Better collaboration workflow

I recently hired a professional proofreader to go through my paper. It seems I did not explain Paperpile well enough to to them and the end result was that they made changes directly on the document rather than in the Paperpile references. Of course, this is a big problem for me because if I were to use Paperpile to reformat my references I would loose all of their changes. So I spoke with them about what went wrong to try to get feedback for the developers and here are some notes to try to help someone like them in the future.

Here is what they wrote:

“Yes, I can edit the stuff in PaperPile–I just couldn’t find it easily. When I click on a cite, I get taken to the full list, and then there’s no search function and it was taking me awhile to find the first couple of cites, so I thought I’d save time by just editing within the document… Lesson learned.”

A few thoughts on this:

  1. Directly editing a PaperPile citation should result in an error message or warning.

  2. When clicking on a citation it should take the user directly to that citation in the list of citations. They shouldn’t have to search for it.

  3. Perhaps offer a way for an editor or reviewer to simply flag citations as being problematic or leave a comment rather than editing them.

Also, they made the following comments about the errors that they fixed:

Some of the bibliographic problems are not really about ‘correcting’ the bib data, but about the formatting of it–for example, city doesn’t seem to be automatically entered, as it was missing from nearly every reference. And journals had the name of the publisher added, which is not normally in a reference, and certainly not in chicago style. Just something to watch out for in the future.

I think they are wrong. This is about data. The city isn’t entered because it is often missing when the metadata is pulled from Amazon.com, and I think the publisher is added incorrectly because Google ads it there. I believe PaperPile creates correct citations if it has correct data, but the quality of the data is pretty lousy and a big problem for a large bibliography.

The end result is that I now have a lot of work to do going through and trying to figure out what edits my editor made and I have to put them back into Paperpile manually myself. I think Paperpile has put a lot of thought into all of these questions already, but I think more work needs to be done if it is to really work as part of a professional workflow, especially when someone who hasn’t been working with Paperpile for a while comes in and collaborates on a paperpile formatted document.

Thanks so much for this report. It’s really helpful to see how this work with a proofreader worked out for you and where the problems where.

A few random thoughts;

  • I think we are pretty good already. Just sharing a link of a Google document and as a proofreader you have everything to modify and edit the citations (this will get even better with our free add-on). Imagine they would have never used EndNote and you would have to set up and explain a sane workflow with Word and EndNote.
  • The linking is a bug. It should open only the references in the citation. We are aware of that and will fix that.
  • It’s technically not possible to avoid someone editing a link (it might be if she or he has the addon-enabled but even there it’s tricky to implement)
  • Regarding the data. I think it’s the right thing to save all data available and CSL should figure out how to format the citations. I’m not sure why this does not work for publishers of journals in Chicago. We have not heard any complaints but I’m not really familiar with that.
  • If you see constantly books being not imported correctly let us know. @andreas can have a look at it and maybe there is a generic fix. I’ve said this before we do lots of magic to get journal data right but our algorithm is not as sophisticated for books yet. But we don’t get many complaints. That either means it’s not that bad or users just accept that data needs to be manually improved. We don’t really want accept that except in cases where it’s really not possible.