2 ms·

TIOBE can remain transparent and also be seriously flawed in its methodology, which has been already questioned for many years. No popularity indicators would b

TIOBE can remain transparent and also be seriously flawed in its methodology, which has been already questioned for many years. No popularity indicators would be entirely free from flaws, but anyone who are aware of TIOBE's possible flaws always quote multiple indicators including PYPL [1] and RedMonk Top 20 [2]. That's how you can remain convincing even with only possibly flawed data sources. While C/C++ will remain an important language for many years, its continuing decline is also clear from those indicators. In fact even the most recent TIOBE has reported the lowest ever position (4th) for C, and that fact was already well known for other indicators: PYPL indicator for C was roughly in decline for a decade, so did its RedMonk ranking as well. They both estimate the current use of given language by looking at popularity of tutorials or questions, while TIOBE estimates the cumulative use of given language. All things equal, TIOBE is going to be systematically delayed compared to others, yet still TIOBE ranking for C is now falling and there is no reason to believe otherwise. [1] https://pypl.github.io/PYPL.html https://pypl.github.io/PYPL.html [2] https://redmonk.com/rstephens/2024/09/12/top20-jun2024/ https://redmonk.com/rstephens/2024/09/12/top20-jun2024/

rramadass 2y ago
What you are pointing out is nothing revelatory. That a statistical index is dependent on data and methodology is a vacuously true statement and not an argument. That there are multiple indexes trying to measure the same thing is also true and not an argument. The point was to show an index which is well respected (criticisms not withstanding) as a counter to silly claims. I personally do not place much stock in these rankings since all of them are flawed in their sampling methodology due to using only publicly accessible indicators like google searches, stack overflow questions, job postings, Github and similar public repositories etc. C programmers on average are more experienced and hence have no need for these. They are already aware of most of the Good/Bad/Ugly about the language and are used to working out problems for themselves and hence don't show up in these metrics. Thus C/C++ rankings might appear to be waning when they are holding steady or rising much more slowly w.r.t. others.

rramadass 2y ago

What you are pointing out is nothing revelatory. That a statistical index is dependent on data and methodology is a vacuously true statement and not an argument. That there are multiple indexes trying to measure the same thing is also true and not an argument. The point was to show an index which is well respected (criticisms not withstanding) as a counter to silly claims. I personally do not place much stock in these rankings since all of them are flawed in their sampling methodology due to using only publicly accessible indicators like google searches, stack overflow questions, job postings, Github and similar public repositories etc. C programmers on average are more experienced and hence have no need for these. They are already aware of most of the Good/Bad/Ugly about the language and are used to working out problems for themselves and hence don't show up in these metrics. Thus C/C++ rankings might appear to be waning when they are holding steady or rising much more slowly w.r.t. others.