Producing a volume of a journal in LaTeX

Question

As a matter of academic curiosity, I've been wondering for a while how one could produce an entire volume of an academic journal (not just a single journal article) using LaTeX. It seems clear to me that a number of publishers use LaTeX to produce their journals (IEEE, American Physical Society, and Elsevier come to mind) especially since many journals provide LaTeX document classes, but it's not clear to me how they go about it. Specific questions that I have include:

1. Supposing every article is submitted in LaTeX format, how are they combined to produce the volume of the journal? Is it possible to use \input{} or \include{} on the standalone articles (and are there any existing document classes that allow one to do this), or is it necessary to edit the preambles of the individual articles?
2. How does one produce a table of contents for a document consisting of multiple parts, each with their own title and author(s)?
3. How does one handle potential name conflicts between references in different articles (cross-references, citations, etc.)? The only way that I can think of offhand (for cross-references, at least) is to adopt a naming convention for labels, e.g. \label{author:label}, which should prevent many (but obviously not all) name conflicts.

Answer 1

the ams generates journal issues using a combination of unix scripts and a "template" produced from a database, identifying the articles to be included, the starting/ending page numbers, and other relevant data. articles are processed individually -- there are no "mutual" cross-references, so that complication is avoided -- and combined only after the dvi/ps/pdf output is generated. more scripts generate tables of contents, indexes, etc.

every publisher, no doubt, has their own procedure; some, i know from hearsay, resolve all data into an xml archive. the ams is possibly in the minority for using the latex files as the underlying archival data, but until either latex/xml/mathml translators become more reliable, or good, preferably free, xml-authoring tools for math are available, the volume of math in ams journals makes latex the most practical archival format. we keep looking ...

Answer 2

There are various approaches. At one end of the spectrum, you have small-scale stuff which can be hand-edited into a single file. We do that for the UK-TUG magazine Baskerville. However, that would soon get impractical.

The next step is to typeset each document separately, with the class set to to pull in things like page numbers, then combine the pages either using TeX or something else. TUG do this for TUGBoat, and I suspect other journals which do typeset from TeX work in a similar way. (Barbara may have more to say on this!)

There are then a lot of journals, at least in my field, that accept TeX submissions but do not typeset the authors source directly. I have a bit of insight here from my work on achemso. What seems to happen is that the LaTeX source is imported into a database, in the same way that Word sources are. The papers are then typeset from the database, which may itself use TeX or another layout system.

Answer 3

From what I understood during a cumbersome submission, they have their own "home" typesetting programs. Providing the latex interface is just the top layer to help authors write their papers; afterwards, they translate it into their bottom layer language. Not sure however...

Answer 4

I'm the managing editor of an academic journal and I've done some experiments in this area. For getting author names into the table of contents I created a new macro, \Article, which takes two arguments: the author and the title. This macro includes an \addtocontents{toc} command, which allows you to insert this information into the table of contents.

Answer 5

Combine package is the answer which handles all these problems.