On Thursday, 26 March 2015 at 19:37:30 UTC, Ola Fosheim Grøstad wrote:
On Thursday, 26 March 2015 at 19:16:54 UTC, bachmeier wrote:
You're making a big assumption about which programmers and projects count and which don't. I wonder if outside of Google

It doesn't matter what the programmers think, what matters is how the development environment affects the project in measurable terms. Having all kinds of features does not necessarily benefit projects. That's the difference between a fun toy language and one aiming for production and maintenance.

Programming is - for now - still a human activity, and what is important in human activities may not always be measured, and what may be easily measured is not always important. That doesn't mean one should throw away the profiler and go back to guessing, but it does suggest caution about adopting the prestigious techniques of the natural sciences and applying them to a domain where they don't necessarily fully belong.

I say this as someone coming from the financial markets, where we have all experienced quite recently the effects of mistaking being quantitative for thinking soundly - what happened ought not to have been a surprise, and of those who saw 2008 coming and spoke publicly about it, I don't think a single one based their view on the quant especially. Yet the field of macroeconomics is much more fully developed than that of assessing programmer productivity and quality of output.

It is not scientific to depend on an approach that has not yet proven itself in practical terms over the course of time and in different environments.




Reply via email to