I have been playing around with PubPub lately (see this example document), and see that it has many of the functions that we want in the community:
Media-rich
Multi-file download
Built-in citation support
…
I originally thought that it would not be possible to make an anonymous submission. However, I see that it is possible to put in a fake name, and it will not link to my profile. So that means that it could also be used for original blind submissions.
It also has support for both inline and end-of-document comments. So it could also serve as a solution for peer commentary as discussed in another thread here on the forum.
My main reservation is how we would handle the longterm archiving. I guess a PDF + media files would be what could be archived on Zenodo?
What do you think? Could this (finally) be a way of moving to a new template/format?
That looks quite good actually! I noticed you can download the document in various formats. Did you edit your example document in latex, md, or word? If PubPub is capable of generating the latex source regardlessly, I guess it would be even better to archive the source rather than the PDF. That way you could compile the next NIME retrospective book straight from Zenodo
This document I just wrote straight into the editor. I tested importing a fairly simple LaTeX document, and that worked well. However, it choked when I tried to import a previous NIME paper. I guess it had problems with all the weird template functions.
Would be great if someone else could try to import and test.
Being able to work in LaTeX would be a big plus in my opinion.
I like the idea of having rich media incorporated into the papers. The biggest reservation I have about any non-traditional format is that it makes NIME papers stand apart from other venues in a way that communities outside our own might not recognise. There is something to be said for a PDF that looks like what computer scientists expect conference papers to look like.
Would there be a way to keep the current NIME PDF template alongside a media-friendly HTML version?
You can import LaTeX into PubPub, and also add LaTeX blocks inline. It is also possible to export a PDF, but we don’t have much control over how it looks.
Then Authorea would give us more options. We tested that last year (see test document), and it imports our old ACM template. It can also export a two-column (non-ACM-looking) PDF.
I also like a LaTeX approach, but around 50% of NIME authors prefer the MS Word template. And the looks of the two templates (LaTeX and Word) are quite different. That is why I think a WYSIWYG-approach like PubPub (or Authorea) could be the best way forward.
Looks interesting. I’m wondering about scraping/indexing. Google scholar scrapes sites like arXiv but not Zenodo or Figshare. By contrast, the latter two can assign DOIs, whereas arXiv cannot. What is the situation with PubPub?
I made a comment about pubpub here (perhaps it should be moved here). I’d just like to add that in my experience, and according to this github issue, google scholar does not index zenodo. Very annoying…
Thanks @NiccoloGranieri. In fact, it turns out that it is, indeed, possible to make anonymous submissions with PubPub. See this test article that I also mentioned in the PubPub thread.
What is nice about OSF, though, is that it allows for storing all sorts of other stuff together with papers (since it is a bucket-based system), complete with DOIs. At the moment, it seems like we cannot have it all. But to summarize:
PubPub could give us a strict and multimedia-rich template with very nice reviewing/commenting features.
OSF could give us a solid, bucket-based solution, with DOIs, but would not solve the template problem.
So I am leaning towards PubPub > Zenodo being the overall best solution.
I rather like the current NIME PDF template. But I can absolutely see the value for media-rich, modern open-source formats, and PubPub might be a good choice for a NIME journal. Regardless, I think any new format for the conference should be rolled out carefully to avoid disruption.
Some questions to consider:
Will some NIME authors be confused by the PubPub authoring system on first encounter? Will it disrupt established workflows for early drafts and in-house peer review? Many papers are written at the last minute. Will we need a deadline extension?
Will the PubPub reviewing system become the NIME conference reviewing system? If yes, does it have all the features we need/expect of NIME reviews? If no, how do they relate to one another?
(How) will the resulting articles be indexed by Google Scholar?
What is the expected way in which authors self-archive papers from PubPub? Download and archive a copy of the PDF?
Will the results look professional and consistent enough that they are recognised as high-quality publications by academics outside NIME?
I expect all of the above can be addressed. There may also be other issues we wouldn’t think of until we try to deploy it in a production environment. I wonder if we might have a dual-format option for NIME 2021, so authors could choose to use the old template or the new tools. With appropriate volunteer time we could convert LaTeX/Word papers to PubPub after the conference if things work well.
Thanks for your good questions @andrew. I will try to answer some of them:
We will have to communicate the shift very clearly, I think. But working with PubPub is fairly intuitive and it is also possible to import (parts of) documents from MSWord and LaTeX. So I would not expect this to be a major issue. We would need to have a “support team” that is available to help with submissions, though.
