from Github Language Stats ~0.08%
from Github Language Stats ~0.08%
Very nice. Here is a relevant graph of Fortran, Julia, MATLAB and R for pull requests:
And it’s actually quite comparable!
Has anybody figured out how this site works or what the x-axis is? I get the same curve for Fortran (the one you’re showing above), no matter what year or quarter I choose.
Fortran fell to 30th in March, with another rating drop as well. I will try to remember to store the monthly data and may post again when the news is good. I doubt that the situation had ever improved as much as the ranking rise indicated or that the recent ranking fall reflects a big deterioration.
| Month | Ranking | Rating | Change YOY |
| Jan. 2021 | 30th | 0.46%
| Sep. 2021 | 17th | 1.01% |
| Oct. 2021 | 18th | 1.08% |
| Nov. 2021 | 19th | 1.19% |
| Dec. 2021 | 17th | 1.04% | +0.59%
| Jan. 2022 | 19th | 0.77% | +0.31%
| Feb. 2022 | 23rd | 0.58% |
| Mar. 2022 | 30th | 0.39% |
Apr. 22 | 31st
May 22 | 30th
Jun. 22 | 26th
In Tiobe’s survey about July, Fortran experienced some improvement compared the earlier months:
| month | rating | rank | |---------+--------+------| | 2022-01 | 0.77% | 19 | | 2022-02 | 0.58% | 23 | | 2022-03 | 0.39% | 30 | |---------+--------+------| | 2022-04 | 0.35% | 31 | | 2022-05 | 0.38% | 30 | | 2022-06 | 0.55% | 26 | |---------+--------+------| | 2022-07 | 0.76% | 18 | | 2022-08 | 0.81% | 19 | | 2022-09 | 1.03% | 15 | |---------+--------+------| | 2022-10 | 0.79% | 19 | | 2022-11 | 0.57% | 24 | | 2022-12 | | | |---------+--------+------| last update: [2022-11-08 Tue]
«Fortran and Matlab entered the top 20 once again at the expense of Lua and Prolog.» as by Tiobe’s July statistics accessed on 2022-07-09 (reference). Because this metric explicitly includes Youtube, I speculate the video FORTRAN in 100 Seconds contributes to this increase. (The video was discussed earlier here.)
To put this a little bit into perspective, I picked some 100 s of … videos of Jeff Delaney’ channel Fireship (1.43M subscribers), namely about Fortran, Julia, Python, Lua, and COBOL. It could be I did not recognize R, or that this topic is not (yet?) considered by him. As by today (2022-07-11):
| language | published | views | likes | |----------+------------------+-------+-------| | Fortran | [2022-06-10 Fri] | 429k | 22k | | Julia | [2022-05-19 Thu] | 332k | 21k | | Python | [2021-10-25 Mon] | 1.07M | 53k | | Lua | [2022-02-14 Mon] | 774k | 48k | | Cobol | [2020-04-11 Sat] | 723k | 32k | |----------+------------------+-------+-------|
Though the video about Fortran was released more recently than Julia, the difference in the numbers of this snapshot is less important than I anticipated.
2022-11-08_Tiobe_index_Fortran.csv (7.3 KB)
Cobol solved the 2-language problem.
Another day in hepatoscopy: Fortran is #22 on PLDB, whatever that means.
Looks like this entry is available on GitHub:
Might be worth to at least update the example snippets to free form.
Seeing Fortran (15) almost tied with MATLAB (14) in the most recent ranking of TIOBE is worth a mention here.
I think its best rank since 2020 was 13th in August 2021. Maybe we can renew that record. The GSoC has triggered a good activity, and the translation process has started again, for example.
This is .true. both in percentage, as well as in the classement (I started updating post 86 on a monthly basis, the .csv includes the Tiobe data all back to June 2016, the first time archive.org recorded a copy of the site).
This is a great result. Is there an explanation for this result? I noticed that there aren’t many new Fortran repos in GitHub or new Fortran questions at Stack Overflow.
The explanation for this is very simple. Tiobe is a completely unreliable benchmark with some obviously incorrect results. The less attention payed to it, the better.
For a very simple example, what remotely believable benchmark would say that Classic Visual Basic (which was deprecated in 2008) was the 11th most popular language last year, had it’s use cut by 2/3rds between October 2021 and March 2022, doubled in popularity by June 2022, and is currently more used than R? If you believe that is true, I have several bridges to sell you.
I do not think new Stackoverflow questions would be a good measure, especially for Fortran, which has been around for a long time. Any drop in Stackoverflow usage is also partly due to the migration of many such questions to this forum, reddit, and other communities (e.g., Intel, IBM, NVIDIA, Google, … Fortran forums).
Probably, much ado about nothing! I believe it measures (imperfectly) the noise we make on the web as a community. Imperfectly, as there is no reason for the lows and highs, for example the low in April 2022. Probably we are just above the noise level and the signal is difficult to detect. Moreover they sometimes change their methodology.
In any case, the TIOBE index is relayed on many sites and the return of Fortran in the top 20 is a good signal for people that may be interested by the language.
I totally agree that the TIOBE index is not accurate. However, the Julia community seems to consider it an important indicator of language adoption, and I do not see you questioning its accuracy when it comes to Julia’s rise in the TIOBE index.
I don’t think I have specifically called out how terrible it is in that thread, but that’s only because enough other people already had. See Thoughts on TIOBE's Language Ranking Methodology - Julia Computing and the majority of the comments in that thread being how tiobe is useless.