From: Donald G P. <don...@ni...> - 2012-08-20 18:21:06
|
On 08/20/2012 01:44 PM, Brian Griffin wrote: > It's actually a documentation issue the prompted me to promote the > idea of allowing *any* tcl proc to be available as an expr function. > Take an average developer that is maintaining or enhancing code. If > they run across an expr statement with a function call that they > don't recognize, how do they find it? How many man pages, greps, > etc., do they have do pour through to understand the code? This is a general problem with the extensible parts of Tcl (that is, most of it!) If we give the programmer the power to add new things to Tcl, whether that's a command, or ensemble subcommand, or expr functions, at an equal status with the built-ins of Tcl, the second programmer coming to the interp of facilities from multiple sources has to navigate multiple documentation sources. Packaging conventions, including good use of namespaces, can help, but ultimately a dynamic language is going to face some documentation difficulties. This reality held back ensembles for many years, AIUI. I don't have any brilliant insights into a solution, but it's not a deficit unique to [expr]. -- | Don Porter Applied and Computational Mathematics Division | | don...@ni... Information Technology Laboratory | | http://math.nist.gov/~DPorter/ NIST | |______________________________________________________________________| |