I just got back from a very positive strategy meeting with my partner (I do the code and he does the ops and the sales) and this essay struck a chord with me. Why are we making so much progress when I'm a decidedly average programmer? I think it's because I'm a decidedly average programmer in enough languages to know what their strengths and weaknesses are and thus to be able to pick the right tool for the job:
- hardware-accelerated video decoding on the ARM - C obviously, with a small amount of ASM.
- web server - Python using web2py
- web interface prettiness - jquery and javascript
- server processes and admin automation - Python
- core apps - Luajit because it's so fast on the very limited ARM boxes we're using
- customer data munging and reports - emacs lisp because it's right there
- desktop apps for Windows, OSX and Linux - Python again
- some bash and Perl for sysadmin stuff
What Steve doesn't say - although it is implied - is that he knows enough languages at a practical level to make these kinds of judgement calls. And I know for a fact he's a far better coder than I am because I use his Penlight library for Lua every day.
Perhaps Flub can never be a force for change the way Lisp was because modern languages are incorporating enough features from Lisp to bring them close enough to the top of the hierarchy not to matter as much as before.
- hardware-accelerated video decoding on the ARM - C obviously, with a small amount of ASM.
- web server - Python using web2py
- web interface prettiness - jquery and javascript
- server processes and admin automation - Python
- core apps - Luajit because it's so fast on the very limited ARM boxes we're using
- customer data munging and reports - emacs lisp because it's right there
- desktop apps for Windows, OSX and Linux - Python again
- some bash and Perl for sysadmin stuff
What Steve doesn't say - although it is implied - is that he knows enough languages at a practical level to make these kinds of judgement calls. And I know for a fact he's a far better coder than I am because I use his Penlight library for Lua every day.
Perhaps Flub can never be a force for change the way Lisp was because modern languages are incorporating enough features from Lisp to bring them close enough to the top of the hierarchy not to matter as much as before.