Last call: Work-combination improvements

Okay, instead of moving on to author combinations, I'd like to ask about further improvements to work combination.

This is kind of your last change. The code is all fresh and plastic in my mind now. It won't be in a week or two, and getting back into it will be onerous. I won't do it, barring bugs, for some time.

So, what changes do you want to see in how work—only work—combination works?

I was thinking we need a better way to combine arbitrary stuff. For example, there's currently no good way—short of a hack—to combine works you find here and there, especially works without an author.

My thought is that there should some sort of "remember this" link in the combination area. It would make a little pocketbook of works. Then you could go to the combination page, and they'd be right there, to check or not.

Anyway, that's my idea.

The biggest problem with combining works is that people don't want to ask for help doing it, so they split the author, alias the work into another author, then combine. This leaves that empty link that I have long despised, which you think is fine/dandy hunky/dory. I continue to fix them at every opportunity. :p

Okay, that's a hack, though. So if there were a way to collect works and do it appropriately, that would solve that, right?

omargosh wants a warning when combining works within? between? both? series.

Okay, I can do that.

I suspect this is a bit out of scope: Would it be possible to "never" proposed work combinations? (As with nevering author combinations, it just removes it from the list of suggestions. It should still be possible from the author page.)

There are quite a few userscripts on greasyfork.org that just streamline the UI, but don't actually change the underlying process:

https://greasyfork.org/en/scripts/11603-librarything-work-editions-grid
https://greasyfork.org/en/scripts/18437-librarything-better-combine-works-button
https://greasyfork.org/en/scripts/18457-librarything-sort-and-re-link-on-combine...
https://greasyfork.org/en/scripts/20060-librarything-ignore-articles-punctuation...

I think there's some overlap in those, but the big things for me are:
- Ignore articles when listing works on the combine/separate pages. It's so helpful to see "The Lord of the Rings" next to "Lord of the Rings". The userscript sorts after the page has loaded; it would be faster if you could do it before sending the page, perhaps using the most popular non-zero sort character.
- Make the separation link middle/ctrl-clickable (so it can be opened in a new window). When there are a lot of editions to separate off of a work (after an ill-advised combination, for example) it's much easier if you don't have to wait for the page to reload after separating each and every edition.
- The second one up there makes a floating "combine works" window that lets you make many combinations without reloading the initial combine/separate page. It opens the usual confirmation page in a popup window, so none of the checks are bypassed, and disables the works that have been combined until the page is reloaded, so you can't accidentally try to do another combination on a work that's been combined in to something else.

Like the pocketbook idea. Also like the idea of ignoring articles, as in >6 r.orrison:. (However, I think nevering is an abomination.)

What I'd like is a way of toggling (show/hide) the subsidiary editions within the works on the combination page.

What I'd like is a way of toggling (show/hide) the subsidiary editions within the works on the combination page.

For ease of sorting through options?

I think that, if you check one, it needs to expand, so you are looking at it.

>10 timspalding: Cool. I think you need a 'show/hide all' but also the option to show/hide each individual entry. I'm used to using the term 'twisty', with a right-pointing triangle ► for hidden and down-pointing triangle ▼ for shown.

I've gotten things to work mostly how I want to with the userscripts, and would be happy to see any of those implemented by default on the site.

One idea that isn't fixable with userscripts is to have an option to not display works on an author's combine/separate page where they are only an "Other Author" on that work. This makes the page load very slowly for somebody who's happened to write an introduction on a very popular work.

>6 r.orrison:

Would it be possible to "never" proposed work combinations?

Oh, god, that would be wonderful! I am so tired of seeing zillions of unrelated titles on the editions page of a work, just because some idiot publisher decided that it would be a good idea to use the same ISBN for every work in his "Susie Q retells every kiddie story you ever heard of" series.

>12 omargosh:

have an option to not display works on an author's combine/separate page where they are only an "Other Author" on that work.

Yes! Good one!

>12 omargosh: and >13 lilithcat: I was just about to propose that too. I work with a lot of illustrators pages, and if they've done all the classics the combination page never ends.

