citability

[proohit]

Dear authors, thanks for your work on evolving WebAssembly. I am working on my thesis right now, which is all about WebAssembly. I wanted to cite this proposal because it solves problems I encountered during my testings.

Is there an official way to cite this and possibly other proposals? Thanks in advance.

[lukewagner]

That's a good question, thanks for asking. For my part, I don't know what an official convention would be other than simply including the GitHub URL of this repo (which should stay valid).

[tschneidereit]

@linclark worked through this for the WASI repo in webassembly/WASI#356. Perhaps the solution used there can be applied here, too?

[proohit]

Thanks for your pointer! I think this is a good way to persist the repository, since it represents a specific commit/state as well. I'm just wondering if this is the solution for all proposals as they may be cited at any time. My automation-heart whispers CI in my ear :D. I guess at any given point, the repos can also be DOI'ed by forking and uploading to zenodo. On the other hand, forking may violate license rights.

(In case anyone's wondering, the DOI from https://github.com/WebAssembly/WASI/issues/356 can be found here: https://zenodo.org/record/4323447)

[RossTate]

Another thing to consider is what should be used as the "author(s)" of the citation. Based on what other communally developed languages do, I would suggest "The WebAssembly Community Group".

[rossberg]

Another thing to consider is what should be used as the "author(s)" of the citation. Based on what other communally developed languages do, I would suggest "The WebAssembly Community Group".

I disagree. Citing the proposal authors' names is good practice, see e.g. C++ proposals. There is a difference between group decisions and individual authorship. In most cases, the authorship and most of the actual work involved in a proposal is due to few specific people, typically the (co-)champions, sometimes with additional collaborators.

[RossTate]

I'm not sure which C++ proposal you're referring to, but sampling through a few I find that their scale is so so so much smaller than the scale of a wasm proposal. (Not to mention that Wikipedia is not particularly great about authors in citations.) And while the champions are conceptually the project lead, "authors" of a work are generally the project contributors, not just leads, and where contributions include not just to the text but also to artifacts that inform, validate, and support the text (e.g. software and experiments). The DOI provided for WASI above lists over 20 authors, which is in line with what is common in research communities where projects are similarly large scale.

To give a non-wasm point of comparison, citing just the champions of a proposal would be like citing just Gilad Bracha for JSR 14, the proposal that added generics to Java. There was a whole paper that led to the first major draft of that design. Then the community tried to put that draft to practice and found it far too limited to be viable, identifying important use cases it failed to satisfy. So then a whole second paper was written to articulate and motivate the second major draft of the design. That then was further iterated by the community until arriving at something that was practically viable and then shipped with Java 5. While Gilad was certainly a huge part in that happening, citing him as sole author would seem to me to underappreciate the many many others involved and their contributions. The convention I have seen in this case is that the citations of the two research papers list their specific authors but citation of the proposal lists the community in some way. That some way was often just Sun Microsystems, as Java is a proprietary language, but other languages have adopted different norms.

So, for citations of a proposal as a whole artifact, I would be in favor of either a broad term, as I suggested above, or a long author list that recognizes the many contributions individuals in the CG make to each proposal, as the DOI above does.

[proohit]

I think it's valid to include all participants, i.e. all authors that contributed to the proposal up to now (the time of citing). Maybe in an order similar to how GitHub sorts them to emphasize the leaders of the proposal.

On the other hand, I think it's not common to include all participants. Given I'm writing my thesis, then I myself cite people that supported me to do so. However, if anyone was to cite my thesis, those supporters of my thesis would not be mentioned as authors, even though they contributed.

[RossTate]

Heh, yeah, theses are especially weird in that way.

[proohit]

I am no friend of applying pressure, yet here I am bumping, because my deadline is getting closer.

I thought about some more challenges regarding Zenodo. Zenodo works with WebHooks and whenever there is a new release, a new DOI is created for that release. That leads to the requirement of always keeping the releases up to date with the current state. That sounds very cumbersome for you as the authors if you ask me. Also, if I want to reference more proposals like gc, each one would need to release their states.

Maybe, I will fallback to citing GitHub as an internet resource for now.

[rossberg]

I think citing a GitHub repo is fine. We don't have proper papers for proposals.