X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil] [nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil]) Date: Thu, 19 Jul 90 17:33:00 PDT Reply-To: LaTeX-L Mailing list From: Don Hosek Subject: FWD: RE: Multiple refs in a \cite{} with page refs To: Rainer Schoepf Status: R X-Status: X-Keywords: X-UID: 206 From Oren Patashnik: Date: Wed, 18 Jul 1990 13:06:06 PDT From: Oren Patashnik Subject: RE: Multiple refs in a \cite{} with page refs In-reply-to: Your message of Tue 17 Jul 90 13:24:42-MDT To: "Nelson H. F. Beebe" , dhosek@HMCVAX.CLAREMONT.EDU Message-id: X-Envelope-to: dhosek nb> On the LATEX-L discussion list, a question has arisen about nb> the handling of a citation that would appear as: nb> Input: From \cite{..}, we can see nb> Output: From [19, p. 23, 89, Ch. 2], we can see In my opinion, it's better to write input so that your output looks like From Knuth [19, page 23] and Ullman [89, Chapter 2] we can see . . . Perhaps that's why Leslie didn't include the feature you're suggesting in the original LaTeX. (In fact, I believe even more---that the page and chapter numbers almost always belong in the reference list, not here in the text---but I guess I've been dealing with bibliographic issues too much of my life not to have strong opinions about them.) Nevertheless I agree that, given that there are people who will want to use the feature, it should probably be doable. nb> The \cite[]{} variant permits getting a page or chapter nb> reference into a single citation, but doesn't support nb> multiple refs of the type illustrated. Have you any nb> thoughts on how BibTeX might be extended to deal with this? As Don Hosek suggests in his reply, this is more properly thought of as a LaTeX issue, not a BibTeX issue. However, if Leslie/Frank/Rainer don't want to include this in the new LaTeX, I'd give it further thought. nb> It was also observed that the [] part illogically appears nb> before the {} part, when in the output, the order is reversed. I agree that "locally" it's illogical. But in the context of LaTeX as a whole, where in general the [] part of the command precedes the {} part, it makes more sense the way it appears now. I think that, in this case, the global consistency of LaTeX is more important than the local logic. (Sorry if I've misinterpretted anything.) nb> This suggests that perhaps an alterate format might be supported; nb> examples could be nb> \cite{Jones1989:protozoa[p. 23],Lamport:latex[ch. 2],Knuth:texbook} nb> \cite{Jones1989:protozoa{p. 23},Lamport:latex{ch. 2},Knuth:texbook} nb> However, any such extension could require revision of .bib nb> bibliography file tags to remove the possibility of nb> syntactic confusion. Given the special nature of braces in nb> TeX and BibTeX already, it seems to me that the second form nb> might be introduced into BibTeX 1.0 rather innocuously, nb> since .bib tags are highly unlikely to already contain braces. The first form isn't possible, since [] are legal characters in the citation key (and I wouldn't want to introduce an incompatibility between version 0.99 and 1.00). The second form would be a possibility, though, as long as LaTeX could be convinced to parse it correctly. dh> Actually, no. The parsing can (and should) be handled exclusively by dh> LaTeX. The only thing that I think that BibTeX should be modified to dh> handle is (1) individual chapter reference lists and (2) separate dh> references and bibliographies in a document. Albeit somewhat clumsily, BibTeX/LaTeX can handle both situations now (assuming that by (2) you mean roughly that some document (say a book) has several reference lists, as well as one long bibliography at the end). The idea is that you run BibTeX once for each reference list/ bibliography you have (each one should be in a different \include'd file in the LaTeX (.tex) file; you wind up running BibTeX on the .aux files of the \include'd files). It's a bit of a pain, but Leslie and I discussed this at some length during the design stage, and we decided that the extra work involved for the author, under this scheme, is usually in the noise of the work in dealing with the (presumably) large document anyway. My gut feeling now is that unless people are encountering problems I haven't heard of with that scheme, the difficulty in implementing multiple bibliographies outweighs the marginal gain in ease of use; but I'm open to evidence to the contrary. dh> Incidentally, Oren, are you interested in being added to the LaTeX-L dh> list? Since we keep wondering into the bibliography zone, it might be dh> useful for you to hear what we have to say (if this is the case, send a dh> note to Rainer Schoepf, RZ92@DHDURZ1.BITNET). Thanks, but I think I prefer to not be on the list. (I'm trying to move to a situation where, once I'm finished with BibTeX 1.00, the frozen version, I will be in a DEK-like mode, available for bug fixes but not much else; not being on mailing lists (aside from TeXhax, which I had previously promised people I would read until 1.00 is finished) is a step in that direction, I hope.) Feel free, on the other hand, to send me any BibTeX-related questions that come up, even if you're not sure if they're BibTeX related. And feel free to forward my responses to whomever. --Oren