Closed funderburkjim closed 3 years ago
Copies of md.txt and md_error.txt (adjusted to have just one suggestion per line) are in temporary repository: https://github.com/sanskrit-lexicon/temp_corrections_ap90
These are ready for @sanskritisampada to pull, process, and push.
Note to self:
I have not yet processed Sampada's corrections to MCI.
This review looks to have a few additional cases to correct, although most suggestions will probably be 'OK' for one reason or another.
Oh, ok.
@sanskritisampada has done the md corrections. Still for me to install.
Corrections installed.
31 lines changed. No print changes to report.
@sanskritisampada Now finished with this round of 'md' corrections. My few notes now in md_error.txt.
Now finished with this round of 'md' corrections.
Do you have a list or table to track who is there next on the list or off the list?
I am going alphabetically down the list of xx_error.txt files in https://github.com/sanskrit-lexicon/CORRECTIONS/tree/master/english_error/output.
alphabetically down the list of xx_error.txt files
So half way, but big ones ahead. And English errors will not help in German dictionaries. Remember you had the old orthography libraries running for PWK and PWG?
For the German dictionaries, we first would need to have someone fluent in German. @sanskritisampada also knows French, but not German, I think.
Since @drdhaval2785 is generating the xxx_error.txt files using, I think, Pyenchant, he could also do a similar generation for the German dictionaries (and for the French).
However, for the German dictionaries, there is an additional difficulty dealing with systematic differences between the German language at the time of Bohtlingk and the current German language.
he could also do a similar generation for the German dictionaries (and for the French).
@drdhaval2785 can you? Jim forgot he did it 2014 already https://github.com/sanskrit-lexicon/CORRECTIONS/issues/18
German dictionaries, there is an additional difficulty dealing with systematic differences between the German language at the time of Bohtlingk and the current German language.
No, it's not an issue. If the Pyenchant
will do it's part, I will start the German reading and will find a German to continue it.
I can definitely do for the French. My German is not good enough for these corrections.
On Tue, 6 Apr 2021, 03:31 funderburkjim, @.***> wrote:
For the German dictionaries, we first would need to have someone fluent in German. @sanskritisampada https://github.com/sanskritisampada also knows French, but not German, I think.
Since @drdhaval2785 https://github.com/drdhaval2785 is generating the xxx_error.txt files using, I think, Pyenchant, he could also do a similar generation for the German dictionaries (and for the French).
However, for the German dictionaries, there is an additional difficulty dealing with systematic differences between the German language at the time of Bohtlingk and the current German language.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sanskrit-lexicon/csl-corrections/issues/59#issuecomment-813759082, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACTKJX7DPT6DFOONLNRNZJLTHJQARANCNFSM4Z6WUGPQ .
We've already been through English errors once several months ago with MacDonell.
This work examines latest suggested corrections md_error.txt prepared by Dhaval.
This review looks to have a few additional cases to correct, although most suggestions will probably be 'OK' for one reason or another.