I'd sure like that editions toggle too!

>6 r.orrison: Also throwing my hat in the ring for a 'never', please!

Would it be possible to "never" proposed work combinations?

Well, yes and no. It would be possible, but as works change numbers, the nevers would expire. I'm not sure people would understand that.

I don't know if this would be helpful to many people, but I often look for possible combinations from the Author's page instead of the combination page because it's smaller. If I see singletons, I go to editions for just that one and try. What would improve this process would be making it possible to sort works alphabetically as well as by number of copies on the Author page.

>8 Noisy: et al. Yes! I despise scrolling down long pages and Ctrl+F is nearly useless when searching work combination pages because of all the editions.

>10 timspalding: I'm okay with it expanding if I click on it. That's probably ideal, actually.

Posting from my phone and after a long day at work, so I apologize in advance for any mistakes!

I don't use any scripts, so my views are based solely on LT behavior.

If works have a disambiguation notice as well as a canonical title (maybe with parenthesis et. al.?), make them not combinable with each other. Create a thread like the "Sorry, too much love" one to discuss.

Ditto for work-to-work relationships.

Ditto for the suggestion how to separate works more comfortably and quickly! Being able to separate more than one edition at once would be awesome! I admit I have passed on some works needing separation simply because I feared the "rabbit hole".

On "Other authors": yes, please make them expandable, i.e. don't show right away! Work combination pages are long as they are, why have all those huge entries only because someone wrote a preface to a Jane Austen or Charles Dickens work?
In that vein: being able to sort an individual work entry alphabetically would be great! Most popular works have way too many other popular works on the "potential work combination" page to be able to navigate well.
By the way, I'm getting a 504 error quite often these days, especially when working on popular works!

I'd love to see another round of deleting zero-editions.

Thanks so much for taking suggestions, I can't wait to see the outcome!

>19 leselotte: Every time one of the works in the combination is 1000 or more copies (or thereabouts), I will get a 504 the first time I try and it will go through on the second attempt.

Other from that:
- One thing that makes it hard on the big authors is that "Name" and "The Name" cannot be sorted next to each other. So I am usually forced to have a second list open on the "The" titles so I can see if I can find matches.
- As >19 leselotte: Allow to see the list only of the author, not all the works they are an other author (try to work on a SF author that has stories in hundreds of anthologies).

When separating editions out, they could automatically go into that notebook you're talking about, making it easier to recombine them with eachother once you're done.

I've added an option to "abbreviate" the combine-works display. When abbreviated, you can't see all the editions, unless you click to include them.



I'm hoping this makes it easier, because extra options for sorting with or without articles is fiddly, I think.

Every time one of the works in the combination is 1000 or more copies (or thereabouts), I will get a 504 the first time I try and it will go through on the second attempt.

Can you save me one next time you find one? I want to benchmark all the various parts.

>23 timspalding:

Sure - next time I see the 504, I will just back out and not make a second attempt so you can see it.

BTW Tim, are you fasttracking things again, around new years eve, when people are busy with shopping and doing stuff? Hope you give us a bit more time to discuss and think about improvements..

I've added series notices:

A couple of ideas:

1. Replace the individual (separate) links on the Combine works page with (▢ move) check boxes allowing the identification of a subset of entries from a given work to be separated and combined in one operation. Then, on the Combine works: Combine the following works? page, the selected individual entries would be listed under their current work as candidates for separation/combination with the other selected works. If no other work had been selected, the sub-set would be separated and combined into one new work.

2. Add check boxes to the individual items on the Combine works: Combine the following works? page which would allow individual items to be excluded from the current combining operation without having to go back to the Combine works page to redo the entire selection just to eliminate an item.

>2 timspalding:, > 5

I'm liking that remembering pocketbook!!

Okay, I made a very minimal implementation of what I'm calling "The Workbench."

You can see it on work pages:


Adding books takes you to this:


So, basically, you can rack up works you're interested in, and then move to the combine page. It doesn't auto-combine them--you have to click the ones that you want to combine.

So far it's minimal—just works, only available from work pages, and so forth. But it may be useful in work combining.

