I met Simon on a CS conference for schools in Swansea around 2012. I was a nobody but he made the time to talk to me and listened to my interests in programming. He was so nice and personable - a very humble person who takes the time to build others up. A great guy who has a gifted mind. I was like wow I spoke to Simon :)
I'm even going further: One of the greatest minds in Computer Science. His enthusiasm is such a beautiful infection. I can only recommend each Junior scientist to view his talk about "how to write a good paper".
38:00 I actually prefer Strings to Template Haskell, because Template Haskell changes between GHC versions, because format Strings are more similar to the target code and because you can watch the result and ship code that does not need Template Haskell, at all, so it is very portable between GHC versions.
23:00 Specifically for GHC: Yes, I also find it too monadic, and too monolithic. It would be cool to use individual parts like the parser, the type checker, the code generator for other tools and projects. Even the 'base' library is bound to GHC. Instead there should only be a small 'ghc' library.
53:00 why has Haskell persisted without growing explosively? I strongly suspect that's because it's supported by a core of academic adherents. R probably has a similar profile.I think there are a bunch of reasons it's a good academic language (and maybe not a great industrial language for most purposes), but without getting into detail, that's the hypothesis I'd pursue.
Am still a Haskell novice but the problem I heard is not so much with Haskell but with Cabal and Stack, and the gazillions of unmaintained (dangling) packages.
Could I talk to this guy? I'm working on a computer architecture that is like nothing ever seen on earth before! On the level of assembly programming, it's still a completely different feel... not for the fact it's different, hard to explain. The memory (more than just memory) structure is insanely different, some of the benefits are power savings, very literally unlimited simultaneous Parallelism (not a joke), oh and my architecture doesn't require a cache, but 6ou could simulate one if you wanted to lol. I don't know haskell but I know of it and it's importance. If you could write a haskell compiler for my architecture that would be great, and you won't see any reduction of execution time.!
Simon certainly isn't gonna write a compiler for you. He's much too busy with GHC you know. But you've barely provided any information whatsoever about your achitecture. If you want to appeal to someone with the knowledge of writing a compiler, you need to advertise your architecture like a programmer, not a marketer. Buzzwords don't really do anything. Maybe you should write a paper on your architecture. I'd read it.
Its certainly interesting. Of course being a co-creator of Haskell, Simon is naturally going to see things with rose tinted glasses. The reality is that Haskell is niche and will pretty much stay that way. However there are some really fantastic ideas in Haskell which are "trickling" down to mainstream languages and I think that's Haskell's main strength to be a "laboratory" for cutting edge language and type design exploration. Things that prove useful can be then implemented and adopted by the wider industry. I don't see that as a negative relationship but rather a pragmatic and mutually beneficial arrangement. Of course the Haskell fans will lament otherwise saying that its used in industry and that its perfectly fine as a general purpose language etc. While there may be a few instances of industrial use in comparison to even to any of the top 10 languages its miniscule to the point of irrelevance.
@@SEKUNHO Haskell is around 30 years old, if look at the trajectory it's essential remained in permanent stasis. The numbers simply do not lie. Now in terms of why I think it will continue with the same stasis as always has is because: It's a niche language with a very small community, in order for languages to grow the community needs to grow dramatically, something that not only has never happened with Haskell, it doesn't look like it will happen any time soon. As a result of not getting "fresh faces" into the community, this means that the rate of growth of libraries remains stagnant. Essentially the rate of growth determines and underpins the languages entire ecosystem,.e.g libraries, tooling, documentation, and support. Secondly, mainstream languages have already adopted many Haskell/FP features from pattern matching to even things like Maybe/Results Monads as well as having type classes and ADTs etc So combining stagnant growth rate as well as mainstream already adopting FP features there is zero commercial incentives for companies wanting or even needing to make very risky and hugely costly migration away from they're using today.
@@Anhar001 1) There are currently 66.7 thousand members on the Haskell subreddit. That's 1/3 of the size of the C++ subreddit and over 1/4 the size of the Java subreddit. To suggest that Haskell is a niche language is simply ridiculous. Furthermore, the hype around Haskell is only growing and I would posit that there has never been as many people and enterprises adopting Haskell as there are now. 2) Python is also 30 years old and until a few years ago was widely regarded as a joke language that was good for nothing more than writing shell scripts. 3) While many languages have adopted features from Haskell and continue to do so, suggesting that such a trend will make Haskell irrelevant is as absurd as suggesting that Java would do the same to C++. 4) You seem to be under the impression that Haskell's features are stagnant. This couldn't be further from the truth. Thanks to the ease with which users can write their own language extensions Haskell continues to introduce new features at a blistering pace. 5) For certain use cases, Haskell's libraries are second to none. Text parsing and web servers immediately come to mind though there are a great many projects currently underway to extend Haskell into new frontiers. I'm currently working on a Django codebase and if I had my way, I would replace it all with Servant or Yesod in a heartbeat. Out of all the web frameworks I've used, my experience with Haskell has consistently out-performed the competition.
@@allseeingeye93 The Haskell hype peaked at around 2014, its pretty stagnant now. The only real places where there is some growth is Blockchains and Crypto Currency e.g Cardano etc. But outside of that no one cares for Haskell. Using Reddit members as a measure it utterly weak! that doesn't mean you have 66K Haskell developers LOL even so if you want to play that game Python sub as 800K. If you find Haskell useful more power to you, just realise out there in the real world Haskell is so niche even VBA is more desired then Haskell.
It is a common misconception that adding more and more cool features makes a language better. Security means that certain things are not possible and that you get some warranties about code you see. In this sense it can be better to miss some features.
The fact that Simon hasn't yet convinced any big corporations to adopt Haskell for 30 years, compared to other opensource language such as Rust, shows it's a failure.
It depends on what failure means for you. If you go to 17:36, you might learn a bit about Haskell's philosophy and why Haskell hasn't failed in what it wants to achieve. Additionally, Haskell has, of course, been used (and is being used) by Facebook, GitHub, and others.
@@gintsd2 I wouldn't keep boasting of Facebook's trivial use of Haskell...and if it was so great then why they haven't expanded to more projects other than the spammer? I'm a big fan of ML languages and I really hate to see the slow death of the only living ML language which might quicken its pace with the success of Rust... even FP complete turned to Rust..The Haskell community should have a sense of urgency now..
Rust is just a really good language. It doesn't have to be "this or that". The majority of the haskell community really likes rust. Haskell's goal has always been more about influence than anything else. Rust is such a beautiful testament to said influence. I'd be surprised if SPJ ever actually tried to convince big corporations to adopt haskell - but the influence and impact of haskell is felt everywhere. Everything on the mainstream is picking up ideas straight from haskell. Not just mainstream either, I'm sure you're aware already. Haskell has been the foundation of almost every lang born from the functional community within the last couple of decades. What is this, if not success, at its truest meaning?
I met Simon on a CS conference for schools in Swansea around 2012. I was a nobody but he made the time to talk to me and listened to my interests in programming. He was so nice and personable - a very humble person who takes the time to build others up. A great guy who has a gifted mind. I was like wow I spoke to Simon :)
One of the greatest minds behind Haskell. Thanks for sharing!
I'm even going further: One of the greatest minds in Computer Science. His enthusiasm is such a beautiful infection.
I can only recommend each Junior scientist to view his talk about "how to write a good paper".
Amazing! Thank you!
Very interesting! WOW amazing!
This is awesome
38:00 I actually prefer Strings to Template Haskell, because Template Haskell changes between GHC versions, because format Strings are more similar to the target code and because you can watch the result and ship code that does not need Template Haskell, at all, so it is very portable between GHC versions.
I would like to have a 'let ... in' in spreadsheet calculators.
Why will Simon leave Microsoft Research and where will he go to?
Does anyone know who he's talking about when he says "Richard Paulson is teaching Haskell in secondary schools" or what book this might be?
23:00 Specifically for GHC: Yes, I also find it too monadic, and too monolithic. It would be cool to use individual parts like the parser, the type checker, the code generator for other tools and projects. Even the 'base' library is bound to GHC. Instead there should only be a small 'ghc' library.
