Verse popups all wrong...

I'm a new Logos user and apologize if this is the wrong place to put this note or if it has already been reported. I've read 4 books and have noticed a recurring error thus far. The algorithm used to popup the Bible verse seems to be wrong. When the author lists a verse without the book reference, Logos tries to identify what book to use. It is frequently wrong. I have sent this to customer service, but they thought I was identifying one error; I am trying to report a repeating error in all of the books, not a single typo.

Here are two examples, far from the only experiences in my first four books:

In James Montgomery Boice's "Foundations of the Christian Faith." On page 432 in the section following the header The Christian Way of Knowledge, the author is referring to John's first epistle and lists many verse references, all without stating the book. Because he writes, "John says..." the algorithm assumes it is the gospel of John and the popup of every verse referenced is wrong.

Another book is John MacArthur's Bible Handbook. While studying Ephesians, on page 398 under the title Background and Settings, he starts a paragraph, "The fledgling church begun by..." Mid-way, he references 1 Timothy 1:3, 20. In subsequent sentences, he references 1:4, 4:3 and 1:7. MacArthur is still talking about, and sometimes quoting from, 1st Timothy but Logos points us to those verses in Ephesians in the popups.

As I mentioned, I'm a new user and maybe everyone already knows of this recurring error. If not, it is disconcerting to me to pay so much for a product that is frequently wrong. In every book sales page on their website, Logos tells us "all Scripture passages link to your favorite Bible translation." Unfortunately, after reading four books, I've found this statement to be more wrong than right when the specific book is not referenced but only the chapter and verse.

I hope someone in the Logos organization reads this and reports to the engineers that their algorithm doesn't work and to the quality control people that they have a big issue on one of the prime features.

Find more posts tagged with

Comments

Jack Caviness

Have not seen this problem to the extent you are reporting. the proper way to call this to Logos' attention is to report typo. Select the incorrect text, right click, and select "Report Typo" from the left side of the menu. Then type "This should reference …" in the bottom field. Unless you tell them where the errors are located, no one will be able to repair them without spending a large number of precious man-hours.

Jonathan Pitts

There is a lot of this about, and I find it frustrating too.

The links are coded within each resource rather than generated as you read them. I get the impression that software is used to insert the links automatically, and it often gets it wrong. Mostly it is where there has been an oblique reference to a book without an actual chapter and verse reference. In commentaries it is then often assumed that subsequent number-only references (3:16) refer to the book the commentary is about rather than the book obliquely referred to. However, I note that one of your examples shows this the opposite way round.

I have also occasionally found situations where it appears that someone has forgotten to change the chapter setting on the link-inserting software. Then every link in the format v.8 is linked to the previous chapter and not the one being discussed.

Much as I love Logos and appreciate the work that the team do, I do find these frequent errors unsatisfactory. I cannot imagine any respected paper publisher accepting a similar frequency of important errors. Proper proof-reading is not just ensuring that everything is spelt correctly but ensuring that every link is to the correct information and makes sense in its context. Logos software is all about the ease of linking resources together, and these too-frequent errors cripple that core functionality.

I used to meticulously report every linking error as a typo, but it gets pretty tedious when you have repeated errors in a book or chapter. I now do most of my reading on the iPad and don't have option to report typos.

Although Logos has laudably relinked some resources (particularly useful in the Anchor Yale Bible Dictionary), I get the impression that correcting links scattered throughout many resources is not a high priority. I wouldn't mind taking the time to go back to my PC and report every error if I knew that it would rapidly be corrected to save the next reader frustration. The same applies to adding the links to resources that weren't available in Logos when the book was created.

I did have a suggestion on Logos Uservoice regarding this, but it seems to have disappeared. Surely there are other users out there who would be willing to be part of an elite team of proof-readers and link-inserters, correcting things as we go through a book that we are reading anyway. But we would need to know that the corrections would be made soon after we reported them.

I'd like to hear from Logos on how it intends to tackle this and get linking up to the highest standards.

NetworkGeek

https://community.logos.com/discussion/comment/372535#Comment_372535

Bad links and typos are a difficult problem. As a user I see a lot of them too, and even though I see updates to resources coming out (I have received probably 5 to as many as 10 updated resources each week-to-ten days over the last month), I realize there are two realities:

1, To users the number of errors is high enough that's it's very noticeable, and can be a drag/impediment on reading/work;

2. As a percent of total quantity of text coming out, the number of typos is absurdly small. If Logos has around 20,000 resources, 200 pages/resource, 500 words on a page, 5 letters per word, that's 10 billion characters! A 1% error rate in transcription/conversion is 100 million errors! 0.1% error rate (absurdly small by the way) would be 10 million errors in existing resources. The metrics are a guess here, but you get the idea - very small percentage of errors is a HUGE number of errors from a user's eyes.

