\overfullrule0pt \def\\{\hfil\break} \font\ttt=cmtt10 \def\tt{\ttt\hyphenpenalty10000} \def\subsection#1{\leftline{\let\sc\scit\sl #1}\par\noindent\ignorespaces} \let\subsubsection\subsection \def\BibTeXi{B{\iteight IB}\kern-.2em\TeX} \font\sb=cmbx8 \def\BbiBTeX{B{\sb IB}\kern-.2em\TeX} \hyphenation{abbrv} \centerline{\bf Towards \BbiBTeX\ style-files that implement} \centerline{\bf principal standards} \section{1: Introduction} \BibTeX\ can relieve an author of a lot of the work that is traditionally involved in compiling a list of references. The public domain bibliography style-files certainly produce neat bibliographies. However, documents often have to conform to a house style, to the conventions within a subject, or to an externally defined standard. \BibTeX\ users may not always be working in the same subject areas as \BibTeX's originators. This article considers the extent to which \BibTeX\ and the public domain style files are suited to this more exacting task. \section{2: Standard conventions} % Recall what speaker (a) said about the standard conventions [2 mins] % (reference-by-number, author-date) `Reading between the lines' of the \LaTeX\ manual suggests to me that the main influences behind the \LaTeX/\BibTeX\ scheme for classifying bibliographic references were van Leunen's book (van Leunen, 1978), and the precedent set by {\it Scribe}. However, I have the impression that the `principal standards' for citations and bibliographies are generally those defined by Butcher's book (Butcher, 1981), the relevant British Standards (BSI, 1989; BSI, 1983; BSI, 1978), and the {`Chicago Manual of Style'} (Chicago Manual of Style, 1982), rather than van Leunen's and {\it Scribe\/}'s conventions. This impression derives partly from the numbers of times that I've seen citations of the works concerned. I've only seen van Leunen cited by the \LaTeX\ manual, whereas Butcher's book is cited by Hart (1983), BSI (1988), Chicago Manual of Style (1982) and Williamson (1983); the British Standards are cited by BSI (1988), Butcher (1981), McLean (1980) and Williamson (1983); the {`Chicago Manual of Style'} is cited by Butcher (1981), Turabian (1987), Williamson (1983) and Lamport (1986), and forms part of the `instructions for authors' for the journal {\it Electronic publishing: origination, dissemination \& design}. BS 1629 also seems to correspond to an ISO standard. More parochially, the British Standards are mentioned in various Nottingham University guidelines about theses and dissertations. Therefore, it seems more important to help authors conform to the recommendations of Butcher's book, the British Standards and the {`Chicago Manual of Style'}, than to help them conform to van Leunen's recommendations and the {\it Scribe} precedent. This may involve supporting the author-date (`Chicago Manual of Style', 1982; BSI, 1978; Butcher, 1981), reference-by-number (BSI, 1978; Butcher, 1981) and short-title schemes (Butcher, 1981). There is also an author-number system (Butcher, 1981), but it does not seem to be used much. \caption{Table 1: BS~1629(1976)/6371 entry types compared with \BbiBTeX's}\smallskip \def\hline{\noalign{\hrule}} \tabskip0pt\halign {\offinterlineskip\strut\vrule\enspace\tabskip=0.5em#\hfil&#\hfil&\tabskip0pt#\hfil\tabskip0pt\vrule\cr \hline Category & Subcategory & Nearest \BibTeX \cr & & entry-type \cr \hline Book & Single-vol. work & {\tt book}, {\tt booklet}, \cr or other & & {\tt manual} or \cr separately & & {\tt proceedings} \cr issued & Multi-vol. work & {\tt book} \cr publication & British Standard & {\tt techreport} \cr & Technical report & {\tt techreport} \cr & Translation & \cr & Govt. publicn. & \cr \hline Portion of & & {\tt inbook}, \cr above, other & & {\tt incollection}, \cr than sep. & & {\tt inproceedings} \cr contribution & & \cr \hline Periodical or & Periodical as & \cr other serial & a whole. & \cr publication & & \cr & Limited run or & \cr & specific vol. & \cr & or issue & \cr \hline Article or & in book or\dots & {\tt incollection}, \cr contribution & & {\tt inproceedings} \cr &in periodical\dots& {\tt article} \cr \hline Patent & Patentee and & \cr & inventor & \cr & & \cr & Patentee only & {\tt techreport} \cr \hline Unpublished & & {\tt booklet}, \cr & & {\tt mastersthesis},\cr & & {\tt phdthesis}, \cr & & {\tt misc} \cr \hline } \section{3: Assumptions and abbreviations} The remainder of this article will assume that the conventions described in Butcher's book, the British Standards and the {`Chicago Manual of Style'} are indeed the `principal standards', and that the \BibTeX\ user wants to be able to produce documents that conform to them. (Take care to distinguish this use of `standard' from the \LaTeX\ manual's use of `standard' to mean `{\tt plain}, {\tt unsrt}, {\tt alpha} or {\tt abbrv}'.) The following abbreviations will be used: \item{BS:} a British Standard, or a combination of British Standards; \item{Chicago:} the {`Chicago Manual of Style'}; \item{public domain styles:} {\tt plain}, {\tt unsrt}, {\tt alpha}, {\tt abbrv}, {\tt aaai}, {\tt acm}, {\tt ieeetr}, {\tt siam} and {\tt apalike}; \item{principal standards:} the conventions for citation and bibliography layout specified by Butcher (1981), the British Standards (1989; 1983; 1978) and the {`Chicago Manual of Style'} (1982). \section{4: The problem of divergence} % On the assumption that attendees have some handouts [5 mins] % of output from Patashnik's and other public domain style files, % review the extent to which they meet standards such as the BSI and % Chicago. \subsection{4.1: Categories of document} Table 1 %\ref{1} compares the classification of documents used for the British Standards with the classification used by the public domain \BibTeX\ styles. The table assumes that, where the 1989 edition says that `other information may be added in whatever position is most appropriate', one would actually put the information where specified by the 1976 edition. Table 2 %\ref{2} gives a similar comparison for Chicago. You will see that in neither case is there a one-to-one mapping between the `\BibTeX\ category' and the `principal standard' category. Thus, if the \BibTeX\ user wishes to conform to the `principal standards', (s)he starts with the handicap of not being clear about which section of the \BibTeX\ documentation corresponds to the required section of the specification of a `principal standard'. \caption{Table 2: Chicago entry types compared with \BbiBTeX's }{\smallskip \offinterlineskip\tabskip0pt\halign {\strut\vrule\enspace\tabskip=1em#\hfil&#\hfil&\tabskip0pt#\hfil\tabskip0pt\vrule\cr \hline Category & Subcategory & Nearest \BibTeX \cr & & entry-type \cr \hline Book & & {\tt book}, {\tt manual}, \cr & & {\tt techreport} \cr & & {\tt booklet}, \cr & & {\tt proceedings} \cr \omit\vrule\tabskip1em&\multispan2{\hrulefill\hskip1em}\vrule \cr & [chapters & {\tt inbook}, \cr & \hfil or parts] & {\tt incollection}, \cr & &{\tt inproceedings} \cr \hline J'n'l article & & {\tt article} \cr \hline Unpublished & Theses, & {\tt mastersthesis}, \cr material & lectures & {\tt phdthesis}, \cr & and other & {\tt manual}, {\tt misc}, \cr & unpublished & {\tt booklet}, \cr & works & {\tt manual}, \cr & & {\tt techreport} \cr & Manuscript & {\tt unpublished}, \cr & collections. & {\tt misc} \cr \hline Public & United States & {\tt unpublished}, \cr documents & UK, Canada, & {\tt misc} \cr & International & \cr & \hfil Bodies & \cr \hline Nonbook & \hbox to 0pt{Musical scores} &{\tt misc} \cr materials & \hbox to 0pt{Sound recording} & \cr & \hbox to 0pt{Video recordings} & \cr & \hbox to 0pt{Computer programs}& \cr \hline }} \medskip \subsection{4.2: Elements of reference} Similarly, there is no one-to-one mapping between the elements of a reference as defined by the `principal standards' and the fields used by the current public domain \BibTeX\ bibliography styles: \item{\tt address} is used in three different senses. \itemitem{1}For `published works', one would presumably use \BibTeX's {\tt address} for the `place of publication'. \itemitem{2} For `unpublished works', one would use {\tt address} as `locator element' (BS 6371) or `location or sponsoring body' (Chicago, Section 16.129). \itemitem{3} For {\tt proceedings} and {\tt inproceedings}, {\tt address} is used in \BibTeX\ 0.99 as `where the conference was held', which the principal standards would probably treat as part of the title. In case 1, one might use \BibTeX's {\tt unpublished}, {\tt misc}, etc.