php-math-parser

IMHO precedence climbing (a refactored version of recursive descent) would be much simpler and makes parsing seem almost trivial, since it can be made table-driven: https://www.engr.mun.ca/~theo/Misc/exp_parsing.htm Speaking of parsing simplicity and software design in general, you've coincidentally found a great example of how not to (ab)use OOP... one line of actual work wrapped in a dozen lines of boilerplate and replicated across multiple files with tiny variations: https://github.com/ircmaxell/php-math-parser/tree/master/lib... To put things into perspective, a precedence-climber for 13 levels of precedence from C's grammar, including a tokeniser, fits comfortably in <300 LoC of C, the bulk of that being the tokeniser. PHP's higher level nature (especially around string handling) would probably make a corresponding implementation eve