Review sorting bug

If a review is edited and sorting by thumbs is set, the edited reviews sort after all the thumbs, as if there wasn't any thumbs. I have three reviews sorting incorrectly if an example is needed.

OK, a little more info, if a review is edited after being thumbed it's sorting gets set as no thumbs.

That makes sense to me. After all, the thumbs were given for the original review, not the edited version.

That being said, it is confusing to see the thumbs when they don't really count.

Well cD did say, he Assumes most editing of a review should improve it so that is why he leaves the thumbs in the first place.

Hmm, in that case, maybe someone with a different assumption was mucking around in the code. Wouldn't be the first time that Tim and cD didn't see things from the same angle. :)

I'm so glad someone else noticed this! I found this group specifically to report this bug. That is, I assume it's a bug. I know my edits are only for minor things, usually grammatical errors; it seems a bit ridiculous that one can't fix those without losing one's thumbs. By the way, I do not think the thumbs go uncounted if one edits the review on the "Edit Book" page.

Recently I've noticed that if I sort by votes, reviews with flags come above reviews with neither thumbs nor flags. Which isn't how it used to be, and doesn't seem quite right.

Example: http://www.librarything.com/work/47291/reviews/

7>

Looks fine to me (the top review when sorting by votes is one with a mind-boggling 17 thumbs-up); are you sure you aren't accidentally clicking twice on the "sort by votes", and ending up sorting inversely? If you're sure that isn't the problem, it's probably browser-specific.

8, what 7 was saying was that of the reviews with no thumbs at all, the one that had also been flagged was the highest of the bunch - currently #4 when sorted by votes, right under the three that have received thumbs up.

ah...here's the thread. Thanks readafew for pointing me here. I've noticed this as well and think it's very strange that a revised review doesn't sort with the thumbed reviews even though it still has one. If revising assumes the thumb is no longer pertinent, it should kill it altogether.

I find this really confusing. If I happen to spot a spelling mistake in one of my reviews, I want to correct it, without either losing the thumbs or the ability to sort by them.

Actually I just edited a whole load of my reviewed books to add a "reviewed" tag (didn't touch the review, just the tags), and now all of them have vanished from the sorted list.

I agree with the two previous posters. Now, I'm rather new to this group, so what do we do? Wait for the developers to do something about it? Or contact them directly?

since it isn't mission critical we usually wait and bump this thread now and then.

Thanks, readafew. :)

I'm glad I found this thread... this has been driving me crazy! Any word on a solution?

I think that revising a review should discount the thumbs (because the review is changed from the one that someone thumbed) but agree with you all that the thumbs just shouldn't show at all then.

Also, there is no reason they should sort below flagged reviews.

16> So, people can't fix spelling errors in their reviews without giving up their thumbs? That seems kind of dumb. I've edited almost all my reviews for simple, easily-fixed errors, and now none of the thumbs are counted. If you were in a similar situation, you would be miffed too.

If it's a good review it will get more thumbs. The alternative is to allow people to completely rewrite their review and still keep the thumbs. Or should LT allow you to change up to 10% of the review without losing its thumbs?

How often do you think people are going to entirely re-write a thumbed review and replace it with something worse?

I'd wager that that sort of thing is exceedingly rare compared to the number of edits that get made to correct spelling, commas, minor word choice, etc. I think the thumbs should stick.

I agree. I've never changed a review of mine wholesale. It's just been phrase, grammar and spelling changes. Sometimes I add material because I've reread the book. All for the better IMO. Negating thumbs up for the sake of improvements seems counter-intuitive to me.

I'm glad the thumbs are still there, but wish they'd show up on the sort. When I read others' reviews from that page, I like to see which folks found the best and it would be a shame to miss some because the writer made them better.

"If it's a good review it will get more thumbs."

^^ That's not necessarily true. Reviews that get the most viewings are, I believe, those that are most recent and those with the most thumbs. (I know mine rarely get thumbs after the first few days.) If an old review loses its thumbs, it will get buried in the stack and likely never seen again. Plus, the people who originally thumbed it won't have any idea that there vote does not "count."

Looks like I'm experiencing this bug, too; though I'm not sure if it's because the review in question was edited. (If so, only in a minor way.)

When I sort my reviews, Harry Potter & the Deathly Hallows (with 5 thumbs up) doesn't appear at the top, though no other review seems to have more than 2 thumbs. Instead, the HP review is far down the page, nestled amongst other reviews with no thumbs.

Argh! Not sure whether this is a bug or a misfeature, but it's definitely not the behavior I'd expect.

Since Tim and Chris seem to be on a bug fixing spree, bump.

As soon as the downrated thumbed review gets a new thumb, the old ones count again and it sorts properly.

I take back my earlier comment that one is less likely to lose one's thumbs if one uses the "edit book" page; obviously that is not the case, because I went in to simply change the additional authors (whose roles were showing up as "9" and "14", rather than "Introduction" and "Cover Art") and without touching my review, lost the thumbs for it.

Gah.

