[_delirium]

It looks like they've been doing it this way since 1917: ftp://howard.astro.ucla.edu/pub/obs/drawings/

Some information on the process: http://obs.astro.ucla.edu/150_draw.html

That doesn't in itself explain why they wouldn't be generating these diagrams automatically nowadays. Perhaps nobody's written a sunspot-finding computer vision algorithm that consistently produces the results they want, in a format vaguely compatible with the traditional drawings?