\centerline{\bf Standards} \medskip \noindent \TeX\ is a standard. It is rigidly defined; every implementation of \TeX\ must pass the `trip' test before it has the right to describe itself as `\TeX'. Even the components of \TeX\ are standards; {\tt dvi} format is defined rigorously; even the format of {\tt pxl}, {\tt pk}, and {\tt gf} is defined. Thanks to this standardisation, every user of \TeX\ (\LaTeX\ and \AmSTeX) knows that he or she can expect the same results from the same input, no matter what equipment they use to prepare their document. This is a degree of standardisation which is unheard of outside the \TeX\ world. There are areas within the \TeX\ world without adequate standards. Some of these have been at least considered by the \TeX\ community. Both device driver standards and macro-writing standards have been the subject of working parties: the driver standards working party has made some preliminary announcements, but the macro-writing standards working party seems to be moribund. Other areas require consideration, including the handling of {\b\tt special}s, but this may be tackled by the driver standards people. But there are wider standards which affect us in the \TeX\ world. The \TeX\ world is only a small (and many would argue, privileged) part of then `document' world. There are international and {\it de facto} standards which are of critical importance to us. We have to acknowledge the importance of the {\em de facto} standard, \PS{}. In a sense, \TeX\ `sits above' \PS{}. No-one (well, almost no-one), would ever dream of writing a document in straight \PS{}. Normally, we write a program which generates \PS\ -- for example, \TeX{}. Nevertheless, we must be aware of the way in which \PS-% compatibility is crucial if we are to be taken seriously by the rest of the world. We must have an acceptable answer to the question `Can you generate \PS?', even if we feel that the question is ill-posed. The first `international' standard which is of importance to us is \sgml{}, the Standard Generalized Markup Language. \TeX\ is itself a markup language, but \sgml\ takes this one step further to become divorced completely from the realm of typesetting. \LaTeX\ has a closer affinity to \sgml, although it does not go quite far enough. While it is possible to argue that the basic paradigm of \sgml\ is flawed, its widespread acceptance and use (by, among others the US DoD, and the EEC) demands that we do not ignore it. Many \sgml-based systems use \TeX\ (or \LaTeX) as the document formatting engine. But \sgml\ is an existing {\sc iso} standard. An evolving standard of which we must be keenly aware is {\sc oda} (Office Document Architecture). One of the objectives of {\sc oda} is to permit the electronic interchange of documents over open systems. In the \TeX\ world we would argue that this has already been achieved. One aspect of concern to {\sc oda} is mathematical text. Again, we would argue that this has already been achieved. Sadly, the way that national and international standards are created does not ensure that the best {\em de facto} standard becomes enshrined in the ultimate {\sc iso} standard. At present, the various {\sc oda} national committees and panels are considering the input of mathematics. The principal submission which has been received suggests the use of \eqn{}. This is a somewhat limited and limiting approach. It is of the utmost importance that we in the \TeX\ world promote the other alternatives that we know (and love). We can live outside the `standards world'. It is possible. But it is uncomfortable, and ultimately it will lead to atrophy. We should ensure that decisions, like those to become part of the {\sc oda} standard, are made with reference to a wide spectrum of possibilities. It may be that \TeX\ is inappropriate to {\sc oda}. But that conclusion must be reached by active and informed debate, not by ignorance and apathy. \smallskip \rightline{\sl Malcolm Clark}