Thanks for the heads up Elle! Looking quickly through the code, I do not see much optimizations. Instead, it seems to implement an alternative CMS which does not depend at all on LCMS2.
For me this seems to be overkill… what I actually need with high priority is an optimized, fast implementation of ICC conversion between matrix profiles in relative colorimetric intent, as well as RGB ↔ Lab conversions, which cover 99% of the conversions commonly performed during photo editing.
LCMS2 already provides the infrastructure for reading or creating ICC profiles, and for retrieving the useful information (type of profile, colorants, TRCs, etc…) from them, and I am not planning to move away from it.
Moreover, the LCMS2 machinery is still very handy for more complex conversions (LUT profiles, non-relative rendering intents, partial adaptation, abstract profiles, etc…).
I have a big respect for Pippin, but in this case my impression is that he is re-inventing the wheel… are there plans to completely drop the dependency of GIMP from LCMS2? If yes, why?