I apologise for adding to a thread that's been dormant for five years, but I've encountered this issue today and this is the only thread that came up in my search parameters. A review for the Bertrand Russell Autobiography made the Hot Review page on 15/09/14. It gained five 'thumbs up' before I realised I'd made a small mistake. I changed the word 'The' to an 'A'. When I came to check on it on 16/09/14 I'd lost those five thumbs and only had one remaining, which was presumably given after the edit. How is it that editing one, single, word is considered such a substantial change that is worthy of having lost five people's time and approval in having read it?

How is LT supposed to know what changes are substantial and which are not?

jjwilson61, presumably a staffer at LibraryThing has been, or is, aware of the fact that, dating from 2009, editing reviews does change 'thumbs up' ratings that have already been applied. I was *not* aware of this issue until it occurred to the review mentioned above, and then only discovered it by searching the talk topics to find one that related closely to my experience. If I'm not aware of a pre-existing issue; am I supposed to guess how it works and whether it is possible to resolve it, without asking questions about it? As I am not an expert on the coding of LT, perhaps someone who is knowledgeable can say why there is no possibility of differentiating between the correction of a single spelling mistake and the deletion of an entire review to be replaced by another, entirely different one?

28>

As I am not an expert on the coding of LT, perhaps someone who is knowledgeable can say why there is no possibility of differentiating between the correction of a single spelling mistake and the deletion of an entire review to be replaced by another, entirely different one?

Well, I'm not an expert on the coding of LT either, but I do know coding, and I can say that's the sort of thing that would be technically possible using one of a variety of string-distance measurements, but computationally expensive and thus not really worthwhile for the small gain. What would probably be easier would be to treat thumbs differently from flags - flags are preserved when reviews are edited to prevent spammers from easily shaking them off (it's trivial to get undeserved flags removed by posting about the issue, and usually draw a few sympathy thumbs-up in the process).

Proofread your reviews before posting them, and if you really care about thumbs post to wherever you posted that drew the thumbs-up in the first place (I find that unless reviews are specifically linked from a high-traffic thread like one of the challenge groups they very rarely get thumbs; they're mostly due to people specifically seeking out the review rather than stumbling on it organically) if you've made edits, and I'm sure you'll get them back.

>28 Belochka: The thing is, what you're asking seems simple, and if it were a human looking at the review each time, they'd be able to recognize the difference between correcting typographical errors, and fundamentally changing the meaning of a review.

This is not possible to do programmatically. There is no way that software can recognize that you've only changed something a tiny bit, and that it still says what it did before. I've corrected a review I'd written, and I've also gone back in and changed something completely. Unless someone is specifically watching for a particular review, and voting it up (or flagging it) after every change, this will just have to be one of those annoying things.

BTW, sometimes there will be a person who does indeed watch for a review to change. I've seen it. I'm not always sure why I write reviews; sometimes I think it's mostly for myself. ;-}

30>

This is not possible to do programmatically. There is no way that software can recognize that you've only changed something a tiny bit, and that it still says what it did before.

Sure there is. The most obvious ones are longest common substring (what is the longest sequence of characters that the two strings share) and edit distance, which measures the number of character-by-character changes needed to transform one string into another.

The problem is that when you're talking about something as long as a review, these are computationally intensive. It's not that it's impossible, it's that it's not worth the trouble.

>31 lorax: I should have known you'd pick on me. It's early. I'm still drinking coffee. ;-}

I might add that when I saw that you'd replied, I *knew* that it was going to be that I'd used the words "not possible" instead of very difficult. I almost thought about changing them.

BTW, sometimes words don't carry proper meaning. Reading what you'd written made me laugh. Thanks for the smiles to start out my day. I lift my cup of coffee in salute.

32>

Glad to make you smile. And it's certainly the sort of thing where it's much harder for computers than it is for people, because computers are stupid.

Lyndatrue, thank you for the answer, and also to lorax, for going further in explanation. I appreciate both of your replies for gaining a better understanding. As it stands; I do proofread and spell check reviews, but it's also easy to have the eye slip over a 'The' instead of an 'A' after a few revisions, only to pick it up after the original posting. I'm just sorry that what was a very pleasurable experience in having the only review I've ever had featured under the Hot Reviews section be somewhat dampened by the loss of a visible sign that people reading it had an appreciation/enjoyment/recognition of the review. I will, obviously, be very aware of this now for any future reviews.

From what I'm reading in this thread, it doesn't even matter if you proofread first, simply editing the book at all counts as editing the review. Which is pretty crap, imo.

35>

I've not encountered that, but it was possibly the case back in 2009 when that complaint was made.

Edited to add:

I just tested it, with a book of mine where mine was the only review and it only had one thumb (so no great loss if it did vanish). The thumb stayed.

If you edit the book (but not the review), the review will be pushed to the top of your reviews on the "About you" selection on the Dashboard. I just verified this by editing my collection of poetry by Longfellow, and changed not one word of the review (merely added something in the comment section).

Since that particular section is entitled "Your Recent Reviews" I would think that is what they ought to be, not "The Reviews from books you've recently edited" (but that's just me).

I do recognize that the recent complaint had to do with "thumbs" and not just sorting of reviews, but it's still an unexpected behavior. So it goes.