Cleanup of multi-language processing.

There are at least 4 cases where the processing can be greatly improved.

Case 1: When processing a non-English Bible without a Reverse-interlinear, I am still offered the option of Bible Word Study although the only things built are Topic and Textual Searches. Note the Topic is built from English:lust rather than German:lust and therefore would often be erroneous. I would prefer (a) that the Topic section not be built when the source language is not English (or the variant of this that applies to the Logos structure), (b) that the Bible Word Study not be offered when the only section available is the Textual Search section and (c) if the BWS must be offered, the Textual Searches include the language of origin.

Case 2: This case shows the BWS as requested from the screen shown above. Note the Topic Section built off the English rather than the German. Running the second Textual Search option (on the NRSV) I get completely erroneous results because the search argument is simply "lust" rather than the correct "German:lust".

Case 3: Modifying the Search to the correct search argument "German:lust" and searching all resources fails to find my starting point, the Luther Bibel, because the language parameter ignores the metadata language and searches only for languages tagged within non-German resources. The metadata language should be respected and reflected in the search results. My bad - I had Match Case set from yesterday's testing.

Case 4: In the Luther Bibel, when I double click to obtain a dictionary definition the metadata defining the language is ignored and the German:lust produces results for the English:lust. Dictionary lookup should respect the language of the select word whether defined by metadata or internal tagging. Edit: Bradley identified this as an error in the dictionary rather than the Search.

This is in no way limited to German. I actually started in Finnish. However, I chose German as the language of illustration in hopes that some of the German speaking Logos users could expand on the issues.

Find more posts tagged with

Comments

Rosie Perera

Case 3: Modifying the Search to the correct search argument "German:lust" and searching all resources fails to find my starting point, the Luther Bibel, because the language parameter ignores the metadata language and searches only for languages tagged within non-German resources. The metadata language should be respected and reflected in the search results.

Case 4: In the Luther Bibel, when I double click to obtain a dictionary definition the metadata defining the language is ignored and the German:lust produces results for the English:lust. Dictionary lookup should respect the language of the select word whether defined by metadata or internal tagging.

[Y]

I recognize that some of this report was likely inspired by a question I asked, and your response to it. Thank you for probing further and articulating these shortcomings in Logos.

Bradley Grainger (Logos)

Case 3: Modifying the Search to the correct search argument "German:lust" and searching all resources fails to find my starting point, the Luther Bibel

Can't repro; do you have "Match case" turned on?

Bradley Grainger (Logos)

Case 4: In the Luther Bibel, when I double click to obtain a dictionary definition the metadata defining the language is ignored and the German:lust produces results for the English:lust. Dictionary lookup should respect the language of the select word whether defined by metadata or internal tagging.

Your diagnosis is precisely backwards. [:)]

The dictionary lookup is performed for the German word; NDCS is opened because it is incorrectly tagged as having German headwords.

MJ. Smith

https://community.logos.com/discussion/comment/707256#Comment_707256

Case 3: Modifying the Search to the correct search argument "German:lust" and searching all resources fails to find my starting point, the Luther Bibel

Can't repro; do you have "Match case" turned on?

You are correct - my bad [:$]

MJ. Smith

https://community.logos.com/discussion/comment/707258#Comment_707258

Your diagnosis is precisely backwards.

The dictionary lookup is performed for the German word; NDCS is opened because it is incorrectly tagged as having German headwords.

I'll have to take your word for it as I can't confirm it:

Will I have to load SQLite workbench and start snooping? [:D]

Rosie Perera

https://community.logos.com/discussion/comment/707264#Comment_707264

The dictionary lookup is performed for the German word; NDCS is opened because it is incorrectly tagged as having German headwords.

I'll have to take your word for it as I can't confirm it:

It actually looks as though you've just disproved it.

