Serious problems with Arabic and Quran

[miketisdell]

There are a number of serious bugs that have been introduced into Accordance 13:

 

1. If I display the characters window when using an Arabic text (Quran or Van Dyke) and choose a letter on the displayed table, Accordance crashes instantly.
2. If I search for a lexem like اسم, Accordance inserts spaces between attached suffixes/prefixes i.e. بسم becomes ب سم (see Q1:1 in the attached screenshot)
3. I cannot use a consonantal text in the search box if the text I am searching for includes a hamza wasl i.e. searches for الله fail to produce any results. This is even more problematic because I have no way to enter a hamza wasl from the keyboard or from the characters window. 

 

 

[miketisdell]

Accordance, any updates?

[miketisdell]

Accordance, any updates?

[miketisdell]

Accordance, any updates?

[Silas Marrs]

1. I'll fix this crash, hopefully in time for the next release.

 

2. This seems to be a bug.  I kinda looks like it has moved the space from the right of the  ب  to the left of it.  I'll fix it as soon as I can.

 

3. I'm honestly not sure what this means.  Is there any way to explain it that I could understand if I know no arabic?  When I copy/paste الله into the search for Van Dyke, it gets about 2142 results.

Edited December 11, 2019 by Silas Marrs

[Helen Brown]

#2 is not a bug. It is working like the Hebrew texts which have morphological separators between the prefixes and suffixes and the main words. These are invisible when viewing the text, but you can see them in Instant Details and the parsing tab, as well as any time you try to search a selection. They greatly help to look up words in a lexicon, and to expand the search to not just include the specific example, but all examples of that lexical form (if you remove the prefixes and suffixes from the search, or select only the main word before you search).

[miketisdell]

#2 is not a bug. It is working like the Hebrew texts which have morphological separators between the prefixes and suffixes and the main words. These are invisible when viewing the text, but you can see them in Instant Details and the parsing tab, as well as any time you try to search a selection. They greatly help to look up words in a lexicon, and to expand the search to not just include the specific example, but all examples of that lexical form (if you remove the prefixes and suffixes from the search, or select only the main word before you search).

 

Just to be clear, this isn't working like it does in Hebrew; I do understand that there are morphological separations, but the spaces are being added into the main text i.e. it would be like תשופנו was changed to תשוף נו in the main text when a search for the lexeme was performed. In the Arabic Quran text a space is added and the letter form changes in the main text widow (See the image I posted in the original post). 

[miketisdell]

1. I'll fix this crash, hopefully in time for the next release.

 

2. This seems to be a bug.  I kinda looks like it has moved the space from the right of the  ب  to the left of it.  I'll fix it as soon as I can.

 

3. I'm honestly not sure what this means.  Is there any way to explain it that I could understand if I know no arabic?  When I copy/paste الله into the search for Van Dyke, it gets about 2142 results.

 

 

I will try and explain number 3 a little more clearly:

 

In Hebrew it would be similar to trying to search for אֱלֹהִים by using only the consonantal text i.e. אלהים and having those searches fail because I did not enter the alif with the pointings i.e. אֱ; The Arabic searches are treating this as if it were a completely different letter. I can highlight the word الله in the text and paste it into the search and then it returns ONLY those instances that have those same marking, but I have no way of entering the Alif with the markings i.e. the hamza wasl from the keyboard so any word that has this must be located first and then the text selected and copied. Additionally, when الله has different markings on the alif then those examples are not found at all. Also when typing the consonantal text of الله into the search field, the popup box with suggestions never includes الله as an option. 

 

If you look at this example of Allah from the Quranic text, you will see the marking above the Alif that kind of looks like a tadpole, it is the alif with this marking that cannot be entered from the keyboard (this is the hamza wasl). 

 

You can see form above used in the first verse (Sura 1:1), and the consonantal form entered in the search box; when the search is executed, there are no entries found.

 

If I copy and past the text for الله that is found in Sura 1:1 then the entries displayed here are not found because the markings on the alif are the same. I found these by searching for الله enclosed in quotes i.e. "الله"; both this example and the example in Sura 1:1 are the same lexeme but there is no way to do a lexeme search that finds both examples. 

[miketisdell]

1. I'll fix this crash, hopefully in time for the next release.

 

2. This seems to be a bug.  I kinda looks like it has moved the space from the right of the  ب  to the left of it.  I'll fix it as soon as I can.

 

3. I'm honestly not sure what this means.  Is there any way to explain it that I could understand if I know no arabic?  When I copy/paste الله into the search for Van Dyke, it gets about 2142 results.

 

Based on your search results you provided, I attempted to do the consonantal search from the Van Dyke and the tagged Van Dyke NT and both worked; the bug described in item 3 appears to only be a problem with the Quran-T.

 

This works.

 

This also works.

[Joel Brown]

Thanks for the reports, Mike.  Your reports (1) and (2) are now fixed for the next build.  I think item (3) is a change to the text file itself, which we are working on now.