\centerline{\bf BCS-ep} \smallskip\noindent The BCS Electronic Publishing Specialist Group's meeting of December 5th was concerned with publishing standards, and in particular with Standard Generalized Markup Language (\sgml{}), Open Document Architecture (ODA) and Computer-aided Acquisition and Logistic Support (CALS), the acronyms for which are frequently quoted, not always with a clear understanding of what they are about. The first speaker of the day was Professor David Barron from the University of Southampton on `Why use \sgml{}?', although perhaps `What is \sgml{}?' is really what he discussed. He began by reminding the audience that Electronic Publishing is not just Desktop Publishing, that is, in spite of all the hype there is a place for non-\wysiwyg\/, that is procedural, systems (although, of course, the two do not have to be mutually exclusive). Professor Barron then gave a brief explanation of the development of GML at IBM and of MarkUp, the two precursors of \sgml{}, together with some details of the ISO working groups and a summary of what \sgml{} is. Rather he stated what \sgml{} is not, once he had explained the principles of generic markup. \sgml{} is not a language -- it is a metalanguage -- and \sgml{} is not a formatter, although it specifies how coding systems written according to the rules of \sgml{} can be parsed and parsers can be written to output files in such a way that they can be input directly into formatters.\looseness1 Professor Barron then explained the significance of the Document Type Definition (DTD) and went through a simple DTD describing the meaning of the coding. Finally he showed how the power of \sgml{} could be used in a project like the OED and how it is invaluable when either transferring documents between different systems or producing a range of documentation in a single format or related range of formats. For those who missed this talk and would like a relatively straightforward introduction to \sgml{}, Professor Barron has published a paper which is similar in content to his talk in {\sl Electronic Publishing: Origination, Dissemination and Design} (1989) 2(1), 3--24. The second talk was on `The Text Encoding Initiative' (TEI). Lou Burnard of Oxford University Computing Services explained what this is. TEI is essentially an aid to scholarship and covers aspects outside those considered by \sgml{}, such as those concerned with textual analysis. TEI is not really concerned with coding for publishing; it is concerned with the representation of already-published texts, for example what actually appears on the page, both its content and representation. However, it is not always possible to know in advance what is to be described, so the standard \sgml{} DTD formulation is not flexible enough unless the DTD can be violated. Lou Burnard finished by describing some of the practical problems: the lack of software; the lack of success of previous standardization efforts in this area; and finally the lack of public awareness. Stephen Downie, Secretary of the \sgml{} Users' Group and previously of Unixsys UK Ltd, gave a rather different introduction to \sgml{}, by pointing out that when data and text are input the uses to which they may be put are not always all known (as TEI illustrated very aptly). If a text is input according to \sgml{} conventions and the DTD has been written properly, then the text can be used subsequently for uses which were not envisaged when it was originally input. Pathways give new meanings to data. Stephen Downie then went on to describe the use of \sgml{} in producing the Official Journal of the European Community (although in fact no actual issues have as yet been produced). The principal problem which has had to be solved is the production of nine editions in nine languages, but with synoptic pagination. Input is now carried out using the SoftQuad \sgml{} Author/Editor (on which a DTD has been written at Pindar Infotek using the associated Rules Builder). Using a specially written DSSSL (Document Style, Semantics and Specification Language), the documents are converted to SoftQuad {\tt troff}, using which all text heights are measured on the first run. Then using the \Unix\ utility {\tt awk} a master list is produced which is used in the final pagination run. There have been problems, as might be expected: producing the DSSSL has been very complex; the journal rules had not previously been clarified; US software (as usual) does not concern itself with accents and non-latin character sets; and finally, although both SoftQuad {\tt troff} is up to the task and \sgml{} reduces the complexity of the problem, high programming expertise is required.\looseness1 CALS, Judi Morrison of Interleaf explained, is a program to acquire and manage weapon systems information more efficiently. At present 96\% of manuals are not computerized, one in four contain errors and it takes up to 20 months to make corrections -- and the cost is phenomenal. In addition 10\% of the cost of a weapon system is paperwork. CALS was therefore developed to try to save money, reduce development times and improve quality. CALS therefore affects almost everybody eventually, in that contractors and subcontractors to the military purchase materials and expertise (including software) from other industries, and by 1990 all proposals to the Department of Defense have to be CALS-conforming. Of course there are many aspects of CALS, but the documentation aspects, which include \sgml{} for text and various graphics standards such as CCITT Group 4 (as Fax), IGES (International Graphics Exchange Specification) and CGM (Computer Graphics Metafile), are probably the most currently important. The final aim is to create a single database which can be accessed by both military and contractors. Whether this will actually happen remains to be seen and there are bound to be problems of security as well as of a technical nature. After lunch (which through an interesting bureaucratic mistake was different from usual!), Patrick Hull and John Dixon of Deuce Systems discussed more detailed aspects of CALS, in particular all the various standards involved, who produces them and what they mean. It is difficult to summarize what was said, but the two speakers made it very clear how complex the field is and what a challenge is there for those involved. In neither of the talks on CALS was there any discussion of what quality of weapon systems means in real terms, but it was clear that defence (or defense) and military applications have ramifications far beyond the immediately obvious. The final talk of the day was from John Stratton of the National Computing Centre and was on The Open Document Architecture (ODA). This is another ISO standard which is like \sgml{} but not like \sgml{}. It grew not from the world of publishing and computer text processing and programming, but more from the world of ECMA, the European Computer Manufacturers Association, and was initially concerned (as is probably obvious from the name) with Open documentation. The aim was to be able to provide open exchange of documents, either in a fully formatted way or in a form which can be edited and reformatted. ODA therefore considers a document in terms of both its structure and its content or, put another way its logic and its layout. The logical view is similar to that provided by \sgml{}, while the layout view is in terms of blocks, pages, page sets etc. In addition there are document management attributes and standard ways of handling graphics. In summary, a day full of acronyms, but most people probably came away feeling that they understood at least a few of them and the concepts behind them. \rightline{\sl David Penfold}