Unfortunately, working in Physics means giving quite a few presentations. This would be bad enough, as it takes significant time away from experiments. But, adding to this, everyone uses PowerPoint. The program is wretched for this application.
For instance, say we want to display some kind of atomic structure. Atomic structures, if you don't know, are made up of atoms, which are normally drawn as spheres. Now, you would think that drawing a sphere in PowerPoint would be rather simple. Draw a circle, make it into a 3D object, voila, right?
Wrong. You get this:
Now, I have seen presentations that included spheres. How can this be done? The thing about Powerpoint is that it isn't really a graphics program. It displays graphics, but it's not so good with generating them. The only straightforward way I found of making a sphere in Powerpoint is to
1. Find a suitable 3D picture of a sphere
2. Draw a circle
3. Map the image of the sphere onto the circle.
To make a 3D picture, I took an eps from BALSAC and stripped it down to only one ball by editing the PostScript manually (You can download the stripped version here) After changing the color (around line 184), I run it through epstopdf to make a vector-graphics pdf
> epstopdf ballgen.eps
and then through imagemagick to give me a high quality bitmap
> convert -trim -density 300 -geometry 50% -rotate 180 ballgen.pdf this_color.png
After which you can get this
which looks very nice.
This illustrates that the only really reliable way to use PowerPoint is to make it a program that only displays text and bitmaps. Any other scalable media, it invariably has major problems.
For instance, I used to like Tex4ppt, which purported to convert a PowerPoint text box with TeX code in it to a scalable equation in a native Microsoft format. However, many of the structures (notably integrals) would get horribly warped if you had any animations. The solution? A Perl script which takes TeX equations and renders them as high quality PNG files. These are then added to PowerPoint as bitmaps, which it won't screw up. (Code for this is here. Requires Perl and Imagemagick, naturally.)
PowerPoint is probably great if all your graphs are in Excel, all your equations look good enough to you in MS Equation Editor, and you don't draw anything other than flow charts. For us working in science, though, it's a fucking nightmare.
