There is always a bibliography in a research article, for each quoted paper there is plenty of information (title, authors, year, journal, etc..) and this is very convenient for the reader.

But in the article itself, those articles are always named using the author(s) name(s) rather than their title. An example with this randomly chosen paper :

A random article

This is very common, at least in computer science and math. I think it's not a good practice because it's what has been found which is important and not who has found. It would be better use articles titles, if someone needs the authors names he can look at the bibliography.

I assume that titles are well chosen and give a good first idea of the contents, but anyway it can not be more meaningless than authors names.

Someone can argues that name is often shorter than title. I think it's not a good point because sometimes there are three or four names and because we could just use the bibliography reference number in order to be as short as possible, but it's not the aim of an article.

The same occurs with theorems and algorithms names. It's quite difficult to memorize what Kosaraju algorithm is, it would be much more easier to memorize what double traversal algorithm is.

Furthermore, science is timeless and this intrusion of a temporal thing is very unaesthetic.

On the other hand, don't be mentioned by other scientists could be less stimulating for authors. In addition, it could be difficult to find meaningful and short title for each article.

So, is using author name in order to refer to an other research article a good practice ? The idea is to complete my pros and cons lists and then possibly to conclude.

  • 3
    "Science" may be timeless, but the "pursuit of science" and "scientific writing" are very much human endeavors. Considering that metonymy is long ingrained as a part of human language, and is recognized to play a role in human cognitive development (see e.g. this collection), in the context of papers written for and by human beings such usage certainly is not bad practice. (And don't ask about aesthetics: that's "primarily opinion based".) – Willie Wong Jun 22 '15 at 11:43
  • 5
    I'd be terrified to see an in-text citation of a title like "Hydrogen bonding in diols and binary diol-water systems investigated using DFT methods. II. Calculated infrared OH-stretch frequencies, force constants, and NMR chemical shifts correlate with hydrogen bond geometry and electron density topology. A reevaluation of geometrical criteria for hydrogen bonding". – Massimo Ortolano Jun 22 '15 at 12:01
  • 1
    "A math article should be about math and only about math, not about..." c.f. the much-praised article by Bill Thurston in this regard. – Willie Wong Jun 22 '15 at 12:20
  • 1
    @user3702657: No, absolutely: I like the way it is: a name and a year (or a number), or simply a number. Anything longer would lead to unreadable papers. – Massimo Ortolano Jun 22 '15 at 12:45
  • 1
    @MassimoOrtolano What I tried to say is that you never just give a reference whithout anything else, you still introduce a reference : why it is connected to your work, why it is interesting, etc.. So, a good title could subsitute this explanations and could be as readable as they are. – François Jun 22 '15 at 13:04

I think it's not a good practice because it's what has been found which is important and not who has found.

Basically, I agree with this assessment. This does not mean that it does not matter who discovered or developed something, but it does indeed seem unreasonable to memorize metadata such as author name or publication year.

Like many conventions and customs, what is done in publications nowadays is inherited from earlier times and other disciplines. Some of these conventions might indeed not be suitable forever or for all disciplines, but it takes a lot of time to change commonly applied customs.

Two concrete examples, as you mention the field CS:

  • Many subfields of CS order authors by degree of concrete contribution. In many cases, the first authors are young researchers who conducted implementation and studies of few or a single project themselves (as opposed to more seasoned researchers, who are rather involved with several projects on a higher level of abstraction, thus rather bringing them towards the end of the aithors list). With large numbers of doctorates awarded each year (providing an impression of the numbers of young and still fairly unknown researchers around), it is questionable whether pointing to a publication in the form "first author et al." is likely to be more helpful than a totally nondescript numerical reference.
  • Presumeably, CS is slightly special compared to many other fields in that it lends itself to creating concepts that start a life of their own. For instance, concepts presented in papers frequently get fancy names, and reference implementations may be directly built upon by people who have never met or communicated with the original authors. In quite some cases, these concept names may become much more recognizeable than any of the many people who worked on the concept.

Thus, while there are various ways to meaningfully refer to such papers without relying on the author names and by rather relying on the content of the paper, it is difficult to standardize them:

  • Not every concept gets a name. Some authors prefer not to name their concepts.
  • Likewise, not every paper title is descriptive. Some titles explain what the paper does, some explain what problem is tackled, some just grab the readers' attention.

The author name, on the other hand, is always around for articles, which is why many authors might rely on using thosr rather than deciding on a by-case basis what is the most recognizeable pointer for a given work.

  • You point out exactly my problem, this unaesthetic mixture between temporal metadata (unreasonable to memorize for ever) and the timeless article. And you point out two bad agreements to stamp out before : unnamed concepts (If there isn't word, we can create one respecting ethymology rules) and bad titles (including descriptive but too long titles). – François Jun 22 '15 at 13:14
  • And the little meaningful information suggested by this metadata is indeed timeless information which could be suggested by a timeless title as well as in the metadata and even better. – François Jun 22 '15 at 13:30
  • What are the (currently two) downvotes for? Especially as the OP accepted this answer as the, in their opinion, best answer, it would be interesting to know what factual errors other users see in this response. – O. R. Mapper Jun 29 '15 at 7:18

I think it's not a good practice because it's what has been found which is important and not who has found.

That is not the perspective that many mathematicians take. If you read the American Mathematical Society Ethical Guidelines, they state:

The correct attribution of mathematical results is essential, both because it encourages creativity, by benefiting the creator whose career may depend on the recognition of the work and because it informs the community of when, where, and sometimes how original ideas entered into the chain of mathematical thought.

