Ask any marketer — it's (almost) never the marketing. Or, rather, it's never the marketing of the "hype and subterfuge" kind. No product dominates the market fo

Ask any marketer — it's (almost) never the marketing. Or, rather, it's never the marketing of the "hype and subterfuge" kind. No product dominates the market for decades due to hype and subterfuge, let alone hype and subterfuge deployed decades earlier. It's always the product. It's just that what matters about the product to most people is often not what matters to some minority that insists that “technical excellence” is defined by their own preferences. In the famous VHS vs. Betamax war, the technically superior product won, it’s just that it was superior in matters that appealed to many and inferior in matters that appealed to few. The few famous cases where it was the marketing (diamond rings, maybe?) stick in our mind because they're the exception, not the rule. Blaming marketing problems is often an excuse to overlook product problems. Second, the problem with the message of "code written in Haskell is more correct than code written in other [more popular] languages," isn't that the industry doesn't care about correctness. Correctness translates to cost (to a degree), and the ability to write software as correct as it needs to be more cheaply is actually quite compelling. Rather, the problem with it is that it's simply not true (or, rather, hasn't been shown to be true), certainly not to any degree big enough to matter. That, in and of itself, does not make it a terrible message; after all it's just harmless hype, right? What makes it a terrible message is that it is very clearly not (shown to be) true and yet some Haskellers seem to actually believe it, drunk on their own kool-aid, which makes you think of them as out of touch with reality. At best, it's an emotion (it "feels" more correct), which you've mocked, rather than any sort of "critical thinking," that you present as the desirable standard. Third, the mistake in the perception of the economics of software is not the importance of correctness — that can have an easily measurable impact. It's in the importance of code. Programmers in general, but Haskellers in particular, are obsessed with code. As a result, they grossly overestimate its importance to software. Ask yourself how long it takes you to produce, say, 1000 lines of code of sufficient quality, and then look at a software produced at your company and divide the size of the codebase by the number you came up with. You'll find that the true cost is several times that. The "pure" cost of producing code is not where most of the cost of developing software is. Maybe it's in what you mockingly call "emotions", but it certainly isn't in programming-language semantics. Also, before presenting managers' priorities, it's better to actually speak to actual managers; they just might have some useful insight here, plus, it's what a culture of technical excellence demands, no? All of that means that unless your language has some clear and direct effect on the bottom line — which, these days, usually means it has good performance and/or observability, a good ecosystem, or a monopoly of sorts on some distribution platform -- there is no "message" that can make it popular, other than "it's easy and it's fun." While perhaps not a sufficient requirement for success, this can create a supply of qualified developers that at least won't put your language at a disadvantage. And here's another tip: if your retrospective introspection isn't painful, doesn't discuss any serious and conspicuous flaws with the product, and still paints you as being somehow right after all and everyone else as being somehow wrong — it's probably not true. Worse, it seems untrue. And not calling your own practice one of "engineering excellence" and others' as "throwing enough bodies" can't hurt, either. In other words, I don't know what a good marketing message for Haskell would be, but this post is a good example of a bad one.