\ for the principal standards' `unpublished works', so there is probably an overlap of the function of {\tt address} with that of {\tt howpublished}. \item{\tt month} does not seem to appear as such in the `principal standards'. It is sometimes necessary to use `date' as a means of specifying a particular issue (BS 1629 (1976 edn.), Section 4.4(b); Chicago, Section 16.124), but the {\tt month} field is not quite suitable. The non-obvious \begintt month = jul # "~4," \endtt trick has to be used. \item{\tt volume} It is not clear from the documentation whether the public domain styles distinguish between `a volume of a single work' and `a single work that is a volume in a series'. Chicago shows variation of positioning between the two cases. \item{\tt portion} Chicago 16.49--16.53 mentions the specification of chapters, parts or page-ranges. BS 1629 (1976 edn.) has the concept of `portion' for these specifications. `Specifying a portion' would be easier than getting involved in tricks like \begintt chapter = "1.2", type = "Section" \endtt with \BibTeX's |chapter| and |type| fields. \noindent In summary, the following groups seem to be merged in the principal standards: {\tt booklet}, {\tt manual}, and {\tt proceedings}; {\tt conference}, {\tt incollection} and {\tt inproceedings}; {\tt mastersthesis} and {\tt phdthesis}. The following \BibTeX\ fields seem to hold information that the principal standards would handle differently: {\tt address} and {\tt howpublished}; {\tt institution}, {\tt month}, {\tt organization} and {\tt school}. The principal standards don't seem to involve the concepts of `open' and `closed'. \medskip \subsection{4.3: Order and presentation of elements} Given the divergence between the schemes used for classification and element-definition, it is difficult to make a general comparison of features such as the order of elements, their punctuation and the fonts used. However, we can make such a comparison in simple cases, for example in the cases of references to books and to articles. If you compare BS (1989) and Chicago (1982, p.439), you will see that there is a degree of consensus about the style in which references to books are to be laid out for reference-by-number. Unfortunately, none of the relevant public domain styles ({\tt plain}, {\tt unsrt}, {\tt abbrv}, {\tt aaai}, {\tt acm}, {\tt ieeetr}, {\tt siam}) format entries in line with this consensus. For reference-by-number to journal articles, BS and Chicago differ in the order suggested for volume and year. Of the public domain styles, {\tt acm} seems closest to Chicago. Nothing seems particularly close to the BS. The public domain author-date styles are {\tt alpha}, {\tt apalike} and {\tt aaai}. Of these, {\tt alpha} has unconventional labels, and {\tt aaai} gives an unconventional bibliography (with both the short form of author-date and the long form). The remaining style, {\tt apalike} seems fairly similar to BS and Chicago for journal articles, but not for books (because of the order of place and publisher). \medskip \frame{5pt}{The {\tt \lowercase{MASTERSTHESIS}} and {\tt \lowercase{PHDTHESIS}} entry types now take an optional {\tt type} field\dots Similarly, the {\tt \lowercase{INBOOK}} and {\tt \lowercase{INCOLLECTION}} entry types now take an optional {\tt type} field, allowing `section 1.2' instead of the default `chapter 1.2'. \dots{\tt \lowercase{PROCEEDINGS}}\dots\ If you want to include the publisher's\dots\ address, put it in the {\tt publisher} or {\tt organization} field. \dots feel free to be creative in how you use these entry types\dots \dots don't take the field names too seriously\dots \dots don't take the warning messages too seriously\dots \dots if you want to include information for the day of the month, \dots {\tt month = jul \char'43\ "\char'176 4,"} will probably produce just what you want. } \nobreak \caption{Figure 1: Some hints from `\BbiBTeX ing'} \medskip \subsection{4.4: Consequences} Thus, although \BibTeX\ is very useful for reducing the work involved in producing a bibliography, and the public domain styles produce neat bibliographies we have to note that the entry-types and fields used by the public domain styles diverge from