although long strings of braced arguments for a single control sequence
is the accepted latex convention, i do think that using the key=value
syntax (as suggested by sebastian) is preferable in a number of these
situations. omission of an unneeded key is easier for the user, and
omission of an obligatory key is just as easily detected.
sebastian has already pointed out the absence of the author's full name.
at ams we extract the front matter to be used for tables of contents and
indexes. it would be very helpful to be able to identify unambiguously
an author's surname -- it is now impossible to handle Brian Hamilton
Kelly without manual intervention, and the proposal does not improve
the actual publication information (although it would be added in
production, not by the author) should be accommodated for the purpose
of contents lists and indexes: year, month, volume, issue, pages, pii
or similar identifier for use as an identifier for electronic publication.
we also publish translation journals. for this we need the publication
information for the original (possibly including the math reviews number)
and translation information (translator, translation editor, possibly
other notes on the translation such as "Proof of Theorem 23 added by
the author after original publication").
in addition to keywords, our journals carry subject codes in accordance
with the math reviews subject classification.
we agree with sebastian's comment that the abstract is part of the
topmatter; for one of our journals, the abstract is printed at the
there are not enough addresses; in addition to those shown we have
encountered current affiliation, temporary address, current as well
as permanent e-mail address (to recognize "visiting" positions), and
i'm not convinced that itemization is the best way of handling the
address lists, although some counting mechanism is certainly needed.
this shouldn't be considered an official ams list, merely a compilation
of the most obvious things that i happen to have come across.
-- barbara beeton