Talk:Semi-serious language list

Can hypercomputable languages be added? --PkmnQ (talk) 14:56, 22 April 2025 (UTC)

AI

should AI-generated languages be on this list? —aadenboy (talk|contribs) 02:17, 11 December 2025 (UTC)

One of the problem with attempts to generate an esolang using AI is that you typically end up with about 10 different specifications that all contradict each other, making it hard to determine what the language actually is. So it can be hard to tell whether the result of such an operation even is a language or not (and if so, which language it is), which in turn makes it hard to determine whether the rules for the list are met or not. I think that at the least, there should be a requirement that the content of the wiki page matches the content of the external resource or the behaviour of the interpreter (minor mistakes are OK, as is leaving details out especially for complex languages – but hallucinations probably aren't OK). For this particular language, I'm having trouble working out whether it complies with the rules or not (because it can be so hard to find anything substantive in an AI-generated repository). So I don't really have an answer (and would appreciate input from other people). --ais523 02:26, 11 December 2025 (UTC)

Turing completeness

Why must they me turing complete? I see no logical reason for it to be forced. Many languages can be semi serious without being turing complete, so I see no good reason. And it can, for certain languages, be incredibly hard to prove. --Yayimhere2(school) (talk) 05:41, 11 December 2025 (UTC)

I've been considering changing that rule (it wasn't me who created the rule originally). But I think a) ethere should at least be a usability-for-computation requirement (i.e. you could actually use the language to write a range of useful programs), and b) the page should be organized by computational class (i.e. separate sections for the Turing-complete languages, the LBAs/PDAs, the "usable but unclear whether or not it's Turing-complete" languages if we even have any of those, and the weird-but-precisely-defined computational classes like ELEMENTARY and Nellephant). Perhaps usability-unknown languages (that obeyed the other restrictions) could be added in their own section. I do, however, think that things like example-based languages, and conceptual languages which have 0 or 1 valid programs should be excluded from the list, as you can't meaningfully program in them – currently those languages are excluded by the Turing-completeness rule, and if the rule weren't there, they'd have to be excluded a different way. (Likewise, finite state machines should probably be excluded unless programming them is more interesting than "write out all the possible inputs and the corresponding outputs" – perhaps the minimum strength required should be that of a finite state transducer.) I'd be interested to hear whether other people have feedback. --ais523 05:55, 11 December 2025 (UTC)

I'd say perhaps is more powerful than an FSA? There is also the possibility of putting the requirement that some amount of example programs are written. I dont have a very good precise idea, but I certainly think putting away non-TC languages is a not good restriction. I think maybe the restriction "must be able to do varying computation". Like for example, doing a fizzbuzz and a truth machine would be varying, as it takes pretty different paths to make. But like 99 bottles of beer, fizzbuzz, and looping counter, arent varying(they all just count upwards forever). --Yayimhere2(school) (talk) 06:08, 11 December 2025 (UTC)