In many cultures, family names are high-entropy and given names are lower entropy (two random people are more likely to share a given name rather than a family name). Under this assumption, the family-name citation trend makes more sense as there is a lower chance of collisions.

For Korean and Chinese names in particular, it's sort of the opposite: family names are lower entropy (e.g. Lee, Wang, Kim) than given names, which drastically increases the chance of family-name collision. I’m a PhD student of Chinese descent, and I share a family name with lots of academic peers in my specific subfield of research, and have even co-authored with unrelated individuals who share a family name with me. Family-name citations are really ambiguous and confusing to me, so I prefer omitting them altogether and using numeric links.

I don't buy the argument - when I pick the wrong "Wilson" in "Wilson (2002)" then by the time I read the bibliography, I can see that I make a mistake. Has that ever happened? Not that I know, but I saw something else, without stepping in the trap, namely that the _same_ Wilson wrote more than one paper in 2002, and they cited the _other_ paper, not the one I first thought. Again, the reason I found out is that I read the references, otherwise I could not relate this incident.

It's important to scan the references to check oneself if one agrees with the selection of references, or if the author(s) omitted important work that one knows and the authors did not know or did not want to cite due to a hidden agenda (selection bias), in which case one may email them to share additional references so that they cannot say "we never knew".

I could accept numeric citekeys if and only if the HTML rendering shows me the full entry when hovering over it with the mouse cursor (as a "tool tip"); there is absolutely no need in the 21st century to manually scroll/jump to the end without coming back when you can shop the reference inline. Hypertext is there for you to make use of it (Berners-Lee [1], 2000)!

References

[1] https://search.library.wisc.edu/catalog/999902745302121/cite ;-)

One of the nice things about Tufte-Latex is that refs get put into margin notes right at the point of reference.

Edit: example https://mirror.apps.cam.ac.uk/pub/tex-archive/macros/latex/c...

If your article is digital, just make it so that the full citation is revealed on hover (as a "tooltip"), so users don't have to jump back and forth.

edit: Tooltips supplement the traditional reference section. They are not to be printed.

Two things that I don't see mentioned is that:

(a) [Name 2005] is much easier to mentally track if it appears repeatedly in longer text than [5] (at least for me). [5] is just [5]. [Name 2005] is "that paper by Name from twenty years ago".

(b) By using [Name 2005], I might not know which exact paper this is, but I get how recent it is w.r.t. what I am reading. In many cases, this is useful context. Saying "[5] proves X" could mean that this is a new result, or a well known fact. Saying "[Name 1967] proves X" clearly indicates that this is something that has been known for some time.

"Fundamentally, the problem with this is that it's actively encouraging guesses to override the communication of ethically required citation information."

This is silly. 99% of the time, if you cite Autor et al. (2013), it'll be that Autor et al. (2013). The other 1%, it'll be another David Autor paper. The case when you guess something wrong, and really it's a different author, who then gets hurt, probably happens once a year. Meanwhile, Autor et al. (2013) immediately lets me understand what you're referencing, which [57] does not.

I wish hackernews supported references and footnotes.

Semantically tagged so folks could style them as they wanted.

What about being able to use direct voice when reporting related work? e.g. "Secondname invented a yabayaba". I have seen "[1] invented a yabayaba", but doesn't it look kind of odd?

Coming from a field where [n] routinely means {1,2,...,n}, I don't feel like the [1], [2], [3] system can win me over. But I'm ready to see some advantages it has.

An extremely detailed look at various ways to do bibliographic references in a text. And yes, digital changes everything again.

One pragmatic advantage of the Autor et al (2013) reference style over the [57] style is that it is much easier to use when you don't have software to automatically number references, and renumber then as new refs are added. In my career I did occasionally have to resort to manual reference numbering and it was pure hell.

It is also quite helpful to know from the first glance that e.g. [Wulf68b] is actually from, you know, 1968 so I probably can just ignore whatever ideas he had on what future directions the programming language design will take (heck, it pre-dates K&R C).

Maybe I am wrong, but isn't the argument here, that because you trust in your guess, you don't look up the key in the bibliography and never become aware of your incorrect guess?

I mean, regardless of the Bibkey naming, it'll be the same content in the bibliography/citations section/appendix

Personally I hate autoincrementing numeric bibkeys for the same reason I hate them everywhere -- if you remove, reorder, or combine citations, it's a huge pita to work around.

I get if you're working in latex or something you can have the bibkeys programmatically generated, but still.

So long as the bibkey is unique and has an associated value, ill always prefer something semantically meaningful in the text.

I mean, when I see a citation, I'll read the citations regardless (if it's for work etc, casually at home I'll only read it if it seems interesting or if I'm dubious). The only way I could see it being an issue is if I've already seen the citations before and somehow I'm misremembering it or ignoring some context/nuance and I decide to be lazy and not refresh the cache, but at that point it's definitely a me problem and I don't think going to the citations and looking at the source title or excerpt there will be any better.

A 1% error rate honestly seems unacceptably high if you're in the habit of reading a large volume of papers, each of which has a large volume of citations. To quote TFA, "Science involves many different error-reduction practices that take time. It's part of the job." At the end of the day, it comes down to whether you think there's value in being correct, or if you're content merely seeming correct.