I believe that Mathcad should be viewed from several points of view to see what capabilities it should have. The Whiteboard The first viewpoint I take is that which led to the invention of Mathcad - The Whiteboard. The concept of the whiteboard has moved on over the years to the extent that it is common to encounter active (or at least semi-passive) whiteboards that allow the user to capture what's written on a PC, with automatic text recognition and vectorization routines that 'tidy up' drawings. It's not too difficult to extend the concept to a more active whiteboard that does this live (eg, a fast touch-sensitive screen + supporting tools, eg automatic or on-demand visual 'tidying up' with automatic interpretation). It should have a drag and drop capability, so that, say, a variable could be selected and dragged into a place holder, or a subset of a matrix could be selected and copy-dragged elsewhere. Now imagine yourself either thinking things through by yourself or giving a lecture/presentation and think what features you'd find useful in such an active whiteboard - that's what Mathcad should have, eg, quick text formatting without quotes, or the ability to choose either freehand drawing or the ability to pick from a palette of simple drawing tools. A major implication of this is that notation should be flexible to allow the maths Prof their Jν(x) whilst letting the 'programmer' enter J(n,x). Instantly, this means both left and right superscripts/subscripts and the ability to format them independently of the main name. Detailed Analysis/Implementation It's fairly easy to imagine that the Whiteboard is used to sketch out the design for some particular problem and get buy-in to progress to a more detailed analysis (eg, Whiteboard: "I've got this great idea for a new nozzle design! Here's the outline ..." followed by Detailed Analysis: "OK, looks promising. Flesh it out and generate some test data we can drop into the lab people."). This would be an extension of the Whiteboard with improved data presentation and more automated control over plots and animations, and access to their data/images. Documentation/Publication/Presentation The next stage, having done the donkey work and got some results back from the lab/production people, is to formally document it and/or prepare it for publication or presentation. This requires an additional set of tools, such enhanced text formatting, equation numbering, and style-guides (plus checkers) and layout enhancements, such as 2-column presentation. As you've mentioned elsewhere, the ability to output in pdf or LaTex format would be essential (built-in not bolted-on, so that, for example, default output file names would be automatically related to the worksheet name. I think all of the above are in keeping with PTC's (and Mathsoft as was) aim of having Mathcad at the heart of a one-stop calculation management system. Whilst I'm aware of the dangers of overkill (the UML saga referred to), there are a lot of capable and competent competitors out there and not doing enough is a much a killer as capability bloat. | 