>28 timspalding: That I love!

Userscript interference?

at 11:30 pm (pst) I was able to edd and save book records. Now I can only add a record. Will not save any edits.Can't access chose all for collections. Date map is missing on acquired date. Hope these specifis help

Work with it a while and tell me what it needs. It was intentionally a "minimum viable product."

If works have a disambiguation notice as well as a canonical title (maybe with parenthesis et. al.?), make them not combinable with each other. Create a thread like the "Sorry, too much love" one to discuss.

I think the idea is if BOTH have disambiguations and canonical titles? Or just one? What if both have disambiguations, but not canonical titles?

Ditto for work-to-work relationships.

That was done.

Ditto for the suggestion how to separate works more comfortably and quickly! Being able to separate more than one edition at once would be awesome! I admit I have passed on some works needing separation simply because I feared the "rabbit hole".

Okay, would you separate them INTO the same work--that is would they be separated "together"?

On "Other authors": yes, please make them expandable, i.e. don't show right away! Work combination pages are long as they are, why have all those huge entries only because someone wrote a preface to a Jane Austen or Charles Dickens work?

Meh. I think all the detail should be shown the first time, but be hide-able. If we make the hidden mode the default, we'll get some very bad combinations.

In that vein: being able to sort an individual work entry alphabetically would be great! Most popular works have way too many other popular works on the "potential work combination" page to be able to navigate well.

I'm confused--it sorts them alphabetically now, doesn't it?

- One thing that makes it hard on the big authors is that "Name" and "The Name" cannot be sorted next to each other. So I am usually forced to have a second list open on the "The" titles so I can see if I can find matches.

So, my feeling is that this has been partially solved by a hide-editions mode. It doesn't solve the problem by redordering entries, but it kills 75% of the clutter that made it hard in the first place.

Allow to see the list only of the author, not all the works they are an other author (try to work on a SF author that has stories in hundreds of anthologies).

Okay, I'll look at this.

When separating editions out, they could automatically go into that notebook you're talking about, making it easier to recombine them with eachother once you're done.

Not automatically, but I'll look at making it easy to add them, from the separation page.

BTW Tim, are you fasttracking things again, around new years eve, when people are busy with shopping and doing stuff? Hope you give us a bit more time to discuss and think about improvements..

Door's still open.

1. Replace the individual (separate) links on the Combine works page with (▢ move) check boxes allowing the identification of a subset of entries from a given work to be separated and combined in one operation. Then, on the Combine works: Combine the following works? page, the selected individual entries would be listed under their current work as candidates for separation/combination with the other selected works. If no other work had been selected, the sub-set would be separated and combined into one new work.

I dunno--it seems to me a lot of damage could be done that way. Having individual "separates" strikes me as a check on abuse.

Add check boxes to the individual items on the Combine works: Combine the following works? page which would allow individual items to be excluded from the current combining operation without having to go back to the Combine works page to redo the entire selection just to eliminate an item.

No. Functionality should live one place. If the problem is the complexity of the page, then let's deal with that.

I would definitely find useful a feature allowing the separation of more than one edition at once into a single separate work. This would be very useful when fixing an incorrect combination of two works with several editions each. As far as I know the only way to do that now is to separate out one by one the several editions of one of the works, and then combine those several separated ones back together. This can be laborious when there are 10-20 editions, especially when there are no-copy editions, and incorrect or no author for some editions. I do get what you say, Tim, about it being a potential for abuse to make broader separation easier, but it also makes good work easier.

On the use of expandable/hideable listings for works on the work combination page, I think that would be useful as well (hide the big long list for major works that are unrelated to the cleanup work you're doing) but I agree with you, Tim, that the default should be showing the full list to make sure people know what they're looking at.

I'm not 100% sure what feature leselotte is talking about in terms of "being able to sort an individual work entry alphabetically", but it sounds like the "Editions" page for an individual work. There the editions of that work are ordered by number of copies; having a option to switch to alphabetical as in a Combine page might be useful, it's easier that way to spot incorrectly combined editions. Also the Potential Work Combinations at the bottom of the Editions page don't seem to be ordered either alphabetically or by number of copies; alphabetical might be useful there.

