3 ms·

Interesting question, because ICT has a tendency to do a progress reset every half decade. Look at things like cpan: Tons of interesting librarys in what is f

Interesting question, because ICT has a tendency to do a progress reset every half decade. Look at things like cpan: Tons of interesting librarys in what is for today the wrong language. Java's maven ecosystem, once a center of open source, is quickly becoming an enterprise-only dead zone. There are plenty of smaller examples, like the tons of dos/qbasic code, the VB6 component ecosystem. javascript is one of the current hotspots, but even there entire sub ecosystems are slowly falling away. Each time we win by having better tools, but we lose by dropping all the optimization already done. And things like lisp or ada show there were better tools than the industry average available in the past, except their ecosystems didn't grow enough and/or died.

theptip 5y ago
My mental model for this process is that whenever a team chooses a new language with a less-complete stdlib (eg moving Perl -> early Python) it’s because there’s a net gain for the specific task at hand, even though you take a productivity hit in some areas. Then if enough developers use the new “better for task X,Y,Z” thing then there is incentive/resource to gradually grow the stdlib capabilities to cover all the old use cases. And in doing so, often some of the initial expressiveness/focus is lost, so there’s now a bit more room for a new focused competitor to come in and repeat the process. Some find this exhausting, or think that it represents “fad chasing” without creating any new value, but I think it’s just what progress/evolution looks like in a community-based system of knowledge like software languages. To your point about lisp or ada being better - using an evolutionary metaphor, it’s possible for fitter genotypes to lose to less-fit ones through random chance. But in these cases it’s hard to tell whether it’s not actually the case that these languages, while better for some users/use-cases, are actually lacking something that makes them suitable for broader usage. (For example - if a paradigm was incredibly productive, but requires you to have a maths PhD, then you’d expect to see some users claiming the tools are better, at the same time as those tools failing to get broader adoption. I’m not making a specific claim here about lisp, just painting some possible fact patterns that could explain the phenomenon.)

theptip 5y ago

My mental model for this process is that whenever a team chooses a new language with a less-complete stdlib (eg moving Perl -> early Python) it’s because there’s a net gain for the specific task at hand, even though you take a productivity hit in some areas. Then if enough developers use the new “better for task X,Y,Z” thing then there is incentive/resource to gradually grow the stdlib capabilities to cover all the old use cases. And in doing so, often some of the initial expressiveness/focus is lost, so there’s now a bit more room for a new focused competitor to come in and repeat the process. Some find this exhausting, or think that it represents “fad chasing” without creating any new value, but I think it’s just what progress/evolution looks like in a community-based system of knowledge like software languages. To your point about lisp or ada being better - using an evolutionary metaphor, it’s possible for fitter genotypes to lose to less-fit ones through random chance. But in these cases it’s hard to tell whether it’s not actually the case that these languages, while better for some users/use-cases, are actually lacking something that makes them suitable for broader usage. (For example - if a paradigm was incredibly productive, but requires you to have a maths PhD, then you’d expect to see some users claiming the tools are better, at the same time as those tools failing to get broader adoption. I’m not making a specific claim here about lisp, just painting some possible fact patterns that could explain the phenomenon.)