the document-categories and reference-elements those used by the principal standards. This may mean that one has to resort to tricks (such as those shown in Figure 1, %\ref{1} taken from Patashnik (1988a) to get the output `looking right'. I imagine that, if the entry-types and fields corresponded more closely to those used by the principal standards, one would have to resort to tricks less often. In addition, the punctuation and fonts used by the public domain styles will lead many potential users to quibble that `this is not how we do things in my subject'. \section{5: Towards `principal standard' styles} \subsection{5.1: Extent of consensus about} \rightline{\sl categories\slash entry-types} \noindent Thus the current public domain styles, and the corresponding entry-types and fields, do not seem to make for straightforward adherance to the `principal standards'. How could new entry-types and fields be defined that correspond more closely to the assumptions behind the `principal standards'? \topinsert \hsize6.5truein\tabskip0pt \def\hline{\noalign{\hrule}}{ \offinterlineskip\halign to\hsize{\tabskip=2em minus1em\vrule\strut\enspace#\hfil&&\enspace#\hfil&#\hfil\tabskip0pt\vrule\cr \noalign{\hrule} BS & BS & Chicago &Chicago & Examples common to\cr category & subcategory & category &subcategory & BS and Chicago \cr \noalign{\hrule} Book or other & Single-vol.\ work & Book & All but `chapters & Book \cr separately & Multi-vol.\ work & & or parts' & Technical report$^a$ \cr issued & British Standard & & &UK HMSO pub. \cr publication & Technical report & & & \cr & Translation & & & \cr & Government pub.& & & \cr \multispan2{\vrule\hfill}&\multispan2{\hrulefill}&\omit\hfill\vrule\cr%\cline{3-4} & &Public docs. & UK nonparliamentary$^b$ & \cr \noalign{\hrule} Contribution & In book\dots & Book & Chapters or parts & Collections \cr or article & & & & Conf. proceedings\cr \omit\vrule\hfill&\multispan4{\hrulefill}\cr%\cline{2-5} &In periodical\dots& Journal article & & Article \cr & & & & \cr \noalign{\hrule} Unpublished & & Unpublished & Lectures &British Lib. Add. MS.\cr & & & Duplicated material &Theses$^c$ \cr & & & Manuscript colls. & \cr \multispan2{\vrule\hfill}&\multispan2{\hrulefill}\cr%\cline{3-4} & & Public docs. & USA unpublished & \cr & & & UK unpublished & \cr & & & Canadian archives & \cr \omit\vrule\hfill&\multispan1{\hrulefill}\cr%%\cline{2-2} &Docs. resembling& & & \cr &published works & & & \cr \noalign{\hrule}}} \noindent\strut{$^a$ For technical reports, BS says that `series title and number' are `essential elements'.} \noindent{$^b$ Chicago says that these are usually treated like `privately published books'.} \hangindent20pt\noindent{$^c$ BS suggests that theses resemble published works, should have `location element' like other unpublished material, but should be given a `descriptive element' like books. Chicago suggests treating title as for a journal article, and providing `location or sponsoring body or both'.} \caption{Table 3: BS 1629(1976)/6371 and Chicago categories that are similar} \endinsert Table 3 displays some BS and Chicago categories that seem similar enough to be represented by common \BibTeX\ entry-types. Table 4 shows categories that are only described by the British Standards, while Table 5 shows categories that are only described by Chicago. Thus any new scheme should, ideally, be able to cater for all the categories shown in Tables 3, 4 and 5. One would also have to refer to ISO 690. Compatibility with bibliographic databases on CD-ROMs might be useful too, but there currently seems little consistency between suppliers, let alone consistency between conventions for fields in CD-ROMs and fields in printed bibliographies. It might also be useful to consider the database structure used for computerised library catalogues. \caption{\overfullrule0pt Table 4: Categories defined by BS 1629(1976)\slash 6371 but not by Chicago} {\overfullrule0pt\offinterlineskip \def\hline{\noalign{\hrule}} \overfullrule=0pt \tabskip0pt \halign to\hsize{\vrule\strut\tabskip3em minus2.5em\enspace#\hfil&#\hfil\strut\tabskip0pt\vrule\cr \hline Category & Subcategory \cr \hline Portion of `book or other & \cr separately issued publication' & \cr other than a