On the other hand, let's not forget what I believe to be the reason for FORTH's demise. FORTH is a very elegant language, with unorthodox ideas. It was invented by Chuck Moore, who is having his own eccentric (and fresh) ideas about how one should program.
The reason FORTH didn't take hold (at least in my own projects) was that it lacked standard libraries for the things which I needed. It expected people to reinvent the wheel (and optimize it to their project's needs) all the time. It didn't take to heart Pareto's Law (80% of the computer time/programmer time/memory requirements/bug expenses of software are in 20% of the code). People should optimize and design their own implementations of data structures only when and where they are critical to the software's performance. For non-critical parts of the software, standard libraries are good enough and should be used.
The morale of the story to hash functions in STL: STL should have provided a standard hash implementation (like Perl does). But the standard implementation should (like implementations of all other STL data structures) have provisions for people to substitute their optimized algorithms when those algorithms are really needed for a specific application
| Author | Omer Zak |
| Work | Hackers-IL Post |
| Published | 2018-03-25 |