miketisdell Posted October 8, 2018 Author Share Posted October 8, 2018 Sorry, no, it's a separately licensed database with attendant royalties. Can Accordance make fixing the search issue in the HMT-W4 a priority? Link to comment Share on other sites More sharing options...
Michel Gilbert Posted October 8, 2018 Share Posted October 8, 2018 Hi Mike, In less than a minute in the Construct Search I get 440 hits to check. Regards, Michel 2 Link to comment Share on other sites More sharing options...
miketisdell Posted October 8, 2018 Author Share Posted October 8, 2018 Hi Mike, In less than a minute in the Construct Search I get 440 hits to check. אדנָי.png Regards, Michel That works 1 Link to comment Share on other sites More sharing options...
miketisdell Posted October 8, 2018 Author Share Posted October 8, 2018 While I am really glad to see a solution, it would be really nice to see a vowel sensitive search work correctly. 1 Link to comment Share on other sites More sharing options...
miketisdell Posted October 8, 2018 Author Share Posted October 8, 2018 Hi Mike, In less than a minute in the Construct Search I get 440 hits to check. אדנָי.png Regards, Michel Interestingly, I am getting 457 hits and I think mine is correct i.e. I am getting hits for items like Ge. 18:30 that have a prefix. I wonder why you are getting less hits. Link to comment Share on other sites More sharing options...
Michel Gilbert Posted October 8, 2018 Share Posted October 8, 2018 Hi Mike, Not sure, but Gen 18,30 is on my list. Regards, Michel Link to comment Share on other sites More sharing options...
miketisdell Posted October 8, 2018 Author Share Posted October 8, 2018 Hi Mike, Not sure, but Gen 18,30 is on my list. Gen 18,30.png Regards, Michel Is there a way to export a CSV of the verse references, so that we could compare the lists and identify the differences? Link to comment Share on other sites More sharing options...
Michel Gilbert Posted October 8, 2018 Share Posted October 8, 2018 It’s Thanksgiving Day today in Canada, and we’re on our way out. I think Accordance could probably figure this out. Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 8, 2018 Share Posted October 8, 2018 Hi Mike, You can just select all the hits - in the search tab ctrl-A - and then Copy As -> References and paste that into a file somewhere. What I did with TYA's BW refs was to then combine all of the refs into a single list separated by ; and then queried them in one tab. Then ran my own query in another tab. Then in a third did a verse search for [CONTENTS tab1] <NOT> [CONTENTS tab2]. That will reveal the differences. Thx D 1 Link to comment Share on other sites More sharing options...
Graham Buck Posted October 9, 2018 Share Posted October 9, 2018 Interestingly, I am getting 457 hits and I think mine is correct i.e. I am getting hits for items like Ge. 18:30 that have a prefix. I wonder why you are getting less hits. You probably have "Search both directions" ticked at the bottom of your Construct Search. With it ticked you get 457, without you get 440. In effect it's doing two searches. The one you've attempted to find, along with a search for a 1st-comm-sing suffix within 1-1 of a plural adon. 1 Link to comment Share on other sites More sharing options...
miketisdell Posted October 9, 2018 Author Share Posted October 9, 2018 (edited) You probably have "Search both directions" ticked at the bottom of your Construct Search. Screen Shot 2018-10-08 at 3.40.18 PM.png With it ticked you get 457, without you get 440. In effect it's doing two searches. The one you've attempted to find, along with a search for a 1st-comm-sing suffix within 1-1 of a plural adon. Yep, that is it. Reversing the order and turning off both directions showed where false hits were found. I have updated the search to look like this. However, I now realize that this search is still not differentiating between אֲדֹנַי and אֲדֹנָי. Grammatically both are the same, but functionally they are not. This search is an improvement but it is not a fix. Edited October 9, 2018 by miketisdell Link to comment Share on other sites More sharing options...
Michel Gilbert Posted October 9, 2018 Share Posted October 9, 2018 Hi Mike, That is why I said “to check.” I agree that there should be vowel sensitive searches for cases like this, and for others. At least in this case the overwhelming majority of hits are what you are looking for. Regards, Michel Link to comment Share on other sites More sharing options...
miketisdell Posted October 9, 2018 Author Share Posted October 9, 2018 (edited) More odd results!Adding an <AND> and including "אֲדֹנָי=" provided only 10 hits, none of which should have matched the construct search i.e. all ten hits word אֲדֹנָיו Searching only for "אֲדֹנָי=" 39 hits; all hits included an additional part of the suffix. Edited October 9, 2018 by miketisdell Link to comment Share on other sites More sharing options...
