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, 3 May 90 00:34:28 CET Reply-To: LaTeX-L Mailing list From: "Nelson H. F. Beebe" Subject: Re: Blank lines in math, array, eqnarray, et al To: Rainer Schoepf Status: R X-Status: X-Keywords: X-UID: 95 [Don Hosek responded with an example showing the omission of a closing dollar sign, followed by paragraphs of text, followed by another $mathstuff$ section.] Thanks for your comments, Don, on the reason that blank lines are not permitted in math mode. Your example is similar to those that illustrate why you need \long\def\foo{...} if the definition text contains a paragraph break. (TeXbook p. 205). Forbidding paragraph breaks in such conditions can certainly help to localize errors, but I still wonder if the inconsistency I noted in eqnarray/array/math vs. tabular is not even more glaring. [And, no, I don't want blank lines in tabular made illegal!] Also, when the user has declared an environment the LaTeX way, with \begin{foo} ... \end{foo}, I maintain that action has already been taken to limit the effect of whatever might happen in that environment, and extra steps to prevent possible missing terminators may be going too far. The situation described in my posting is a little different >from $ ... $, because with dollar signs, there is no way to distinguish an opening delimiter from a closing delimiter. With \begin{math} .. \end{math}, there is no such problem. A similar situation exists in programming languages; many of us feel that the unique closing keywords of SGML, Modula2, and SFTRAN3 are superior to the anonymous braces of C, or the BEGIN END brackets of Pascal. -------