separate & \cr contribution & \cr \hline Periodical or other serial & Periodical as \cr publication & a whole \cr \omit\vrule\hfill&\omit\hrulefill\quad\tabskip0pt\cr & Limited run or \cr & specific vol. \cr & or issue \cr \hline Patent & Patentee and \cr & inventor \cr \omit\vrule\hfill&\omit\hrulefill\quad\vrule\tabskip0pt \cr & {Patentee only} \cr \hline }} \medskip \caption{Table 5: Categories defined by Chicago but not by BS 1629(1976)/6371} {\overfullrule=20pt\offinterlineskip \tabskip=0pt \halign to \hsize{\vrule\strut\enspace\tabskip=3em minus2.5em#\hfil&#\hfil\tabskip=0pt\vrule\cr \hline Category & Subcategory \cr \hline Public documents & US Congress \cr & US Executive departments \cr & US Statutes etc. \cr & US States and local \cr & UK Parliament \cr & US Published records \cr & Canadian government\cr & International Bodies \cr \hline Nonbook materials&Musical scores$^a$ \cr & Sound recordings \cr & Video recordings \cr & Computer programs \cr \hline}} \noindent{\strut$^a$ `Follow rules similar to those for books.'} \medskip \subsection{5.2: A division into categories/entry-types}{\tolerance10000 From Tables 3, 4, 5, it seems that a project to provide style-files to implement the `principal standards' might start by defining entry-types such as the following. \item{\tt booklike} The categories of document specified in Section 4.2(a) (publication as a whole) of BS 1629 (1976 edn.) (with the exception of `works issued in series by research bodies\dots'). This seems the same as the group of documents classified as `Books' by Chicago (excluding Chicago's `chapters or parts of a book'). This category might include some of Chicago's `Nonbook materials', e.g. musical scores. \item{\tt report} Publications `as a whole' issued in series by research bodies and similar organizations, as specified in Section 4.2(a) of BS 1629 (1976 edn.). Differs only from {\tt booklike} in that `series title and number' are required. \item{\tt booklikeportion} A portion (other than a separate contribution) of a document in the category specified in Section 4.2(b) of BS 1629 (1976 edn.) (excluding documents for which `series title and number' are required). (Looks as if it could be implemented by a {\tt crossref} to a {\tt booklike} database entry.) Probably equivalent to the optional argument of |\cite|, but worth doing to give the ability to conform to BS 1629 (1976 edn.). \item{\tt reportportion} As {\tt booklikeportion} but with `series title and number' required. \item{\tt contribution} Contribution to a `book or other separately issued publication' as defined in Section 4.4(a) of BS 1629 (1976 edn.). `Chapter or part of a book' as defined in Sections 16.49--16.53 of Chicago. \item{\tt publishedlike} `Documents resembling published works' as specified in Section 10 of BS 6371. This category would include theses. It looks as though: \itemitem{\it (i)} the thesis information that BS 6371 would use to `compile the descriptive element like a book' is the same as that needed for Chicago to treat the thesis like a journal article \itemitem{\it (ii)} the BS 6371 `location element' is much the same as the Chicago `location or sponsoring body or both'. This category might also include some of Chicago's `nonbook materials', e.g. sound recordings. \item{\tt patent} Patents, as specified in Section 4.5 of BS 1629 (1976 edn.). \item{\tt public} Public documents, as specified in Section 16.141 of Chicago, but excluding those `cited like privately published books' (Chicago, Section 16.162) which would be {\tt booklike} and those to which BS 6371 (except Section 10) and Sections 16.158, 16.171 and 16.172 of Chicago apply (which would be {\tt unpublished}). \item{\tt unpublished} To which BS 6371 (except Section 10) and Sections 16.128, 16.131--16.133, 16.158, 16.171 and 16.172 of Chicago apply. \item{\tt periodical} As defined in Section 4.3(a) of BS 1629 (1976 edn.). \item{\tt periodicalrun} As defined in Section 4.3(b) of BS 1629 (1976 edn.). Might be implemented as a {\tt crossref} to a {\tt periodical}. \item{\tt article} This category would include `a contribution or article in a periodical or serial publication' (Section 4.4(b) of BS 1629 (1976 edn.)), and `journal articles' (Chicago, Sections 