So sadly, we are living in two different worlds that I don't think will ever merge. Logos' position to this point has been to do their best, Bob has published recently that this area is getting more attention in the best, but beyond that there are no clever ideas on the table. There is an implied "we are trying our best so deal with it." As the number of Logos users grows significantly like it has over the last 2 years, the noise of this quality issue won't go away with this plan, ever. New users will introduce the same complaint, and old users will chime in that yes, it's an issue, I guess we have gotten more used to it than you, young pup.

I'll add to this that there is a strong perception that printed books do not have this problem. Now they don't have things like links, but they do have footnotes and endnotes; and I would bet most users would say that finding an error in a book is rare. They do it with manual costly, labor, but they invest the labor because quality is important. The e-pub world needs to find a different way of dealing with this.

I introduce all this to again suggest, as others have in the past, maybe a unique approach would work better. What if there was a way where users could do the updates to resources? Would it have to be planned, thought through, tested, software written, etc.? Would there have to be some quality check on Logos' side, or software built that only basic things could be done (change a link, correct a basic typo of smallish size, etc.)? Sure. But consider this:

1. Even if users messed up some changes that were not caught, would they be successful 90% of the time? Would the resources get to a quality level well beyond what Logos could afford to offer cost-effectively?

2. Would the sense of user empowerment minimize the user aggravation about this issue, as now we would all be empowered to work together on this problem?

3. If a good process/tools design were possible, wouldn't quality go down and user perception of quality go up?

4. What if the plan started small, empowering users only to fix/update easy types of typos and/or links? Learn from the experience, do more in the future...

I know there is work, and skepticism in this approach. I don't believe Logos believes in it philosophically, I may be wrong about that however. I do know that the current approach will never work to satisfy users, and to say "deal with it" never works either. Users care about resource quality, Logos people are very clever, why not work together ...?

Jonathan Pitts

https://community.logos.com/discussion/comment/372539#Comment_372539

I introduce all this to again suggest, as others have in the past, maybe a unique approach would work better. What if there was a way where users could do the updates to resources? Would it have to be planned, thought through, tested, software written, etc.? Would there have to be some quality check on Logos' side, or software built that only basic things could be done (change a link, correct a basic typo of smallish size, etc.)?

Much as I dislike these cartoonish smilies:

[Y]

Proper typographical errors have the problem that the text needs to be checked against the printed edition of a book. One also has to have a policy on whether an error in the printed edition should be replicated in the Logos edition or corrected. The original publisher may need to decide this.

However, links are more straightforward. They are solely the domain of Logos. In fact, the Logos user who is reading and trying to understand a book is in a better position to correct these than an Logos employee or freelance proof-reader who may not be engaging with the meaning of the book. We can point a link to the exact part of a text referred to rather than just the nearest page or section.

Links would be the place to start involving users—both correcting them and inserting missing ones. As you say, involving us would make us feel empowered and involved in improving the standard of resources.

A short test piece, to check that a user is able to detect the errors and insert links appropriately would be wise. I would also want more guidance on how to format links to different types of resource (I believe things are somewhat different in the resources than for the links used in notes). I have also read that it is sometimes difficult to find the exact point in the text that a user-reported typo refers to. Perhaps a page of exported text (with the new links clearly inserted) and emailed to Logos by approved users would be better than the in-built typo reporting system. Some of the Logos team would need to keep up with these reports as they came in.

There is also the issue of how often corrected texts should be sent to users. Those with limited bandwidth don't want to be redownloading their entire library every week.

Taking all these factors into account, we're up for it. What's stopping you Logos?

Jack Caviness

https://community.logos.com/discussion/comment/372539#Comment_372539

Superb analysis and suggestions [Y]

Sent Bob Pritchett a link to this thread.

James Clabaugh

https://community.logos.com/discussion/comment/372528#Comment_372528

Thanks for pointing out the correct way. Now, my thoughts...

Will Logos do anything with these reports? The amount of time I will add to my reading would more than double if I reported every incorrect link. It would be quicker if I corrected the link myself and updated the book. Adding hours to my reading of each book would not be preferable unless I really believed it would help Logos fix the problem. One of the reasons I purchased Logos was an assumption of quality. I don't know how much time it takes them to convert a book into Logos but a little more time in quality control would not be noticed by the user who is waiting for the book.

If the author quotes the verse, I don't need the link. If the author does not quote the verse but only references it, I expect it to be correct so I'm willing to wait another week or month for the book to be completed. I hate the idea of taking hours and hours added to the reading of each book, so can someone convince me the Logos will correct the errors if I report them?

George Somsel

https://community.logos.com/discussion/comment/372558#Comment_372558

Will Logos do anything with these reports?

Yes. Perhaps it won't be corrected as soon as you might like, but it will be done.

Hunter Clagett

https://community.logos.com/discussion/comment/372558#Comment_372558

Thanks for pointing out the correct way. Now, my thoughts...

Will Logos do anything with these reports? The amount of time I will add to my reading would more than double if I reported every incorrect link. It would be quicker if I corrected the link myself and updated the book. Adding hours to my reading of each book would not be preferable unless I really believed it would help Logos fix the problem. One of the reasons I purchased Logos was an assumption of quality. I don't know how much time it takes them to convert a book into Logos but a little more time in quality control would not be noticed by the user who is waiting for the book.

