The stdlib project covers algorithms that may someday be incorporated in the language. Would it make sense to create a separate project, maybe called extlib for Extended Library, for algorithms of broad interest but which are not expected be part of the standard? It could follow the same coding conventions as stdlib and would use features of it when appropriate. Anything not envisioned for stdlib but in the Numerical Recipes, NAG, or IMSL libraries would be a possible algorithm.
In the field of quantitative finance, for example, I would like to see codes for the Black-Scholes option pricing formula (which is in NAG), for computation of the tangency portfolio, and for Cross-Validated Covariance Matrix Estimation, but they are too specialized for stdlib.
Rather than a single library, I wonder if it is better to develop such applications stand-alone, but use fpm to make it easy for others to use, and fpm-registry to make it easy to discover?
The stdlib project covers algorithms that may someday be incorporated in the language. Would it make sense to create a separate project, maybe called extlib for Extended Library, for algorithms of broad interest but which are not expected be part of the standard? It could follow the same coding conventions as stdlib and would use features of it when appropriate. Anything not envisioned for stdlib but in the Numerical Recipes, NAG, or IMSL libraries would be a possible algorithm.
In the field of quantitative finance, for example, I would like to see codes for the Black-Scholes option pricing formula (which is in NAG), for computation of the tangency portfolio, and for Cross-Validated Covariance Matrix Estimation, but they are too specialized for stdlib.