A unique feature of PublishPlot is that you never set absolute size of plot elements. Instead, all plot features are sized relative to the size of the plot. This approach mimics the way illustrators used to draft plots before using computers. For example, one design rule-of-thumb is to have plot symbols be 1-2% of the plot diagonal. Similarly in PublishPlot, all size settings are done as a percent of the plot diagonal rather than a fixed size. Even when setting font sizes in points, you are really setting a "virtual" font size that gives the actual size for a full-page, landscape-mode plot (i.e., about a 10 inch diagonal plot), but scales it for other plot sizes. This approach is ideal when resizing plots to fit in space provided by your publication needs.
PublishPlot is a scriptable application and can be controlled by writing AppleScript or Python scripts. Those that are developing their own custom software or scripts can use this feature to provide plotting support for those custom tools. PublishPlot has built-in tools for simple data transformations and selected fitting functions. By writing scripts, however, transformations and fits can be developed for any custom needs. PublishPlot lets you plot any basic function of x. Alternative, a script could be developed to plot more complex functions.
You can edit plot data right in PublishPlot, but those features a basic (by design). The intent of PublishPlot is as a post-processing tool for plotting data created in other software, stored in spread sheet documents, or saved in plain text files. This need is common for scientists, engineers, and other academics who might obtain data by a variety of methods. Normally you return to the originating tool to edit the data rather the use the built-in PublishPlot methods.
If you want great plots and you want them fast, PublishPlot might the tool for you.