The tech world is always changing…except when it comes to programming languages, where the top three has barely moved for almost 20 years according to the latest rating from TIOBE’s Programming Community Index

The latest index, for April 2019, showed Java in the top spot, with a rating of 15.035 per cent, down 0.74 per cent on the year.

Number two was C, on 15.076 per cent, up 0.49 per cent, while C++ climbed 1.62 per cent to 8.838 per cent, take third place. This pushed Python into fourth place, though it grew 2.36 per cent, and had taken third place in the previous month.

TIOBE itself said the month on month shift between Python and C++ was “certainly not because Python is in decline.”

In fact, it said, “Python is scoring all time highs almost every month. It’s just that C++ is getting more and more popular.” That said, C++ is still far off its peak “popularity in the beginning of this century” (does that make feel you old?) when it was scoring north of 15 per cent.

The top five is rounded out by Visual Basic .NET, with 5.795 per cent, down 1.75 per cent. Skimming through the next five languages, JavaScript jumps out with 2.505 per cent, pulling it up a place on the year. PHP slips from seventh last year to eight this year, while Assembly language jumps from 14th place a year ago to 10th place.

Taking a really long view – which you can with TIOBE’s index – Java has held its top spot since at least 2004, while back in the last millennium it took 12th place in 1999. Both C and C++ have been in the top three since at least 1989 – suggesting the fall of the Berlin Wall had zero effect on programmers’ language choices. However, showing some fashions do change, Pascal now languishes at 207, while in 1994 it was in third place and as recently as 2014 it was in the top 20.

The figures show some stark differences from RedMonk’s January programming language rankings, which put JavaScript in the top spot, with Java placed number 2, and Python number three, with PHP then C# making up the top five.

Needless to say, the two surveys use different, though both slightly controversial methodologies, with RedMonk using a mix of GitHub and Stack Overflow data, while TIOBE‘s is based on the number of skilled engineers worldwide, courses and third party vendors, as deduced via a range of search engines.