Yes, @gxia it is correct that PubPub cannot handle the submission+blind review. So we should use a separate system for that (I guess sticking with CMT is good since we have been using that for a couple of years now). However, we need to provide people with a (media-rich) paper template, and having people write directly into PubPub to start with will simplify things. Otherwize, people will need to first submit their paper in one format, and then change to another after acceptance.
So my thinking about the workflow is as follows:
Write the paper in PubPub (leave author empty to preserve anonymity like I did in the test article). People could, of course, also write drafts in MS Word or LaTeX and import into PubPub.
Submit their paper in CMT. Here they would preferably link to their anonymous PubPub document, or optionally upload a PDF if they only have text.
Everything related to the double-blind peer review will happen in CMT.
After acceptance, people update their PubPub paper and make it “camera-ready”.
Once the paper chairs approve of the final version of the paper, it is added to the public NIME paper stream on PubPub.
The peer commentary starts, and people can read and comment on papers before and during the conference.
Authors will get a final chance (1 month after conference) to make changes to their document based on feedback on PubPub.
PDF and Epub files are exported from PubPub and archived in Zenodo, etc.
Hey @yaxu did you get any further in figuring out self hosting? It doesn’t seem super easy to me and I guess we are heading towards using their provided service. But I agree that self-hosting things is desirable.
Just thinking about the word limit thing for '21 submissions. How are we going to assess the number of words in a submission coming through as a PDF? I presume we’re looking for the number of words in the text body (i.e., not references), but that’s much harder to assess than a page limit.
I guess the organisers could do some kind of rough automagic guesstimate somehow, or just use page length as a proxy? Or are there other options?
Good question. First of all, I think it is a matter of trust that authors keep within the limits. For reviewers wanting to check, I guess the best option is to copy the text of the main body and check the number of words in a text/word editor. The paper chairs should also do a check as part of the preparations of the final, accepted papers.
Just thought it might be of interest for people who usually write their NIME papers using LaTeX (for example on Overleaf for collaborative editing) and who have issues importing their .tex sources in PubPub (it isn’t working at all for me).
The .tex document can be exported to html (with tikz figures converted to images in svg format) using :
this pandoc command (assuming your latex source is main.tex) : pandoc --lua-filter tikz.lua -f latex+raw_tex main.tex -s -o main.html
Then one simply needs to open main.html in a web browser, copy-paste the content into the pubpub editor, and insert the created svg images at the correct positions.
After a couple of days to get familiar with PubPub, I have to admit that I am really enjoying it. It’s great to be able to integrate media and interactive contents (using iframe) in academic articles. In particular you can easily get nice interactive plots by creating the plot on a Google Sheet (which should never be deleted from your Gdrive), then publish it, and then bring copy the generate iframe code in pubpub (from media). (more details here https://support.magpi.com/support/solutions/articles/6000187026-embed-google-sheet-graph-and-other-external-items-into-magpi-reports-using-iframes).
However, there’s a downside… when you export he PDF version needed for the anonymous submission (as explained in the call), all interactive content (median, iframes) get replaced with a gray rectangle stating “Visit the web version of this article to view interactive content”. My work can not be properly reviewed without those contents, and replacing the 20+ plots with static screenshot is a massive overhead (to be done twice, one for the submission, then eventually for the camera ready). Printing from the browser seems to work better (it preserves a static image of the iframe), but citations does not work well and the overall page break is a bit funny.
I am wondering if this has been considered before adopting the approach pubpub + PDF submission for review. Is anyone having the same problem and found a good solution?
I think the plan is to anonymize the pubpub document (you can do this simply by removing your name as author and replace it with “anonymous”) and submit a link for review. I am sure one of the chairs can provide more detailed information. @gxia@x_x@Roger
Hi Stefano. I would say the real downside to embedded data is that the we are intending to use Pubpub for “archival publication.” External links are not easily maintained or archivable. This is more of an issue for the NIME organization than the NIME conference, and of course archival publications often include links that disappear over time (although usually this is supplemental material and not, say, essential graphs and tables). I would recommend using static graphs or self-contained animations or interactive graphs hosted on pubpub (if that is possible) to avoid external dependencies.
For reviewing, we want anonymized PDFs. Pubpub documents are not anonymous (even if you remove the author text box). This is a real shortcoming of Pubpub, and Pubpub is aware of their current lack of reviewing support. Our PDFs-for-review approach is a work-around for this shortcoming that we hope will be eliminated in the future. As Alexander says, you can substitute a link for review, but the link should go to an account created with a pseudonym or to a server that does not reveal the content creator by name or “true” email.
Yes, the archival issue is something we are very concerned about. We may not get it right the first time, but will hopefully be able to develop a new best practice together.
I think that archival and anonymized PDF for revision are pushing in two opposite directions.
For archival, we should use only self contained media (stored in PubPub), but these are not visible in the exported PDF for anonymous revision, therefore linking externally hosted media/files may work better. Perhaps you can consider including in the Submission Guide that authors should not use material hosted elsewhere (i.e. when clicking on media do not use anything listed under Apps). This wont allow authors to use the full potential of PubPub, but being able to integrate audio, video, and files for download is already a big improvement compared to images only as it was in the past. As a new user of PubPub I immediately went for the exciting stuff (unfortunately I will have to drop the fancy mouse mouse hovering functionalities in my plots).
However, I did some experiment and this is what i figured out:
when exporting the PubPub to PDF self hosted images shows (as expected). Self contained audio and video shows as an audio and video player widget. No URL is generated to access media contents hosted in PubPub. However, when editing the attached media file (black dialog window showing on top of the page) next to “Source” there is a down pointing arrow. That gives you a link for direct download, which perhaps the authors can include in the caption only for the revision version, and eventually remove it for the camera ready version (the link seems anonymous https://assets.pubpub.org/90mydxfl/11610387917640.wav). Attached files (Media->Other) works well in the exported PDF (you get a link for direct download). In this way the revision and camera ready versions are not too far from each other. Perhaps this is something you can recommend to authors, or may be there’s even something better.
Can we address the fact that Pub Pub is prone to crashing and not saving your work (perhaps when the server is overloaded?). It makes it somewhat nerve-racking to work with especially near the deadline of submission.
I haven’t experienced crashes myself, but had one incident where I didn’t see that it said “Connecting” in the upper right corner instead of “Connected”. This turned out to be because of a network issue and I lost 20 minutes of edits before I discovered the problem. Now I have made sure to check the connection and will report the problem to the PubPub team.
Great if other people could report on issues they have with the system.