[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian-Astro paper on the horizon?

hi Ole, all,

On Wed, 23 Nov 2016, Ole Streicher wrote:

This brought up the idea that in the next half a year or so, we should
write a refereed paper about Debian Astro, with the contributors to
Debian Astro as authors. This paper should give us more popularity
(especially when Stretch is out), and could be a citable reference for
our efforts and give credit to the contributors of Debian Astro.

Who would be willing to help me there?

I would be happy to participate - hopefully on a scale of 6 months
I'll be considered to be a "contributor" :).  Here are some snippets
of suggestions for some of the content:

* astropy:

I think it would be useful to refer to and compare and contrast with
the Astropy paper: https://arxiv.org/abs/1610.03159 . Unfortunately,
the Astropy paper seems to say nothing at all about the role of
free-licensing in the efficiency of astropy development. [Cooperation
seems to happen miraculously - ordinary human conflict disappears
without explanation...]

I think that our paper will have to give a brief description of GNU,
GNU/Linux and Debian history, and explain why a legally solid
free-software licence is a key in enabling a loose network of
typically independent-minded, geographically scattered software
developers to cooperate, despite the inevitability of conflicts in
concrete decision-making in any group of more than [choose from:
1/3/30/300 depending on taste] people.

* "free" astro manifesto paper:

This mainly US astronomer "free manifesto" paper
https://arxiv.org/abs/0903.3971 should be good to cite too. This
has a very timid reference to free-software-licensing in paragraph
2 of Section 5.

I think we should be straightforward about the free-software vs
open-source-definition dichotomy (personified in RMS vs Linus/Eric Raymond :),
briefly stating Debian's stance.

* money and jobs:

I don't think I'm qualified to judge whether we should include
comments similar to those in 0903.3971 and 1610.03159 regarding
the need for funding and institutional recognition for those doing
most of the work.

* Debian size:

A nice paper to cite: as of 2009, Debian was the "largest
component-based [software distribution] system that can be accessed
freely for study" - https://arxiv.org/abs/0901.4904 - II, paragraph 1,
first sentence. This paper also looks quite interesting in terms of
the mostly scale-free property of the Debian dependence network - there
are also remarks about saturation, which I haven't tried to understand
in detail.

* free-licensed/open-source software (FOSS) development efficiency:

I think we should refer to at least a few papers about FOSS development
efficiency.  https://arxiv.org/abs/cond-mat/0306511 shows interesting
simulations supporting FOSS-vs-closed software development - and argues
that a minimum "developer quality" is an important criterion.
https://arxiv.org/abs/1608.03608 is one in a series of papers arguing
for "superlinear production" in FOSS.

* bibliography:

I think we should include arXiv links for all articles that have them
(bibtex format data from the ADS has this), *including* those with
peer-reviewed bibliometry parameters. Peer review is a strong hint
towards research paper quality, but should not be a block against
green open access. Even a mild block to finding a reference (removing
arXiv links) is discouraging to many readers.


Reply to: