3 ms·

There's at least one bit that goes further than you've noted and while it has an analogue in vegetarianism, it's much more important to note in FP. It's this:

by tel 11y ago

There's at least one bit that goes further than you've noted and while it has an analogue in vegetarianism, it's much more important to note in FP. It's this: language choice is often a shared/communal one. Therefore, it's required not only to find personal reasons for the choice but also to convince a not insignificant number of your colleagues to make the same choice. This means that some portion of language advocacy is a war for survival in a way that vegetarianism isn't (at least until you have to choose what restaurant to go to with some vegetarians in the crowd—tension grows). (To be clear, I'm just noting this, not trying to be a proponent of it!) So that fuels asshole behaviors because one wants to find a justification not only that an alternative exists and works but that it is actually better. So it's not just a matter of coming off as arrogant but there actually being a seed of truth to that claim. Now what I feel makes PL debates remarkably dangerous is that theres a prevalent idea that PL all comes down to taste---therefore anyone who is trying to make a PL sale is peddling snake oil at least partially. Every action they take is therefore hostile in that it is impossible for them to offer a legitimate cure. In honesty, that idea is completely false. Language really matters a ton and is merely a very complex decision not a meaningless one. Each language offers its own sets of cures and downsides and actual decisions are tricky. So when an FPer comes with arguments, legitimate but complex ones, that FP significantly improves programming experience/outputs then making the trade-analysis to your personal needs is a big undertaking. When, further, they do this with at least some motive to convert you, it's a forced, big undertaking. Finally, when you balk a bit at the potentially massive investment with uncertain return involved in analyzing a new language for a project and this is thrown back at you as a failure to recognize the legitimate value of FP. Well, that's easy to see as asshole-y. What makes a language community fun? Lots of ways to get started, lots of positive, non-forced examples of benefits, and a community oriented around perfecting their own trade not proselytization. Once you start heading out of those bounds you pick up assholes.

vanderZwan 11y ago
> There's at least one bit that goes further than you've noted and while it has an analogue in vegetarianism, it's much more important to note in FP. It's this: language choice is often a shared/communal one. Well, so is food. I'd say you just gave a good argument for why the analogy is even more on point that it might appear at first!
pron 11y ago
> Language really matters a ton I'm not sure that is well-established at all. I mean, languages probably do matter, but just how much we don't know. Also, the choice of language may well matter more due to availability of tooling (debuggers, profilers, IDEs, ease of deployment etc.) and other extra-linguistic features (GC, separate compilation, dynamic linking) than actual language features. Even measuring outputs is very hard. The average lifetime of a codebase is roughly 10 years, so the development costs must be totalled over that entire period. The cost structure however, is very different in the first year and in the eighth. I don't think we have good answers to any of the questions: how much languages matter, how they matter and why. It is precisely because we don't know that we can argue so much :) Some of the PL discussions do, however, make me sad because while I believe the choice of a language matters (though I'm not sure that by a whole lot, and when it does it's mostly due to extra-linguistic concerns), the choice of algorithms matters so much more. I wouldn't want young developers to think that the abstractions they use to express their algorithms is as important as the algorithms themselves (as long as the code is reasonably maintainable). Also, I don't want young developers to equate software with the code the program is written in. A useful, efficient running program is much more than its code, and sculpting code is not the sole means of achieving high-quality software.
- tel 11y ago
  Oh, I'm completely happy to roll all of "tooling, libraries, community" into "language". It matters a ton which one you choose! But we're not disagreeing here. The relative merits of all these factors are expensive as hell to compare. I'd love to have a longer discussion with you sometime around abstractions and algorithms, though. I think I have a few points on the side of abstractions that pass far beyond mere maintainability, but I also really empathize with your argument. So, it'd be interesting to flesh it all out.

by tel 11y ago

vanderZwan 11y ago

> There's at least one bit that goes further than you've noted and while it has an analogue in vegetarianism, it's much more important to note in FP. It's this: language choice is often a shared/communal one. Well, so is food. I'd say you just gave a good argument for why the analogy is even more on point that it might appear at first!

pron 11y ago

> Language really matters a ton I'm not sure that is well-established at all. I mean, languages probably do matter, but just how much we don't know. Also, the choice of language may well matter more due to availability of tooling (debuggers, profilers, IDEs, ease of deployment etc.) and other extra-linguistic features (GC, separate compilation, dynamic linking) than actual language features. Even measuring outputs is very hard. The average lifetime of a codebase is roughly 10 years, so the development costs must be totalled over that entire period. The cost structure however, is very different in the first year and in the eighth. I don't think we have good answers to any of the questions: how much languages matter, how they matter and why. It is precisely because we don't know that we can argue so much :) Some of the PL discussions do, however, make me sad because while I believe the choice of a language matters (though I'm not sure that by a whole lot, and when it does it's mostly due to extra-linguistic concerns), the choice of algorithms matters so much more. I wouldn't want young developers to think that the abstractions they use to express their algorithms is as important as the algorithms themselves (as long as the code is reasonably maintainable). Also, I don't want young developers to equate software with the code the program is written in. A useful, efficient running program is much more than its code, and sculpting code is not the sole means of achieving high-quality software.

tel 11y ago
Oh, I'm completely happy to roll all of "tooling, libraries, community" into "language". It matters a ton which one you choose! But we're not disagreeing here. The relative merits of all these factors are expensive as hell to compare. I'd love to have a longer discussion with you sometime around abstractions and algorithms, though. I think I have a few points on the side of abstractions that pass far beyond mere maintainability, but I also really empathize with your argument. So, it'd be interesting to flesh it all out.