Many papers use numbered references (e.g.: , ). Is this considered a good style or even a rule or would it be acceptable to use abbreviated names of the authors and year of publication (e.g. [Smith09] for J. Smith, 2009)? I find the abbreviated name reference style a lot more informative as after a while of reading papers on a given topic it’s usually easy to identify the cited publication without the need to look at the full bibliography. I have seen this style used in books and some editorials but it is not common.
The two systems are equally good but used in different communities/journals etc. You therefore need to check what is normally used in your field and when submitting manuscripts, of course, check what the specific journal uses. The fact that you say "most journals use" indicates you are in a field that uses numbered or Vancouver style (author-number) references. The author-date, or Harvard-system, is used by most journals in my field.
You don't get to choose
Although as a reader I vastly prefer the name-year system, as I don't have to look up most of the references, the advantages are rather irrelevant - in almost all cases, you don't get to choose, as you'll have to comply with the citation standard of the publication.
The journal, conference or thesis standard generelly will list the citation style required, and that's it.
In some fields one or the other style is more common, but in any case you may encounter a publication where a different style is required, so check first.
The name/year system is much better, and you should use it whenever possible. In particular, if you work in a field with preprints your preprints should use name/year or initial/year even if the journal will later force you to change it. The reason is that name/year communicates relevant information, while number communicates no information at all. Just giving numbered references means many readers won't know anything about who did what work.
It surely depends on the field we are talking about.
In medicine the number method now is the standard, probably because it improves overall readability of the text. Important references are often cited literally ("Smith et al. demonstrated that... ") and general statements by a collection of other researchers ("Many workgroups found...[23-27,57,89]).
Especially in books and reviews where the citations go in the hundreds You appreciate if the text is not clutterd by parentheses with long names but has only small-print numbers in exponential style.
In math/CS you mostly use
[Lot02,Zai04]. You can choose (unlike what most other answers impose) either of them, note for instance that both
amsalpha exist and either of them can be used in AMS publications.
It's a matter of habits which style the authors choose. The
[Lot02] style is better in most cases. However, there are communities where a numerical style is strongly prefered, moreover with the bibliography sorted by the order of appearance, and with compression turned on, because they cite hundreds of articles. And well, you don't want your in-text citations look like
[ABC00a,ABC00b,ABC00c,ABC00d,ACE00a,ACE00b,ABC01a,BC00a,BC00b,BC01,BC02], when it can be