I think that indexing is a separate issue. We would still keep the archive on nime.org. From there we can point to the HTML version at PubPub and to a locally stored PDF+zip for Google Scholar to scrape (or not), and we can archive everything in Zenodo as we do today. I have received confirmation that NIME will be indexed in Web of Science, we are in DBLP, and hopefully soon we will be back in the ACM DL. So moving to PubPub would not really change anything, but could possibly help us with better indexing if PubPub has/makes some indexing function.
That is a good question and something to discuss a little more. As far as I can see, PubPub is more geared toward open review. So I don’t think PubPub could take over as a general-purpose conference management system (as of now). So we may want to keep a separate system for the submissions (which would just be pointers to PubPub) and for handling the initial (double-blind) reviewing process. Once papers are accepted and “camera-ready” versions are prepared, they can be added to the public NIME PubPub channel and the open peer commentary can start.
Yes, I would think that either PDF+zip of media/data is one option, or an EPUB file that contains both HTML+media. The latter would preserve the content best, but the former is probably the best overall?
That is a good question. Some random thoughts here:
As other conferences and journals are also moving toward new formats, I am not sure that the idea of “paper as PDF” is going to stay forever anyways. Many journals are HTML-based these days and with the possibility of downloading a PDF for archival purposes. I guess conferences will move toward this practice as well, once the tools are available.
Hopefully, people will judge content instead of form.
As of now the PDF output of PubPub is only single-column (see the PDF of the test document). But I guess it may be possible to talk to the PubPub people about whether it is possible to customize the PDF output. Authorea, for example, has a two-column look that may look more academic…?
I don’t really think that our current templates (the LaTeX and MS Word are quite different if you look at the details) look that nice… They look old-school at least, which may be positive in some contexts. The biggest problem, however, is that people use them in all sorts of ways. Browsing through a few hundred of the papers shows how many variations people are able to come up with when it comes to fonts, indents, spacing, figure placemenets, and bibliography formatting… Even though the paper chairs do a tremendous job in improving the final output, the consistency is not particularly high. Finally, the biggest problem may the lack of multimedia support. If we want that to happen, we will need to move toward an HTML-based system with a WYSIWYG editor, I think.
Authors might be more comfortable with investing in it, if it seems to have momentum. If NIME shares their plan early, helping other academic workshops and conferences to embrace it, that could help. I’d be happy to evangelise to the iclc (live coding) committee, for example.
I think it’s worth looking into creating a self-hosted installation of pubpub. However when I last looked a couple of months ago, it seemed to be open source in theory but not really in practice – in that self-hosting would involve setting up cloud services in a certain way that is not documented.
I’d be skeptical about the extent to which it really lets you author in LaTeX or Word. If it does, I think this is probably a one-way street. Personally, I think the PDF culture around word and LaTeX should really be abandoned, if a better replacement is available. I guess better = less time consuming, good support for embedded audio and video, readable on (i.e responsive to) all modern devices…
If a big advantage of pubpub is open peer review, maybe NIME could consider adding a stage of open peer review after a blinded one, which could be active during the conference allowing authors to make corrections before final publication of proceedings. Therefore a conference could have ‘preceedings’ shared for open peer review before and during the conference, and ‘proceedings’ published afterwards.
This scan shows 3 cookies from Alphabet Inc on Nime.org - probably because there is a YouTube embed on the front page Probably should remove that and change to a self hosted video.
The website also uses google fonts, so we should self host these as well.
@andrew reminded us of an important issue. It seems that pubpub cannot yet handle submission&review management (paper bidding, topic classification, double-blind review, score, etc. Please correct me if I am wrong). If a separate system for the submissions is needed, would submitting the work twice on two different platforms with different formats be a burden for authors?
Your suggestion of using a traditional platform (such as CMT) and pubpub separately for different phases of review is very promising. It may work better with a little modification:
submission and double-blind review are done on a traditional platform “as it is”.
once a paper is accepted, we require a camera-ready version on pubpub.
In “pub settings”, the authors can submit a PDF with custom styling alongside their pub. This can solve the format concern. When readers download the pdf, the can choose to download the styled one or an automated exported one. The styled pdf can even be optional. If the authors want to control the formatting, they are suggested to create their own pdf; otherwise, the system would do it for them.
Not sure how to move posts here in the forum. 
Probably should remove that and change to a self hosted video.