Bugs in the Westminster Hebrew Morphology module

[garfunkel]

Hi @ all,

I repost an issue here which has already been discussed in the German speaking forum some years ago: viewtopic.php?f=6&t=4764

There seem to be some issues in the parsing of the WHM, f.e. in Lev 13:58:



Another wrong parsing I found is in Jud 17:2:



I'm not sure if there are similar problems with other verb stems (both of these are pual problems) - if I find other mistakes, I'll post them.

I would very much appreciate an update of the whm database or other bugfix!

Thanks a lot and God bless
Simon

[JG]

your images do not appear, can you fix them

[garfunkel]

your images do not appear, can you fix them

Sorry, I'll attach them to this post.

[JG]

I have forwarded your observations to GrovesCenter.org as they are the ones that supply us the data.
We are not up to date, so this might have been fixed already or it might not be an issue.

[garfunkel]

I have forwarded your observations to GrovesCenter.org as they are the ones that supply us the data.
We are not up to date, so this might have been fixed already or it might not be an issue.

Thank you very much. Do you think the module might be updated to the latest WHM version soon? I read (in the German speaking topic) that the newest version is 4.18, but the theWord module still is 4.12...

EDIT: I just read, that release 4.20 has already come out in Jan 2016 (see here: https://grovescenter.org/westminster-he ... -jan-2016/).

[JG]

I got a reply on the one issue. It is a mistake we made in producing the module, for what I can see.
The reply on the Judges 17:2 was that it is 'Qal Passive'. But we had it as a small q and so just qal. The supplied data shows this ju17:2,7.1 LU75Q.A75X- [email protected]

Waiting for a reply on the other point.


As to the update, that will be for Costas to decide how best to use the small resources we have. It was a complex job as I recall.

[garfunkel]

I got a reply on the one issue. It is a mistake we made in producing the module, for what I can see.
The reply on the Judges 17:2 was that it is 'Qal Passive'. But we had it as a small q and so just qal. The supplied data shows this ju17:2,7.1 LU75Q.A75X- [email protected]

Ah, that's interesting - never heard of a Qal Passive so far. Seems to be a debate among scholars (https://greekandhebrew.wordpress.com/20 ... ive-exist/).

Are you going to fix the bug? There seem to be 69 occurences of a Qal Passive in the WLC (see attached image).

Waiting for a reply on the other point.

Just found out that this is the exact same problem: there is a big "P" in the data, but you just have a small "p". Again, see the attached image.

Maybe there are more bugs like these because of the case insensitivity? Might be good to check the coding in general?!

EDIT: As far as I can see it's the same with Hofal (big "H") and Hifil (small "h").

As to the update, that will be for Costas to decide how best to use the small resources we have. It was a complex job as I recall.

I understand - thanks in advance for a short notification on that.

[garfunkel]

Any updates on the issue?

[JG]

The problem is not with the data as such. When theWord builds the word index it is down casing and so the capital letters are lost from the index.
Costas will need to fix theWord so that it does not do this, or make a custom index.
I will ask him to make a comment on the situation.

[garfunkel]

The problem is not with the data as such. When theWord builds the word index it is down casing and so the capital letters are lost from the index.
Costas will need to fix theWord so that it does not do this, or make a custom index.
I will ask him to make a comment on the situation.

Alright, thank you.

[JG]

Costas asked me to post some explanation back.
The situation is a bit more complex. There needs to be a fix to theWord, and to the word index of the module and perhaps an additional property to indicate to theWord that no down casing should be performed for wordclick/Lookup/popup. It is a complex fix, so will take some time.
The workaround for the time being is to have the morphology codes visible to see that the popup is correct to that data. If it is not then you will have to click the morphology and look in the list for the matching code that has capital letter as you require.
I believe these are the only codes that cause an issue.

whm pairs.zip

(801 Bytes) Downloaded 320 times

[garfunkel]

I understand. I'd really appreciate the update even if it takes some time.
Thanks in advance for all your efforts!

[JG]

Just an update.
The fixes have been done. But we have to do the release of both the module and theWord.

[garfunkel]

Just an update.
The fixes have been done. But we have to do the release of both the module and theWord.

Thanks a lot! Looking forward to the releases...

[garfunkel]

Just an update.
The fixes have been done. But we have to do the release of both the module and theWord.

Thanks a lot! Looking forward to the releases...

Any update on when the releases will come out (even if it's only a beta version)?