In my simple mind, if software has been "released" it is no longer R&D, and "bug fixes" (which should include continuous improvements such as your example) are not research.
I may be way, way wrong though.
That seems too exploitable to pass muster in the court. If you release Beta 0.0.1 of your software after 2 months of development then spend the next 5 years getting it up to version 1.0 that's clearly a development effort not a maintenance effort.
> such as marketing and promotional activities, maintenance activities that do not give rise to upgrades and enhancements, distribution activities
If it leads to a new release, then its software dev. Meaning anything more than a minor patch is going to count.
That's the reason we have courts, to cut through those gray areas.
No. That is why you have auditors who must sign off on your financial books and records. There are fairly strict rules about capitalization of software development. If it is a meaningful number for your firm, then the auditors will review in detail.