The "proj" system was made available in the early 1990's and has undoubtedly been used in both open and proprietary applications as is allowed by its public domain origins and subsequent MIT/X11 license. In nearly all cases, use of proj has been without with my participation or consultation, however, there has been one important exception and I want to make it clear of my disassociation with that operaton.
A distribution of 'proj' (see this site) is available that represent coding practices that compromised the quality of 'proj,' its management and use. For several years I was associated with this group's (mailing list archive) and was a persistant nuisance complaining from the year 2000 of the coding abuse but to no avail. I finally decided that for my mental health it was time to get away from it.
What is the problem? The problem relates to the issue of interweaving into one software collection two separate operations and disciplines: datum conversion and cartographic projections. Indeed, in many cases both operations are important in GIS applications and chart making but the operations are distinct and independent of each other and it is very poor programming practice to tightly combine the processes. That is, we do not alter the math library sine function to somehow facilitate the use of 'sin' in an application.
Why make such an issue about it? The principle reason for making such a big deal over this problem is that it makes the maintenance of the software more difficult and thus more prone to errors and more difficult to include enhancements and updates. People are also more likely to be expert in one or the other aspect of the problem and they should not be burdened with trying to wade though sections of code that have nothing to do with with their background. All the two disciplines should be concerned with is the interfacing between their sections.
After 9 years of bashing my head on this problem I finally gave up.