Canonical citations for art and pyDelphin

ace

#1

Dear all but especially @sweaglesw and @goodmami,

Are there canonical citations for art and pyDelphin, either publications or specific websites to point to? (Asking for an author of a handbook paper…)

Thanks!
Emily


#2

For PyDelphin there is not. PyDelphin has been partially described by the following:

  • Ann A. Copestake, Guy Emerson, Michael Wayne Goodman, Matic Horvat, Alexander Kuhnle, and Ewa Muszynska. Resources for building applications with Dependency Minimal Recursion Semantics. Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016).
  • Michael Wayne Goodman. Semantic Operations for Transfer-based Machine Translation. PhD Dissertation, University of Washington. 2018.

And the GitHub URL is fairly stable so it is sometimes used in place of citations: https://github.com/delph-in/pydelphin

If a canonical citation is desired I can submit it to JOSS or something similar, perhaps with the v1.0.0 release, but that may take a few months.

Do note that the current spelling capitalizes the initial P in PyDelphin; the initial lower-case thing was awkward, particularly at the beginning of sentences.


#3

Thanks, Mike! Generally speaking, I think it would be beneficial to you to have a canonical citation for this widely-used toolset that you spearheaded, so I would encourage you to make that submission to JOSS. In the meantime, do you have a preference among the three options you listed?


#4

I believe having a canonical reference is always good. But is JOOS a good target?


#5

Ok I’ll look into a JOSS submission, although I’d also like to submit to NLP-OSS if there’s a second workshop, as that seems like an appropriate venue.

If it’s about DMRS (conversion, modeling, manipulation, etc.), as well as SEM-I and MRS support, then my dissertation has more detail than the LREC paper, so I prefer that one in general. If it’s about the [incr tsdb()], TDL, REPP, TSQL, VPM, [incr tsdb()] derivations, ACE or HTTP interfaces, or YY token lattice support, then I don’t think any publication is particularly relevant, and the GitHub URL would suffice.

Sorry, I don’t know what FOOS is and internet searches are not telling me anything helpful. Is it a software-focused journal like JOSS?


#6

Sorry, my mistake, I edited the comment. I was trying to ask about JOSS. I haven’t heard about it before. I wonder if the journal has a good reputation.


#7

JOSS is a journal for software projects. They have some requirements about software being well documented and they are generally about giving a citable “publication” to software projects that aren’t themselves some scientific contribution. It is affiliated with the Open Source Initiative and are sponsored by NumFocus, both of which are reputable organizations I think. You can read more about it here: https://joss.theoj.org/about