I got what you were saying. That certainly may be case, depending on application. It looks in this case like its just definition files being loaded into memory,

I got what you were saying. That certainly may be case, depending on application. It looks in this case like its just definition files being loaded into memory, and if they need to compare that version more than once, doing it as a single operation binary comparison will be significantly faster than continually comparing a UTF-8 string (SIMD aside)