If the author quotes the verse, I don't need the link. If the author does not quote the verse but only references it, I expect it to be correct so I'm willing to wait another week or month for the book to be completed. I hate the idea of taking hours and hours added to the reading of each book, so can someone convince me the Logos will correct the errors if I report them?

I appreciate your concern. We do, in fact, look at and address these kinds of bug reports on a continual basis. Many of the small somewhat mysterious updates that Logos downloads are just these kinds of resource corrections.

I recognize that it can seem like we are ignoring these issues because there isn't an immediate response. As mentioned above, many of the changes made have to be run by the publisher/author for approval, which can take time.

Additionally, our staff that would investigate and fix these types of typographical and linkage errors are the same people that are responsible for producing our NEW content. We have to balance corrections and fixes with the needs of our users to get new content out the door.

I have submitted this thread to our resources team, so they should be reviewing this issue shortly.

Thanks,

MJ. Smith

https://community.logos.com/discussion/comment/372558#Comment_372558

if I reported every incorrect link.

When there is a systemic error over a passage I report only the first error and indicate that the error continues in subsequent text. I find this less frustrating than reporting the errors individually.

Although I wouldn't actually want Logos to do it for cost and bandwidth reasons, I often wish resources indicated that an error had been reported. Then I'd know whether or not I needed to report it - or follow the link.

Jack Caviness

https://community.logos.com/discussion/comment/372597#Comment_372597

I have submitted this thread to our resources team, so they should be reviewing this issue shortly.

Thank you

Jack Caviness

https://community.logos.com/discussion/comment/372613#Comment_372613

I often wish resources indicated that an error had been reported. Then I'd know whether or not I needed to report it - or follow the link.

Seems that I read somewhere that the procedure for prioritizing typo fixing included the number of reports. More reports = more users affected. It is possible that I dreamed this, but it sounds good [8-|]

James Clabaugh

https://community.logos.com/discussion/comment/372670#Comment_372670

Jack, when I first reported this to support, they called it a TYPO also. I don't consider it a typo because it is a systemic errror in their algorithm for linking verses. It causes most verses that do not have a Book written with the Verse to potentially be wrong. A typo can be fixed by going to that particular notation but this problem requires either fixing their automatic process or manually proof-reading every linked verse.

If they fix things by the number of times it is reported, this one stands little chance of being fixed unless they accurately define and understand that the issue is prevelant throughout all their books..

Jack Caviness

https://community.logos.com/discussion/comment/375101#Comment_375101

I don't consider it a typo because it is a systemic errror in their algorithm for linking verses.

In compiling a personal book, I have found this to be true. Practically every verse reference that did not have a book name associated with it required a Bible datatype entry. This calls for a great deal of manual proof reading in a large resource.

I agree that Logos needs some major change in procedure here.

Dave Hooton

https://community.logos.com/discussion/comment/372552#Comment_372552

There is also the issue of how often corrected texts should be sent to users. Those with limited bandwidth don't want to be redownloading their entire library every week.

Logos have a 'patch' system for corrections so it is often not necessary that the full resource needs to be downloaded e.g. a 925 KB patch vs. 5.6 MB.

Louis St. Hilaire

https://community.logos.com/discussion/comment/375101#Comment_375101

Jack, when I first reported this to support, they called it a TYPO also. I don't consider it a typo because it is a systemic errror in their algorithm for linking verses. It causes most verses that do not have a Book written with the Verse to potentially be wrong. A typo can be fixed by going to that particular notation but this problem requires either fixing their automatic process or manually proof-reading every linked verse.

If they fix things by the number of times it is reported, this one stands little chance of being fixed unless they accurately define and understand that the issue is prevelant throughout all their books..

For Logos Editions, references like this that can only be understood from context are resolved by a human tagger, not through pure automation, so these errors are actually due to a human being misreading the context rather than a faulty algorithm.

A typo report is thus a good way to bring tagging errors like this to our attention for when the resource is updated. If a whole chain of references is off, a single report with an explanation in the "Comment" box should suffice.

Jack Caviness

https://community.logos.com/discussion/comment/375169#Comment_375169

For Logos Editions, references like this that can only be understood from context are resolved by a human tagger, not through pure automation, so these errors are actually due to a human being misreading the context rather than a faulty algorithm.

A typo report is thus a good way to bring tagging errors like this to our attention for when the resource is updated. If a whole chain of references is off, a single report with an explanation in the "Comment" box should suffice.

Thank you for the explanation. Will continue to report typos when discovered.

fgh

https://community.logos.com/discussion/comment/375162#Comment_375162

There is also the issue of how often corrected texts should be sent to users. Those with limited bandwidth don't want to be redownloading their entire library every week.

Logos have a 'patch' system for corrections so it is often not necessary that the full resource needs to be downloaded e.g. a 925 KB patch vs. 5.6 MB.

Still needs to index, though, which for me is the bad part.