As you can see, there is a strong focus there on the ''who'' and ''when'' of a result. The focus on remembering who proved each result is a deeply ingrained aspect of mathematical culture.

On a separate note, the question states

It's quite difficult to memorize what Kosaraju algorithm is, it would be much more easier to memorize what double traversal algorithm is.

That may be true, but I think it is easier to remember what Djikstra's algorithm is than to remember what the "spanning tree algorithm" is, because there are many spanning tree algorithms. The same would hold for the "prime factorization algorithm" -- which one is that?

  • 1
    In regards to the final paragraphs: there is a further obvious context dependence. One should usually try and refrain from writing "Euler's theorem" without additional descriptors; but everyone knows what "Galois theory" refers to. – Willie Wong Jun 22 '15 at 11:47
  • 1
    It seems to me that it's not very useful to know "who" and "when" and it can even enclose the reader in an historical point of view which restrain him in finding a better point of view. But you point out a first problem to solve : some mathematicians don't really like maths, they like to be mathematicians. But this is a subproblem of a more general current one, individualism. – François Jun 22 '15 at 13:00
  • 2
    i like the focus of your answer, which is on a particular aspect of the question, namely the value of the author vs the message. in this regard the OP's concern seems to be that emphasizing authorship encourages (fallacious) argumentation by authority, whereas your response states that emphasizing authorship expresses recognition were it is due. there is a trade-off here. – henning -- reinstate Monica Jun 22 '15 at 13:23
  • 1
    @WillieWong: The context remark is an important one. Interestingly, as I am not a mathematician, I had never heard of the term "Galois theory", and as it is not self-descriptive, I had to consult external resources to find out what it is about. Contrast this with "binary search", "bucket sort" and "quicksort", which - even if I did not know them - are instantly recognizeable to refer to searching and sorting, respectively. – O. R. Mapper Jun 23 '15 at 11:43

Which in-text citation handles are used depends on the style manual of the journal in which the article is published. Some use footnotes, some are numerical, and many are variants of the author-year style, for example APA and MLA. I have never seen a citation style that uses full titles as in-text citation handles, certainly because they are much too long to be read conveniently. Some styles use keywords from the title, though.

Multiple author names are usually abbreviated "et al." and reference numbers convey even less information than the author name (which often already is a good short-hand if you are familiar with the field).

  • I understand that it depends on the style manual of the journal, but I think many style manual are bad concerning this rule. Ok, the entiere title would difficult to read, but it could be just the reference number in the text and a foonote for the title. – François Jun 22 '15 at 11:27
  • 6
    So essentially you are advocating a system which makes papers harder to read. No thanks, I say. – Willie Wong Jun 22 '15 at 11:35
  • 2
    @user3702657 sure, some people prefer footnotes over in-text citations, probably for this reason. i find this much harder to read, in particular with recurrent citations of the same source, which then take up too much space or require op. cit / supra note or some other inconvenient work-around. in the end, it is a matter of taste. – henning -- reinstate Monica Jun 22 '15 at 11:38
  • 1
    This answer somehow doesn't respond to the question. From the example in the question, it seems that the paper does use numeric in-text citation handles. However, the question asks about the way these are embedded into the text, which in the case of the paper seems toninvariantly happen by mentioning the author name(s), which may or may not make sense depending on the circumstances. – O. R. Mapper Jun 22 '15 at 11:45
  • 1
    @O.R.Mapper it all goes down to: has the reference number a grammatical role in the sentence, if any, and how is your sentence built around the references? As I've said, some might say that my example is "bad English" while others wouldn't care at all. In your examples, the references can be omitted and the sentence would still be correct, my example would not. – PatW Jun 22 '15 at 18:06

Furthermore, science is timeless and this intrusion of a temporal thing is very unaesthetic.

Name and year citations give a lot of information to a reader: knowing when and by whom something was proven often tells you a great deal about what techniques were likely used to prove it and the broader context the result fits in. Knowing the title tells you basically nothing, other than how to look up the paper (and you can look in the references for that).

  • Knowing who wrote something is only useful if you happen to know the author, which can be somewhat unlikely if you only see the last name of the first author alone, as I have described in my answer. – O. R. Mapper Jun 22 '15 at 13:51
  • 2
    @O.R.Mapper: Even then there may be some use (for example, identifying that multiple cited papers are written by the same person, or by overlapping people). But in my experience, that situation is quite unusual: reading a paper in a topic I know, I generally recognize the majority of the names being cited. – Henry Jun 22 '15 at 14:04
  • @Henry: Granted, that may well boil down to personal preference. Personally, I usually fail to recall or recognize even names of authors whose papers I have cited in my own works several times, but I also know some colleagues in the same field who are somehow able to remember names and even use them as a kind of a mnemonic device to remember the works published by the respective people. – O. R. Mapper Jun 22 '15 at 17:57

To expand upon hennings answer there is (I think) a rationale behind the use of abbreviations and/or numerical references. The number of pages in a journal generally falls into a certain range and even with abbreviations some authors struggle to convey all the important information.

Then there is the point of readability. A general title will - most probably - be longer than the (leading) author's name which will just clutter the page. Whilst an informative title of a paper should be the norm not everybody thinks this way. Including the title of the reference into the body of the article will certainly make it harder to read or skim over to asses the importance/correctness of the paper because one will have to filter more text.

Not the answer you're looking for? Browse other questions tagged or ask your own question.