The Flub Paradox

The Flub Paradox
steved-imaginaryreal.blogspot.com

Doesn't this definition make the "power" of all Turing-complete languages equal?

> In other words, Lisp was the Flub of its day. A few years ago Flub was Haskell, and recently the new contender appears to be Rust, judging from the buzz.

If the author thinks Haskell and, of all languages, Rust are most powerful, I think he has significantly misunderstood Graham's original essay.

Among other things, they are statically checked with relatively difficult types that can take a long time to get right. That is obviously not the kind of language Graham had in mind for agily prototyping around corporate behemoths.

The author seems to define Flub as "the most popular language among the elite". Lisp was rarely even that, however, and we also have painfully obvious evidence of the most popular not being the most powerful. As such, I think a large portion of this essay hinges on a strawman: he dismantles a definition of Flub that Graham never used.

> The Flub Paradox can be stated like this: although Flub is self-evidently more powerful and productive than Blub, relatively few people use it and its effect on the real landscape of innovation appears minimal.

Is this really a paradox anymore? We have so many examples of inferior products dominating through inertia that I've completely lost my faith in humanity at this point.

I could add a few more points, but I think that would only be unnecessarily negative. I do like some points from the essay. I agree that we should take another look at Paul Graham's power continuum: for instance, the power of Lisp scales with lines of code (with low LOC, macros don't really enter the equation, and using a "batteries included" language with comprehensive libraries is probably faster). Haskell is easier than many agile languages to refactor, another important dimension to consider. And so on.

Of course not, and, sure, Flash was a better alternative for a lot of time (but mostly for things outside of JS like HTML5 capabilities)

C is not great as well, but between something good that doesn't do what is needed and a bad one that does guess what is used in the end.

When it appeared, C was a revolution in programming languages. It offered a combination of performance and expressiveness that simply outclassed anything else available (still true today, 43 years later, if more narrowly). It instantly became the most popular and held that position to this day, becoming a major influence on virtually every other popular programming language that's come since.

Now, of course, C is much better suited for, say, writing an operating system than it is for writing security-intensive internet services. (A concept which did not exist in 1972). But, all things considered, it was as close to perfection as a new programming language has ever come at the time of its introduction.

JavaScript is not even in the same realm; much superior, existing scripting languages (Python, Ruby, even Perl or some Lisp dialects) were passed over in favor of a deliberately half-assed effort to advertise Java with a language that shared a name and a superficially similar syntax, which (almost accidentally) became the only VM that any web browsers implement.