The sad half-life of phylogenetic superstar software

When I discovered molecular systematics in the mid-1990s PAUP and MacClade ruled the roost. Without them you were an incomplete scientist. Each could do all sorts of things and people were constantly discovering new sorts of innovative applications that you could use them for. A previous dominant, Phylip, was still very useful but did not have a slick GUI, making you feel more like an engineer than a biologist. I attended several workshops in which the coordinators claimed that putting codes in to Hennig84, and the like, was a superior approach because this somehow demanded a better knowledge of what was inside the box than if you had drop down menus - the workshops then proceeded to show us how to use codes in much the same way that you choose menu options. With the explosive proliferation of analytical methods and increasing sophistication of molecular evolutionary modelling other tools became necessary adjuncts from the late 1990s but these always seemed limited and clunky - customized widgets to do single tasks, which again demanded some sort of programming skill - if only esoteric command line statements. Then, both PAUP and later MacClade came out in spectacular new "Leatherman" versions that could do almost anything you wanted, that could be command line or batch file programmed, and that really ruled the roost. Like all empires their crowning moments were probably also where they started to fade. The manual for PAUP* just never arrived - although the brief instruction documents were really useful, they were often out of synch with the current version of the program (which often had additional, unstated commands available). It seems it was just too great a task for a single practising evolutionary biologist to plug the bugs, develop the program and write an all encompassing manual for this behemoth. The era of single author programming in molecular biology had closed. In acknowledgement the authors of MacClade introduced Mesquite, a comparable open-source, modular project, which has yet to achieve the former popularity of MacClade. You can still buy PAUP* and MacClade from Sinauer, I just did myself, but the facade had crumbled long before Mac switched to Intel chips, divorcing most of the phylogenetics market from their favorite program. Once again we are using a plethora of customized, command or code driven widgeoids for an apparently inexhaustible range of essential new analyses. If we ever come as close to a consensus of analytical tools again it is likely to be with a large multi-functional open source project that interfaces with the engines of these singularities. Fifteen years or so as leaders in an open, competitive rapidly growing field is a huge achievement. I salute David Swofford and the Maddisons for their outstanding contribution to our understanding of the tree-of-life.