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, 29 Mar 90 14:41:59 CET Reply-To: LaTeX-L Mailing list Sender: LaTeX-L Mailing list From: PZF5HZ@RUIPC1E.bitnet To: Rainer Schoepf Status: R X-Status: X-Keywords: X-UID: 60 I would like to respond to some of Michaels comments but I think the best strategy would be if Michael and/or somebody else would come up with explicit examples how the syntax for theorem-like structures should (could) look like. >>I have started working on a prototype that I think >>will convince you otherwise. .... Please give examples. >>I don't understand this, unless you mean that the >>user should do something like >> >>\renewcommand{\thetheorem}{} >> >>after using \newtheorem to define the `theorem' >>environment. In that case perhaps it is only a >>shortcoming in the documentation. I don't recall any >>discussion in the LaTeX book of turning off numbers >>this way, only changing the form of the number, and >>so I bet that the average user would not know how to >>do it. But also I think it would be easier for the >>users to have something like \newtheorem* to do this. Yes I meant changing \the... but I accept that this is not explained in detail and it is probably not the best syntax. Using a star form to provide unnumbered versions might be a good idea. >>If I understand correctly what you mean by >>subdefinition, in the examples that I have seen this >>is generally not true. Definition 2.1' is more often >>a complete restatement of Definition 2.1, in modified >>form. Logically speaking, I don't think it should be >>a separate theorem type. However for 2.1a, b, ... I >>could agree with you. I see the point. Probably the best way would be that \newtheorem would generate two environments (i.e. also the star form). The star form could then have an extra argument were the number is specified. Using \label and \ref it should be possible to refer back to the first Definition or whatever so that a syntax like \begin{DEF*}{\ref{firstdef}'} .... would produce that wanted results. But there are some points why I'm not satisfied with such a solution: 1) This would make \begin{DEF*}{2.1'} .. possible which dangerous because changes to the numbering scheme would not affect this. 2) I'm not sure that this should be part of a general interface at least it has to be part of the peripheral part. >>Actually it is not a general design decision, at the AMS. >> >>Theorem 2.1 (Gelfand) and Theorem 2.2 [19] >> >>could occur in the same paper, not to mention in >>different articles within the same issue of a journal >>(where a common definition of \@opargbegintheorem >>would usually prevail). These are two different >>kinds of added notes, one a mathematician's name, the >>other a reference. The possibility >> >>Theorem 2.2 ([19]) >> >>seems OK in my personal opinion but as I said the AMS >>editorial department considers the parentheses to be >>redundant. I don't think this is a proper way to do it, but coming back to the way my style does it, is no problem to set up a \theoremstyle which would check whether or not the argument is a \cite. But I don't want to make this a major part of the interface. >>The point I was trying to think about is that maybe >>according to the author's underlying logic the >>`preskip' amount in this case ought to be the >>subsection preskip rather than the theorem preskip, >>assuming that they are not the same amount. In my opinion the theorem environments are for theorem like structures which come within a section or something. That is there is text below and/or above the structure. In the special case that a theorem is something like a section itself, I would say regard it as a section in this case. If normally the theorem text is set in italic this would mean that it would look different then other (short) theorems but I think this would be correct in this special case. >>I am not sure that I can agree. To a certain extent >>the numbering system in a document is interdependent >>with the logic of the author's material, and >>therefore needs to be under the author's control. >>And the form of the numbers is part of the numbering >>scheme in the sense that it is used to help the >>reader understand the relations between parts. Isn't >>this also the philosophy of LaTeX? Why else does the >>documentation describe how to redefine \thesection >>and so on? No, I don't think that the way things are numbered is part of the document, or let's say is part of the structure of the document. Of course numbering equations every chapter or every theorem would change the structure in some sense but it would not change the meaning of some reference much. Why does LaTeX allow to change \thesection? In my opinion because it is a good middle way between GML and visual layout. Redefinition of \thesection should not be necessary if there exists a suitable documentstyle. But to provide variations for the user such possibilities are available. Nevertheless one should bann them into the preamble so that an editor of a journal could easily undo such changes. >>In reading the theorem.sty article I was puzzled >>about why you provided the `margin' and `break' (and >>therefore `marginbreak') options. I would consider >>these to be changes of a kind that should be done in >>the style file, not in the document file. I prepared different theoremstyles because I thought it unlikly that all professors would like it the same way (and I was right!). Therefore the strategy was: have a default in our standard document styles for master and pdh thesis but allow the user to use a different layout. Remember this is for inhouse production where it is not necessary to have always the same layout. If such a scheme is in general use it would be easy for articles submitted to journals either remove the declarations because they are all at the beginning of the document or to define journal document styles which simply ignore some of the standard theorem styles. Frank