Yes I know that computer language popularity rankings are not to be trusted but I came upon the site github.com/toUpperCase78/tiobe-index-ratings.git which has data for the TIOBE ratings and I thought it would be fun to graph Julia’s rating. So for your viewing pleasure, here is the plot:
For comparison, Python, which is in first place has a rating of roughly 20% and C is in second place with a rating of roughly 10%. To be in the top 20, a language needs to have, approximately, at least a 1% rating. Right now (February 2026), Julia is in 29th place.
If anyone is interested in gaming the TIOBE rating, you should list Julia your stickers and merch on Ebay and Amazon marketplace, and add references to Julia algorithms to wikipedia. Each additional posting there is probably worth a few hundredths of a rating point.
OK, don’t do that, but seriously, eBay and Amazon make up nearly half of the rating and the denominators there aren’t that big.
The noisiness of the metric is great for posting about every increase and ignoring every decrease, though
I agree that there are systematics in the data that are not believable. The bigger question for me is if the long term trends in the data (say over a decade or so) are indicative of what is actually happening in the real world. If you factor in the data from @StefanKarpinski, along with the data from PYPL (which also has systematics that are not believable), consistent long-term growth is noticeable. I would not want to make any quantitative statements but I do think this puts the lie to what some doom sayers (no names mentioned) are saying.
Do we have anything better? On the one hand, it seems like people are generally interested in such rankings. On the other hand, it’s often claimed that they’re “just noise”, meaningless, etc. And yet these rankings are still around and still attract some attention. I guess that’s because people generally like various rankings?
Also, nobody seems to deny that they correctly gauge the popularity of Python, C and C++, but when it comes to less popular languages, they suddenly become “noise”, “not believable” and “no way that’s legit”. I think computing rankings based on search engines (TIOBE Index - TIOBE, PYPL PopularitY of Programming Language index) can’t be all noise and the results should be at least somewhat meaningful.
That’s more concerning and shows the real problems of Julia like a lack of good tooling. Julia is becoming more and more a niche language for fields like scientific computing. But it’s losing its attraction as a general purpose language. It needs more than a nice programming language and an efficient compiler to attract software developers.
But the data from different sources, if anything, is showing long term growth. This implies that Julia is attracting people and not just stagnating as a “niche language”.
Simple things like dwarf tools, better api calling for llvm.jl writing the manual api for creating .bc from .ll in julia was a nightmare and llvm-as can just do it, then you think you find the right julia tooling pkg and its been unmaintained for 4 plus years and was bare support barely deserving the pkg name in the first place. Look at DWARF.jl I wish that thing worked would have saved me 5 months of hammering out a custom dwarf extractor.