Following on to myself, I haven't tried using the new "Workbench" feature but that would make somewhat easier the separation-and-merging of several editions. Not as easy as separating them as a group, but it should streamline things some.

>39 scott_beeler: " There the editions of that work are ordered by number of copies; having a option to switch to alphabetical as in a Combine page might be useful, it's easier that way to spot incorrectly combined editions. Also the Potential Work Combinations at the bottom of the Editions page don't seem to be ordered either alphabetically or by number of copies; alphabetical might be useful there."

Yes and yes.

There doesn't seem to be any rhyme or reason behind the way editions with the same number of copies are sorted. It often seems when I go back to the editions page to separate more copies, they are in a different order. For example, I separate an edition and mentally tell myself, ok, that was a single copy that was right after the ISBN xxxxxxxxxxx copy, so I shouldn't have to go through the whole entire list again, just go right to that last singleton and proceed down the list of editions from there. But...it doesn't seem to work that way.

I've added the ability to do work-combination with just primary authors, e.g.,

>42 timspalding:

Thanks!

>42 timspalding: Thank you! These changes make everything so much easier!

Just wanted to say that even though I know the other way and have used it for years, I'm digging the workbench and feel a lot safe using it. Thanks for all your recent work related to combining, Tim. It's a great Christmas present to all us Combiners!

>42 timspalding: This (and other recent) feature is great, thanks so much. However, I discovered this evening that when I combined from the primary only page, i was returned to the page with primary and secondary. Can we please be returned to the primary only page?

And I hope I'm not too late to make this request . . . Could the alphabetical sort order on the combine/separate page be adjusted to ignore punctuation?

For an example of why, see the Arthur Machen combine/separate page, specifically the two separate "Arthur Machen's 1890s Notebook" that should be combined.

>38 timspalding: Having single separations does take more time but I always accepted this as a good check & balance to try and reduce overall erroneous separations. I know it won't stop a user on a mission but I think it's a good limiter nonetheless.

>23 timspalding: Here's a 504 from a separation of a title from a large work:

https://www.librarything.com/work_separate.php?book=2538251&work=2770499

I love all these tweaks - this has made the process so much better. One thing I'd find useful is a link to the workbench from the combination page so I can clear it before moving on ... but that might just be me and it in no way detracts from the fact that I love all of the changes so far.

Edited to clarify
I mean the regular Combine Works page you land on after making the combination, not the Combine Works from Workbench page, which has a link.

>48 klarusu:

However, having separations one-at-a time means it's far, far harder to undo erroneous combinations. I have wanted a checkbox-to-separate approach for many years.

>51 lorax: Yeah, it's also easier to lose track of what you're doing when you're cleaning up something that needs multiples. I just assumed that it works as a stop gap but it may not actually do any good at all. I suppose you could argue that the same logic applies to combinations - you can do just as much damage by multiple combinations, I guess. Maybe it's time to just suck it and see - it's slower and it might not be doing what it's given credit for.

However, having separations one-at-a time means it's far, far harder to undo erroneous combinations. I have wanted a checkbox-to-separate approach for many years.

Do you fear it might also make it far easier to make erroneous ones? Maybe a compromise, like the user-script?

>52 klarusu:

The same logic applies to combination but even more so. It's less of a problem now with the 200-copy limit than it once was - old-timers still have scars from things like "the time someone combined together every Terry Pratchett work together" - but since a work contains many editions, even combining only two works could combine together a lot of editions that need to be individually separated out.

>53 timspalding:

Erroneous separations can be undone in seconds, since combining is so much easier. I'm not nearly as concerned about that, but would be happy with a compromise since it's better than the nothing we have now.

I just got the warning that my combination crosses authors. Great to see it!

(Yes, I kept on. One of the works had two authors in the author field.)

Just being able to Ctrl+Click (or middle click) to do a separation in a new tab would be very helpful. From there, it's just a couple more clicks to the newly separated work and then add to workbench.

