\title{On the fringe of Boston} I confess that my enthusiasm for the TUG conference in Dedham was limited. In a hotel on the outskirts of Boston; no electronic access; height of the summer; Red Sox out of town; first four day meeting; first meeting with `one-day' tickets; major accent on the publishing fraternity -- persons in suits. Oh dear. It turned out a whole lot better. The theme of this year's meeting was `Inroads into Publishing' and stemmed from a proposal made by Lance Carnes last year to try to arrange a meeting which would be strongly targetted towards the publishing industry. This was seen as part of a longer term policy to ensure the future of \TeX\ by promoting it to a wider audience. The first two days were more strongly oriented towards the traditional publishing business than the others. The conference proper started with a keynote address from Nico Poppelier from Elsevier. In general this was a good talk, but I found it singularly unfortunate at a forum where there was an attempt to convince and impress the conservative business world, that Nico should take time out to list his set of `needed changes' to \TeX. This theme, of `how shall we change \TeX' was one which was repeated again and again throughout the conference. There is undoubtedly a body of opinion which sees the need for change. However, as Mike Spivak's {\sl Contrarian View} shows, there are opposing views too. This meeting again took the pattern of `networking lunches'. The notion is that the lunch tables are labelled with various topics and you go to the table with your favourite topic. It is certainly a useful idea for a meeting at which many people may be new, and who whant to locate others with similar interests. An innovation this year was the introduction of panels. To my surprise it worked remarkably well. There were a few good opinionated panellists. Reverting to tradition, Personal \TeX\ provided the wine and cheese reception, but another innovation -- the wine didn't run out. I don't know if the parallel sessions were an innovation, but they were a pain. Memo to conference programmers: {\sc don't have parallel sesssions!} The other component were the workshops. I felt sorry for Chris Carruthers {\sl Interpreting \LaTeXsl\ Error Messages}, who had a group including Frank Mittelbach, who really ought to know enough not to make \LaTeX\ errors. To hand out a few honours: probably the best function was the Clambake, sponsored by ArborText and Micro Programs -- all those lobsters! The best content -- Yannis Haralambous; the best panel -- historical roundtable. But I didn't get to everything. What about the bowling you will ask. Candle\-pin; ten pin, with very much smaller balls (no holes!) and an odd set of rules. But it was an interesting challenge. So in the end it all worked out quite well.\author{Malcolm