moe is using the open-source algorithm called Ichiran. It is written in pure Common Lisp. Contributions are welcome.
Q: Is the site itself open-source? What is it implemented in?
The site is not open-source, but it's pretty trivial. It's also mostly implemented in Common Lisp. If you want to help with the web design, please contact me.
Q: I got a "rebuilding suffixes" warning. What does it mean?
It means either the server just got restarted, or suffix definitions got updated. For performance reasons a cache of all possible suffixes needs to be maintained. While this is happening the quality of results might not be optimal.