Read on Twitter
astro twitter i have something to talk about that has been in my mind for a while and i finally was able to put it into words, im open to conversation about this since a lot comes from personal experience:
the publishing paradox of software development projects in astro
the publishing paradox of software development projects in astro
since we're in the topic of papers being expected from undergrads applying to grad school , we should talk about how much harder it is to publish software development projects,,, and how common these same projects are becoming, especially for undergraduate research
I constantly have to remind myself that papers are not the ultimate measurement of success when a lot of my experience and hard work has gone into software development projects. as a software developer, you are laying down the foundation for the data analysis others will do
**and in order to work in many of these projects developing software for the community, not only do you need to have the skill to program in languages like python and C , you ALSO need the scientific understanding of the things you're developing software FOR.
and you also need that scientific understanding for testing your software. code developers in astro do a LOT and often have nothing to show for it when it comes to publications because software development papers are tossed aside
by the astro community into "version papers", into JOSS, into zenodo databases, into computer science journals that are so far removed from the astro community...
i've always wondered why we dont address this issue more.
i've always wondered why we dont address this issue more.
especially when the community would benefit from having a place where software development projects get the recognition deserved in terms of papers and publishing WITHIN the community (not general software journals).
and especially when the only chances of publication are these huge aj papers of collaborations like astropy or JOSS articles of a huge collaborating team for this software package, in which an early career contributed a year's work but OF COURSE wont ever be first author to
getting code, models, databases, model grids, etc out there in an aj or apj paper is very difficult if you (and i hate this phrase but theres no other way to describe it) don't have "science to accompany it"
which puts extraneous pressure on the same code developers, often ec
which puts extraneous pressure on the same code developers, often ec
to not only develop this entire software but then also benchmark it with real life examples AND also do science with it in order to be deemed "worthy" of a publication inside the community despite the fact that developing the code itself is already a very complete project
and what's particularly worrisome about this is that we're seeing an uptick of code development being assigned to ec (early career) and undergraduates simply cause of circumstance and us growing up closer to technology than other generations
so when graduate schools expect publications and an undergrad has 4 full projects in software development and no first author papers because of what i just discussed then that's an issue, and a big one. one a lot of students presently have to fight when applying to grad school