16.98--16.127). \smallskip} \noindent As it happens, this approach gives 12 categories, compared with the 14 for {\tt plain}, {\tt unsrt}, {\tt alpha}, {\tt abbrv}. (Studying the unified approach in the 1989 edition of BS 1629 might enable one to reduce the number of categories still further. On the other hand, the sentences about `Other information may be added in whatever position is most appropriate' may lead to retention of the categories suggested in the 1976 edition, in order to ensure that `other information' is placed appropriately.) \medskip \subsection{5.3: Elements/fields} A project that aimed to implement the `principal standards' would also have to define \BibTeX\ fields (within the bibliography entries) that correspond to the `elements of a bibliographic reference' defined by the `principal standards'. With the British Standards, it may be possible to use the BS `essential' and `supplementary' or `optional' elements directly as \BibTeX's required and optional fields. It looks as if these elements would provide the information required by Chicago, but someone would have to work through all the Chicago examples to check this! To support the `short title' system (Butcher, 1981; pp. 177--8), a field such as {\tt shorttitle} seems desirable. This approach might give fields such as the following for the entry-types postulated above. {\tolerance10000 \item{\tt booklike} Required fields: {\tt author}, {\tt title}, {\tt year}. Optional fields: {\tt shorttitle}, {\tt transtitle}, {\tt origtitle}, {\tt edition}, {\tt alleditor}, {\tt alltranslator}, {\tt allillustrator}, {\tt thiseditor}, {\tt thistranslator}, {\tt thisillustrator}, {\tt placeofpub}, {\tt publisher}, {\tt noofvols}, {\tt pagination}, {\tt mentionofany}, {\tt size}, {\tt seriesinfo}, {\tt isbn}, {\tt price}. Here {\tt author} may be `compiler, personal or corporate'. The meanings of other fields should be obvious by reference to BS 1629 (1976 edn.). \item{\tt report} As {\tt booklike}, but with {\tt seriesinfo} required, rather than optional. \item{\tt booklikeportion} Same fields as {\tt booklike} except that: `book trade' information, namely {\tt alleditor}, {\tt alltranslator}, {\tt allillustrator}, {\tt thiseditor}, {\tt thistranslator}, {\tt thisillustrator}, {\tt noofvols}, {\tt pagination}, {\tt mentionofany}, {\tt size}, {\tt price} are neither required nor optional. \item {\tt volpart} and {\tt volparttitle} would be optional and {\tt portion} would be required. \item{\tt reportportion} As {\tt booklikeportion} but with {\tt seriesinfo} required rather than optional. \item{\tt contribution} Required fields: {\tt contauthor}, {\tt conttitle}, {\tt pubauthor} or {\tt pubeditor}, {\tt pubtitle}, {\tt year}, {\tt portion}. Optional fields: {\tt shorttitle}, {\tt conttranstitle}, {\tt contorigtitle}, {\tt contmentionofany}, {\tt pubtranstitle}, {\tt pubedition}, {\tt pubvolume}, {\tt placeofpub}, {\tt publisher}, {\tt seriesinfo}, {\tt isbn}. \item{\tt publishedlike} Fields for `descriptive element' as for {\tt booklike}, with the omission of {\tt placeofpub}, {\tt publisher}, {\tt isbn}. Fields for `location element' as for {\tt unpublished} -- see below. \item{\tt patent} Required fields: {\tt patentee}, {\tt title}, {\tt country}, {\tt designation}, {\tt serialno}, {\tt pubyear}, {\tt restofpubdate}. Optional fields: {\tt shorttitle}, {\tt transtitle}, {\tt inventor}, {\tt intclass}, {\tt natclass}, {\tt appdate}, {\tt pagination}, {\tt platesetc}. \item{\tt public} Following Chicago (Section 16.141), fields for {\tt division}, {\tt body}, {\tt subsid}, {\tt title}, {\tt author}, {\tt identification}, {\tt publisher}, {\tt date}. In addition, an optional {\tt shorttitle} field, for `short title' citation. \item{\tt unpublished} Descriptive element, consisting of optional fields {\tt name}, {\tt title}, {\tt date}, {\tt designation}. Location element, consisting of optional fields {\tt place}, {\tt repository}, {\tt callno}, {\tt locwithin}. Although the individual fields of the descriptive element and the location element are optional, the descriptive element and the location element are both `required' as a whole. In addition, optional fields {\tt otherinfo} (for Section 9 of BS 6371) and {\tt shorttitle} for `short title' citation. \item{\tt periodical} Required field: {\tt title}. Optional fields: {\tt transtitle}, {\tt origtitle}, {\tt firstissue}, {\tt lastissue}, {\tt frequency}, {\tt placeofpub}, {\tt publisher}, {\tt size}, {\tt summarylangs}, {\tt price}, {\tt earliertitles}. \item{\tt periodicalrun} Required fields: {\tt title}, {\tt firstissue}, {\tt lastissue}. Optional fields: {\tt transtitle}, {\tt origtitle}, {\tt frequency}, {\tt placeofpub}, {\tt publisher}, {\tt size}, {\tt summarylangs}, {\tt price}. \item{\tt article} Required fields: {\tt author}, {\tt arttitle}, {\tt serialtitle}, {\tt year}, {\tt volume}, {\tt portion}. Optional fields: {\tt transtitle}, {\tt origtitle}, {\tt mentionofany}, {\tt placeofpub}, {\tt part}. \par} % Critique bit: "Marie Clare van Leunen had undue influence. [5 mins] % The standards should have had more influence." % Hacking at bst files. Structures in the bst files that I understand.[5 mins] % Postfix stack business that I don't. % Bit of critique: "Did Patashnik really have to use this Postfix % stack business? I don't know." % Future work for somebody: [5 mins] % - start with the concepts in the standards. % - define fields. % - define bst files that process such bib files. % - write it all up. \medskip \subsection{5.4: Implementation} The \BibTeX\ style-file language provides many facilities for manipulation of the information that appears in bibliographies. Thus, having defined new entry-types and fields, one can declare the fields to \BibTeX and also define \BibTeX\ {\tt FUNCTION}s to construct the bibliography entries from the fields I anticipate that if the business of defining entry-types and fields that correspond more closely to those defined by the principal standards was followed through, there would be less need for the sorts of tricks mentioned in Figure {1}, and the output would be more likely to be `right first time'. \section{6: Wish lists} \subsection{6.1: \BibTeXi\ itself} The current version of \BibTeX\ makes it difficult to produce the style shown in BSI (1989) for titles that involve colons. For example, BSI (1989) shows the title: \centerline{\it Shetland sanctuary: birds on the Isle of Noss} \noindent Unfortunately, the effect of \BibTeX's |change.case t$| on the likely {\tt bib} file entry would be to give `{\it Birds\/}'. It would be useful if some future version of \BibTeX\ allowed a variation on \begintt change.case$ t \endtt that refrained from `upper-casing' after a colon. \medskip \subsection{6.2: \BibTeXi\ style-files} It would be nice if someone had time to develop style-files that worked in terms of the same concepts as the principal standards, and delivered bibliographies formatted in accordance with these standards! \medskip \subsection{6.3: The combination of \LaTeXsl\ and \BibTeXi} \subsubsection{A document having more than one list} An author may wish to use `bibliography' to mean `works in the field, with some notes about them', and `references' to mean `works I have cited'. Thus, a `bibliography' may not be the same as `a long list of references', and \LaTeX's assumption that {\tt article}s have `references' while {\tt report}s and {\tt book}s have `bibliographies' may be an over-simplification. Some works have more than one list. For example a book might have a division called `references', listing works giving further information about the subjects covered in the book, and a division called `further reading', listing works that cover related subjects. Similarly, the draft new British Standard for theses (BSI, 1988) suggests a `list of references' (for `every work cited') {\em and\/} a `bibliography' (for `all sources consulted but not necessarily relevant to the thesis'). There is no harm in {\tt thebibliography} having a default title, but an official mechanism for replacing the default by a more accurate description would be very useful. Aside: Are the people who are developing non-English \LaTeX\ repeating the over-simplification? Instead of providing a mechanism for getting equivalents of `references' and `bibliography' in many languages, might it be better to just provide one default title in each language plus a mechanism for changing the title? More