Linas Vepštas

... and version 5.10.2 is now available.

I looked at the powerpc build log: https://buildd.debian.org/status/fetch.php?pkg=link-grammar&arch=powerpc&ver=5.10.2%7Edfsg-1&stamp=1631821070&raw=0 it says the `multi-dict` unit test failed. This is not python... it tests the use of multiple dictionaries in multiple threads. No...

Hmm. If a word is capitalized and unknown, we should spell-guess the lower-case version. If a word ends with an s and is unknown, we should spell-guess it. In other...

Post-processing needs to be avoided at all costs. Post-processing is a fail. I cannot write an exact spec just right now. My train of thought, though, is that everything we...

For alternatives to correctly-spelled words ... not sure. we need to add some sort alternatives file. I don''t see how to jam alternatives into the current dictionary format. So maybe...

one super-hacky way of indicating alternatives is to use word-subscripts. So, for example: ``` yisser.your: A+ & B-; ``` means that "yisser" is an acceptable Irish word, and it really...

Ah hah! Silly me, you are right! This works today, without any changes at all to the C code: ``` than.#then: [then.r]0.3333; ``` which gives ``` linkparser> shall we go,...

Re: post-processing: Let me explain it this way: if there is some optional "post-processing" utility that maybe tags words with some additional info, or does some other light-weight processing, that...

So: pull request #425 implements some basic typo support. Things that don't work (can't work without wordgraph support): replacing there by they're, replacing all by all_of

well, instead of sub-dictionaries, the problem would be solved by "dialect support" - #402 : turn off the "bad speling" dialect, and then these rules no longer apply.