Disambiguation Practices

[Elizabeth (Alaska)]

A discussion as broken out in the Milestones thread. It seems reasonable that maybe others interested might not see it there.

https://www.goodreads.com/topic/show/...

[Elizabeth (Alaska)]

In that thread ,Keith wrote: "...disappeared after that account came into being, probably as the result of untrackable and pointless author merges."

As a side issue is the untrackable part. Are they untrackable? Or are they just so deeply buried in the author librarian edits that our eyes glass over trying to find them?

[Suzi]

Elizabeth (Alaska) wrote: "In that thread ,Keith wrote: "...disappeared after that account came into being, probably as the result of untrackable and pointless author merges."

As a side issue is the untrackable part. Are th..."

some of them are totally untraceable after author merges.

unless you find the author merge in the librarians' edits.

[Elizabeth (Alaska)]

Suzi wrote: "unless you find the author merge in the librarians' edits."

I think this may be the issue. On the author librarian edits are included all edits pertaining to any and all of the author works, including book edits. And for GR authors, you can't even get to that page.

I know we're not supposed to post about site functionality, but it would be really helpful if there were a way to separate true author edit listings from all book edit listings on the author librarian edit page. That way, one could review what author edits have been performed.

[Keith]

Aside from providing a slap on some Librarian's wrist, I'm not sure what would be gained by improving the trace on author merges. Once the merge has occurred, with or without a track, the same research and work has to be redone to unmerge. From what I understand of how databases work (more than a layman, less than a professional DBA), there could be no "undo" function from an author merge that included book records, and no need to undo one that did not include book records.

I'm probably biased by the fact that I'm a Super, but I'd almost be happy to have author-merge permissions limited to Supers (except for the fact of the never-ending imports of authors missing a period after an initial, or including personal titles). In any case, given all the stuff that would be useful to Librarians that hasn't changed and won't change, I don't expect this to change. See also: Sisyphus.

In any case, to Librarians who stumble upon this thread, I beg you: be absolutely sure you know exactly what you're doing before pulling the trigger on a merge, and never merge a profile that has full author details into a new profile that is nothing but a name and a handful of books, even if it looks like the name field is "better." The rest of us have been working on this long enough that such a merge is almost certainly wrong.

[Elizabeth (Alaska)]

Keith wrote: "Aside from providing a slap on some Librarian's wrist, I'm not sure what would be gained by improving the trace on author merges.."

Is it just merges that would shown in such a thread? I think not. It Wouldn't it also be author name changes that don't provide a merge? Or details that get added whether or not correctly? And yes, sometimes it *is* to provide a slap on the wrist. The purpose, as I understand it, of these librarian edit pages is to provide an audit trail. GR staff needs to be able to ensure the integrity of the database.

[Keith]

I'm not 100% sure, but when there hasn't been a merge, the Librarian Edit log continues to exist. I think (but cannot at this moment prove) that log would show such edits. But once author records have merged, a lot goes away, which makes it impossible to tell who did what. I'll admit it would be handy to at least be able to tell that who "hey, quit it." But the effort required to get there probably isn't justifiable at a business level.

[Elizabeth (Alaska)]

Keith, we're not talking about the same thing, perhaps. I'm talking about the librarian edit log on the author profile. It contains every edit ever associated with any book by the author as well as edits to the author profile itself. It is mostly not useful for seeing what happens to the profile itself. I don't know where disambiguations should rightfully appear (since that is the topic of this thread).

I once tried to find who was changing the name of a Japanese author. The name appears as family name/given name on covers, but someone kept changing it to the western way (kept changing, as in at least 3 times after I corrected it). I decided to see who was changing it so that I could flag and report. I had to look through more than 20 pages of edits.

[Elizabeth (Alaska)]

The author merge shows on the book librarian edits. I did one yesterday and looked. But of course it gives only author ID's and, as you say, the author that was merged is empty, so the script deletes it. It also appears in the author profile log.