ambitiously, could some future versions of \LaTeX/\BibTeX\ give more general support to documents with more than one list? For example, could some |\cite|\/s generate a list of references in a brief |\bibliographystyle| to be printed under a heading of `references', while other |\cite|\/s generate another list in a different |\bibliographystyle| to be printed under a heading of `further reading'? I think that people producing conference proceedings can currently get references at the end of each chapter by applying \BibTeX\ separately to each chapter: could some development simplify their task? \medskip \subsubsection{A list with subdivisions} Under some circumstances, it may be desirable to divide a list into sections. For example, Figure {2} shows the structure of the bibliography for a history thesis. Chicago (Chicago Manual of Style, 1982; chapter 15) shows some further examples of subdivided bibliographies. \caption{Figure 2: Structure of a history thesis bibliography} {\parindent0pt \baselineskip10pt \font\ttt=cmtt8 \let\tt\ttt \begintt \section{Manuscript sources} \subsection{Nottinghamshire County Record Office} \subsection{Nottingham University Library, Manuscripts Department} \subsection{Public Record Office} \subsection{British Library} \subsection{Borthwick Institute} \subsection{Friend's Reference Library} \subsection{Dr. William's Library} \section{Printed sources} \subsection{National government series} \subsection{Historical Manuscripts Commission} \subsection{Thomason Tracts (British Library)} \subsection{Other printed sources} \section{Secondary works} \section{Periodicals} \section{Theses} \endtt } Could some future versions of \LaTeX/\BibTeX\ support such subdivisions? For example, perhaps |\section| and |\subsection| could be allowed within {\tt thebibliography}, the author could put \begintt @UNPUBLISHED{... subbib = "Manuscript Sources", ... } \endtt in a {\tt .bib} file, and \LaTeX/\BibTeX\ between them could arrange that the item in question would appear under a sub-heading of `Manuscript Sources'. \section{7: References} \let\newblock\null \let\em\rm {\frenchspacing\parindent0pt\hangindent20pt\hangafter1 BSI, BS 1629, British Standards Institution, 1989. {\it For printed matter, the 1989 edition sometimes gives less detail than the 1976 edition. The present article uses information from both editions.} \hangindent20pt\hangafter1 BSI, Citation of unpublished documents, BS 6371, British Standards Institution, 1983. \hangindent20pt\hangafter1 BSI, Citing publications by bibliographic references, \newblock BS 5605, British Standards Institution, 1978. \hangindent20pt\hangafter1 BSI, Recommendations for the presentation of theses, dissertations and similar documents: Draft for public comment, \newblock BS 4821, British Standards Institution, 1988. \newblock {\it The revised standard is likely to be published in 1990.} \hangindent20pt\hangafter1 Judith Butcher, \newblock {\em Copy-editing}, \newblock Cambridge University Press, 2nd edition, 1981. \hangindent20pt\hangafter1 {\em The {C}hicago Manual of Style}, University of Chicago Press, 13th edition, 1982. \hangindent20pt\hangafter1 Horace Hart, \newblock {\em Hart's rules for compositors and readers at the {U}niversity {P}ress, {O}xford}, \newblock Oxford University Press, 39th edition, 1983. \newblock {\it Revised by Hart's successors.} \hangindent20pt\hangafter1 Leslie Lamport, \newblock {\em \LaTeX: A Document Preparation System}, \newblock Addison-Wesley, 1986. \hangindent20pt\hangafter1 Mary-Claire van Leunen, \newblock {\em A Handbook for Scholars}, \newblock Knopf, 1978. \hangindent20pt\hangafter1 Ruari McLean, \newblock {\em The {T}hames and {H}udson Manual of Typography}, \newblock Thames and Hudson, 1980. \hangindent20pt\hangafter1 Oren Patashnik, \newblock {\em \BibTeX ing}, January 1988a. \hangindent20pt\hangafter1 Oren Patashnik, \newblock {\em Designing \BibTeX\ styles}, January 1988b. \hangindent20pt\hangafter1 Kate~L. Turabian, \newblock {\em A Manual for Writers}, \newblock University of Chicago Press, 5th edition, 1987. \hangindent20pt\hangafter1 Hugh Williamson, \newblock {\em Methods of Book Design}, \newblock Yale University Press, 3rd edition, 1983. } \rightline{\sl David Rhead}