miketisdell Posted October 9, 2018 Author Share Posted October 9, 2018 (edited) Hi Mike, That is why I said “to check.” I agree that there should be vowel sensitive searches for cases like this, and for others. At least in this case the overwhelming majority of hits are what you are looking for. Regards, Michel I agree in this case it gets me much closer, but this is still pretty broken. FYI - I was able to construct a search in bible works looking only for instances of אֲדֹנַי and I found that there was only one instance in the OT. That was something I didn't know before. To double check, I built a search that looked at all verses that any word with the suffix נַי that also contained the lemma אדון and the only valid hit was Gen. 19:2. It would be nice to be able to duplicate a search like this in Accordance. וַיֹּ֜אמֶר הִנֶּ֣ה נָּא־אֲדֹנַ֗י ס֣וּרוּ נָ֠א אֶל־בֵּ֙ית עַבְדְּכֶ֤ם וְלִ֙ינוּ֙ וְרַחֲצ֣וּ רַגְלֵיכֶ֔ם וְהִשְׁכַּמְתֶּ֖ם וַהְלַכְתֶּ֣ם לְדַרְכְּכֶ֑ם וַיֹּאמְר֣וּ לֹּ֔א כִּ֥י בָרְח֖וֹב נָלִֽין׃ (Gen. 19:2 BHS) Edited October 9, 2018 by miketisdell Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 9, 2018 Share Posted October 9, 2018 Hey Mike, You are going to have trouble doing a search based on function without additional tagging in the text. The search I provided https://www.accordancebible.com/forums/topic/24642-cannot-search-for-%D7%90%D6%B2%D7%93%D6%B9%D7%A0%D6%B8%D7%99/?do=findComment&comment=121214gets very close to the same results and the functionally tagged (proper name) ETCBC. Even in the BW results that TYA sent me there are more erroneous hits if one takes ETCBC as the standard (and feel free to dispute that - but it's the only functionally tagged option I have available). In particular there number of false negatives (those excluded that should have been included) is higher. The number of false positives is the same as my search. Higher false positives means more weeding of the results manually to exclude them, while false negatives means you don't see hits you should be seeing. In general I prefer a higher false positive rate so long as it's not too high because at least you don't miss results. It appears you are trying to get an exact result where there no false positives nor any false negatives. I am not sure this is possible with either BW or Accordance in the HMT texts, given what I've seen. It is only possible in Acc in ETCBC if, and only if, the tagging of the proper name is accurate and I haven't checked and it would take a while. Thx D Link to comment Share on other sites More sharing options...
Timothy Jenney Posted October 9, 2018 Share Posted October 9, 2018 (edited) Here is a search that works, recommended by Accordance's chief programmer: it is an inflected search for "adon," followed by an inflected search for a yod, limited by the @ sign to find only those cases where the yod also includes a specific character, a qametz. The symbol for a specific character is a period (.), which is actually typed before the qametz. However, the screenshot makes it look like the qametz is before the period. He also pointed out that the Hebrew letter search (literal) does not support specific vowels, which is why the different vowels kept popping up in these searches. Edited October 9, 2018 by Timothy Jenney 2 Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 9, 2018 Share Posted October 9, 2018 This looks very close to the construct I posted the other day. I'll add it to the workspace I've got on this problem and see how its result compare, but off hand how many hits do you get ? thx D Link to comment Share on other sites More sharing options...
Michel Gilbert Posted October 9, 2018 Share Posted October 9, 2018 438 2 Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 9, 2018 Share Posted October 9, 2018 Thanx Michel. I think that's what mine got - I'll have to check. Thx D Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 10, 2018 Share Posted October 10, 2018 Ok, so this produces same hits as my construct. Interesting to see the י in the quotes and the qamats attach to it though. Nice trick. Thx D 1 Link to comment Share on other sites More sharing options...
miketisdell Posted October 10, 2018 Author Share Posted October 10, 2018 Hey Mike, You are going to have trouble doing a search based on function without additional tagging in the text. The search I provided https://www.accordancebible.com/forums/topic/24642-cannot-search-for-%D7%90%D6%B2%D7%93%D6%B9%D7%A0%D6%B8%D7%99/?do=findComment&comment=121214gets very close to the same results and the functionally tagged (proper name) ETCBC. Even in the BW results that TYA sent me there are more erroneous hits if one takes ETCBC as the standard (and feel free to dispute that - but it's the only functionally tagged option I have available). In particular there number of false negatives (those excluded that should have been included) is higher. The number of false positives is the same as my search. Higher false positives means more weeding of the results manually to exclude them, while false negatives means you don't see hits you should be seeing. In general I prefer a higher false positive rate so long as it's not too high because at least you don't miss results. It appears you are trying to get an exact result where there no false positives nor any false negatives. I am not sure this is possible with either BW or Accordance in the HMT texts, given what I've seen. It is only possible in Acc in ETCBC if, and only if, the tagging of the proper name is accurate and I haven't checked and it would take a while. Thx D I understand false positives and false negatives when the pattern matches. But I consider it a bug when I get false positives or false negatives when the pattern does not match the results i.e. like those in some of the examples I provided. Link to comment Share on other sites More sharing options...
Joel Brown Posted October 10, 2018 Share Posted October 10, 2018 Mike, can you repost, or at least point out which post #s (see top right corner of each post - this one is #47) you claim are giving different results than the provided pattern? There's been a lot of discussion swirling around, but I believe most of it is due to the different tagging in the Westminister database. Link to comment Share on other sites More sharing options...
miketisdell Posted October 31, 2018 Author Share Posted October 31, 2018 Mike, can you repost, or at least point out which post #s (see top right corner of each post - this one is #47) you claim are giving different results than the provided pattern? There's been a lot of discussion swirling around, but I believe most of it is due to the different tagging in the Westminister database. I just saw this. I will try to go back and look soon. Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted October 31, 2018 Share Posted October 31, 2018 I just saw this. I will try to go back and look soon. FYI, if you want notification of posts you can set up you My Settings -> Notifications Options to auto-follow posts that you are contributing to and to then get notified of posts to that topic either in you notifications list or by email. Thx D 1 Link to comment Share on other sites More sharing options...
miketisdell Posted November 1, 2018 Author Share Posted November 1, 2018 I just saw this. I will try to go back and look soon. I did some checking under 12.3.0 (just updated). Searching for אדני with "letters" is greatly improved i.e. I am getting many hits that I was not getting previously; however, I don't seem to be able to exclude hits that have prefixes or suffixes. Searching for אדני with "words" still exhibits similar problems i.e. it returns only the hits with the first person singular declension. The popup window does not provide options for either the plural declension or honorific for YHWH. (These all have the same consonants but different vowels on the last letter). Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now