Perhaps on the post-separation page the two new work links could have an (add to workbench) link next to them, to save having to load the new work page?

Okay, proposal:

1. When you click to separate works, it opens a separation lightbox that lists the edition and gives you "separate now" and "add more editions"? If you click the latter, you can add another.

2. If there are two or more editions, the lightbox gives you two options--separate into separate works, or separate and combine into one work.

How about that?

>58 timspalding: Wow! Yes, please!

>58 timspalding: Oh, please, please, please!

>58 timspalding: Nice!

>58 timspalding:

Yes please!

>58 timspalding: Yes, please!

I've added a "Clear workbench" option.

>64 timspalding: Brilliant!

Did the addition of "Clear workbench" disable the ability to remove individual works from the workbench? Was that on purpose?

I think I've just (well, yesterday) seen again that undesirable behavior when combining 2 works from a split author (1 from a split, 1 still in the 'unknown' pool) pours the result back in the unknown pool despite the already attributed version having more copies.

Any way the attribution could stick, especially when the attributed work has the most copies ?

>67 Jarandel:

The combination screen tells you where it will dump them in the case of a cross-split author combinations - did it tell you they are going to the correct split or did it tell you it will go into unknown?

>68 AnnieMod: I'm fairly sure it told me it would go with the first work, which was the attributed one and with the most copies, but that wasn't the result.

>54 lorax: "the time someone combined together every Terry Pratchett work"
Please tell me that that never happened.

>70 Petroglyph: There is a reason the 200 copies rule was added :) And we had a mishap with Shakespeare once (I do not remember if it was all or just about half but it still was "fun" to untangle).

>71 AnnieMod:
*shudder*

I'm getting the same bug as >66 scott_beeler: where I can't delete individual items from the workbench, only clear the entire thing out.

Also, I'm not crazy about the fact that the series notice comes up whenever one of the works has a series. My original request was just a warning when the two works were part of the same series. I could see how it would be useful to have a warning also when two works were part of different series. My concern with displaying a warning whenever there is any one series involved is that it will train me to start ignoring that message, since it's "crying wolf" in a lot of cases. Though perhaps I'm not seeing things from the right perspective.

Fixed the delete-link not working. Thanks.

>72 Petroglyph: Of course, the 200 copies thing does not kick in when someone combines a 270 copies work and a 161 copies one

I just spent an hour untangling http://www.librarything.com/work/1186034 and http://www.librarything.com/work/18899419 .

So yeah - any plans to change something in the separation? That took forever :) The workbench was very useful at getting the copies together after the separation!

>70 Petroglyph:

Oh, it did. It's one of the few times that Tim deployed his internal "separate ALL the things" script. (I can't locate the thread in Combiners! about it, just references to it in 2008 as an "old war story".)

For a while I lived in fear of the day that someone would combine all of Shakespeare because they had a Complete Works, but that never happened, and given the 200-copy limit it's not likely to ever happen now.

Is it too late to suggest something?

A minor irritant is that if you're on an author's combination page, you sort by work copies, you make a combination, and then when you're sent back to the page, it's now sorted by title again.

Turkey.

Now now, there's no need for name calling.

>81 bernsad:

Oh, stuff it. ;-)

>81 bernsad: :-D

>41 PhaedraB: "There doesn't seem to be any rhyme or reason behind the way editions with the same number of copies are sorted. It often seems when I go back to the editions page to separate more copies, they are in a different order."

It may have changed with recent modifications, or now that I've found and added scripts to Firefox on my laptop, but - for the longest time - "editions with the same number of copies" seemed to reverse their ordering with each successive display. So, if you "separate an edition and mentally tell {your}self, ok, that was a single copy that was right after the ISBN xxxxxxxxxxx copy, so {you} shouldn't have to go through the whole entire list again, just go right to that last singleton and proceed UP {not down} the list of editions from there." after the second resorting, it'd be the next one down again ... then up ... then down ... etc. ...

>71 AnnieMod: As I recalled, it was a bunch of C.S. Lewis' Works. Maybe we can't "never combine" Works, but "never again" has a certain appeal.