Likewise the Fourier Transformation.
It started out as pure mathematics that Fourier came across and it was cool and all. And possibly fairly pointless for a while. (I don't nearly know enough math to know if it had an impact in the field on it's inherent merits)
And lo and behold, these days the Fast Fourier Transformation must be one of the most run algorithms in CS. JPEGs, Spectroscopy, MRIs, anything at all to do with frequency analysis uses this maths.
It wasn't pointless at all. It was based on mathematics that had been in development for over 100 years and was developed specifically to help understand heat transfer during the boring of cannons. It was also useful well before the development of the FFT for all sorts of electronics analysis. The FFT was developed quickly after computers became powerful enough to actually do them because the FT was already an extremely useful mathematical tool.