Gordo, is this a trick question? It depends on the Pantone library being referenced.
Pantone Solid Coated = converts from the library L*a*b* colour values through your CMS and various conversion options to your CMYK working space profile. The final values will vary depending on the conversion settings and output space which will also vary with GCR values etc. If your CMYK working space profile is "accurate" for the final print condition, then you may get "better" results for in-gamut colours using this approach. There are no fixed CMYK values in this library, it all depends on the conversion. Obviously, if the Pantone in question is out of gamut for your CMYK print condition - then it comes down to artistic licence. The L*a*b* values for Pantone 173 C in CS5 are 52L* 55A* 49B*. Would this be the same/similar L*a*b* value if one measured a brand new Pantone swatch book? What if your paper had a different white point to the Pantone swatch book? Should/would you go by Pantone values or adjust them taking your stock values into account? Or perhaps split the difference?
Pantone Solid to Process = this has fixed CMYK values, as found in the Solid to Process swatch book. The fixed values for Pantone 173 C from in my CS5 are 0/69/100/4. Whether or not these numbers are good for your printing condition is questionable (what is your printing condition, chances are it is not the same as the Pantone swatch book). Pantone Solid to Process values are what the industry traditionally used, even if the final print condition was not even close to the Pantone swatch book conditions! Many like to quote the values as gospel, they feel safe as the values are published by Pantone and if there are arguments with clients they can say they followed standard values...even if these values are wrong for the print condition at hand!
The "best" or perhaps idealistic answer that I can give is to try to get as close to the L*a*b* values in your CMYK space as you can (within reason, you obviously need a good CMYK profile). This is of course profile/device dependent and the CMYK numbers for ISO Coated are different to say ISO Newsprint. As one can make the target L*a*b* values using different ratios of CMY to K, one has a lot of leeway with how one may mix certain CMYK colours to hit the same L*a*b* values, so there is no single correct answer, in my opinion.
If I recall correctly, this library change took place back in Photoshop 7 or CS. Before this change, there were only Solid to Process library files. One has to be careful whether one selects from the Solid Coated or Solid to Process library in the various creative suite apps. Adobe licence the Pantone library files, and the same library should match in each Creative Suite application. Obviously Pantone change/update their L*a*b* values over time, so Photoshop 7 L*a*b* values may not match CS5 values - even though both reference the same Pantone colour number.
Adobe "only" display simplified integer L*a*b* values for Pantone Colours (sometimes called ICC L*a*b* values?). Other software, such as found in proofing RIPs tend to licence/use more "accurate" values that are not rounded off (CIE L*a*b* values). It probably does not really matter, a*=66 may be hard to compare to a*=66.426!
Regards,
Stephen Marsh