MJ. Smith

https://community.logos.com/discussion/comment/707266#Comment_707266

False friend error in BWS - a variant on the problem above. German Bald means "soon" but selecting "bald", richt click for context menu, select Bible word study brings up English "receding hair line"

delete12066188

https://community.logos.com/discussion/comment/707276#Comment_707276

welches Resultat bekommst du mit der deutschsprachigen Seite?

delete12066188

https://community.logos.com/discussion/comment/707335#Comment_707335

With my Understanding the Problem is that a Software can't read and decide is this german or english. How should Logos know that your bald is bald and not bald without saying him that you read in german. Maybe a search "just german" would helb?

Sascha

Mark Barnes

https://community.logos.com/discussion/comment/707266#Comment_707266

I'll have to take your word for it as I can't confirm it:

It actually looks as though you've just disproved it.

The language specified for the Headwords is independent of the language specified by the headings. You can check the language specified for the headwords by looking at the indexes section of the resource information panel.

Francis

https://community.logos.com/discussion/comment/707354#Comment_707354

Changing the language setting of the software does not change the situation one bit.

The language support needs to be much more thorough-going, whether it is different languages in resources or as the basis for interacting with the software. It would seem that in the present case, the German "Lust" should have been understood rather than the English "lust". Unfortunately, the german: prefix does not work in BWS.

Thomas Reiter (Faithlife)

https://community.logos.com/discussion/comment/707419#Comment_707419

Changing the language setting of the software does not change the situation one bit.

As it is not about the language of the interface... The interface itself is merely a labeled conduit that makes everything else possible, but without owning the RIB and future Datasets (which you cannot yet) or the respective relevant works (which I assume you don't), there is no area where it can produce these results from. That is why you receive the results from your English resources, simply because you do not own any German resources it could pull these from.

These are results that I get when running a search, just for comparison...

Now there is still some issues with this that need to be worked out like language specific casing, "completeness" of results, etc., but with the right resources and Datasets it does/will work!

Francis

https://community.logos.com/discussion/comment/707447#Comment_707447

Thank you for the informative reply, Thomas. I notice that your example is a word distinctively German (Pharisäer would not be confused for an English word by BWS). Will BWS be able to distinguish between Lust and lust? Will a language-based choice be offered -- "commencement" in French or in English, for instance, in various guides and search windows?

I noticed also that Pharisäer is not capitalized in the griechische wheel, though Φαρισαῖος is.

Also, I own Ziefle's dictionary of theological German words. At this point, I can get it to show an entry in topics if I use the context menu on the word in the dictionary to run BWS on it, but otherwise, typing the word directly in BWS does not work. I suppose this will be worked out in the fuller packages you are speaking of? Will people who have an English base package be able to look up German words or would a German base-package be required for that?

MJ. Smith

https://community.logos.com/discussion/comment/707447#Comment_707447

Tom, thanks for the information. What concerns me is not the absence of information when there is none - but the treating of German as English when there is no German. Because I own a Spanish Bible with a Reverse-interlinear, Spanish works correctly in the BWS. And, yes, I have tried to avoid using examples where I don't think the resource is properly tagged for language by using language:word searches to identify test cases.

MJ. Smith

https://community.logos.com/discussion/comment/707455#Comment_707455

Francis, what I am "arguing" is that words found by the search german:bald should behave differently than words found by the search english:bald

Francis

https://community.logos.com/discussion/comment/707482#Comment_707482

Francis, what I am "arguing" is that words found by the search german:bald should behave differently than words found by the search english:bald

I understand that. So, you could rephrase my question to Thomas above: will the future product he alludes to differentiate between German bald and English bald?

Thomas Reiter (Faithlife)

https://community.logos.com/discussion/comment/707496#Comment_707496

Francis, what I am "arguing" is that words found by the search german:bald should behave differently than words found by the search english:bald

I understand that. So, you could rephrase my question to Thomas above: will the future product he alludes to differentiate between German bald and English bald?

That is the plan, as anybody(or you are rightly so) would argue that that is probably very much so essential!

delete12066188

https://community.logos.com/discussion/comment/707522#Comment_707522

here is a another example. Wortstudy Sin in english and a German Wüste Sin.

Sascha