56:00 A really modular GHC would also help to write competing compilers by substituting only parts of GHC.
Очень интересно!
53:00 why has Haskell persisted without growing explosively? I strongly suspect that's because it's supported by a core of academic adherents. R probably has a similar profile.I think there are a bunch of reasons it's a good academic language (and maybe not a great industrial language for most purposes), but without getting into detail, that's the hypothesis I'd pursue.
the god of haskell
Am still a Haskell novice but the problem I heard is not so much with Haskell but with Cabal and Stack, and the gazillions of unmaintained (dangling) packages.
Could I talk to this guy? I'm working on a computer architecture that is like nothing ever seen on earth before!
On the level of assembly programming, it's still a completely different feel... not for the fact it's different, hard to explain.
The memory (more than just memory) structure is insanely different, some of the benefits are power savings, very literally unlimited simultaneous Parallelism (not a joke), oh and my architecture doesn't require a cache, but 6ou could simulate one if you wanted to lol.
I don't know haskell but I know of it and it's importance. If you could write a haskell compiler for my architecture that would be great, and you won't see any reduction of execution time.!
Blog it
Simon certainly isn't gonna write a compiler for you. He's much too busy with GHC you know. But you've barely provided any information whatsoever about your achitecture. If you want to appeal to someone with the knowledge of writing a compiler, you need to advertise your architecture like a programmer, not a marketer. Buzzwords don't really do anything. Maybe you should write a paper on your architecture. I'd read it.
I would be happy to write a compiler for you
how is he charismatic
is he not a programmer?
Its certainly interesting. Of course being a co-creator of Haskell, Simon is naturally going to see things with rose tinted glasses. The reality is that Haskell is niche and will pretty much stay that way. However there are some really fantastic ideas in Haskell which are "trickling" down to mainstream languages and I think that's Haskell's main strength to be a "laboratory" for cutting edge language and type design exploration. Things that prove useful can be then implemented and adopted by the wider industry. I don't see that as a negative relationship but rather a pragmatic and mutually beneficial arrangement. Of course the Haskell fans will lament otherwise saying that its used in industry and that its perfectly fine as a general purpose language etc. While there may be a few instances of industrial use in comparison to even to any of the top 10 languages its miniscule to the point of irrelevance.
Why do you think Haskell will remain as a niche language?
@@SEKUNHO Haskell is around 30 years old, if look at the trajectory it's essential remained in permanent stasis.
The numbers simply do not lie.
Now in terms of why I think it will continue with the same stasis as always has is because:
It's a niche language with a very small community, in order for languages to grow the community needs to grow dramatically, something that not only has never happened with Haskell, it doesn't look like it will happen any time soon.
As a result of not getting "fresh faces" into the community, this means that the rate of growth of libraries remains stagnant.
Essentially the rate of growth determines and underpins the languages entire ecosystem,.e.g libraries, tooling, documentation, and support.
Secondly, mainstream languages have already adopted many Haskell/FP features from pattern matching to even things like Maybe/Results Monads as well as having type classes and ADTs etc
So combining stagnant growth rate as well as mainstream already adopting FP features there is zero commercial incentives for companies wanting or even needing to make very risky and hugely costly migration away from they're using today.
@@Anhar001
1) There are currently 66.7 thousand members on the Haskell subreddit. That's 1/3 of the size of the C++ subreddit and over 1/4 the size of the Java subreddit. To suggest that Haskell is a niche language is simply ridiculous. Furthermore, the hype around Haskell is only growing and I would posit that there has never been as many people and enterprises adopting Haskell as there are now.
2) Python is also 30 years old and until a few years ago was widely regarded as a joke language that was good for nothing more than writing shell scripts.
3) While many languages have adopted features from Haskell and continue to do so, suggesting that such a trend will make Haskell irrelevant is as absurd as suggesting that Java would do the same to C++.
4) You seem to be under the impression that Haskell's features are stagnant. This couldn't be further from the truth. Thanks to the ease with which users can write their own language extensions Haskell continues to introduce new features at a blistering pace.
5) For certain use cases, Haskell's libraries are second to none. Text parsing and web servers immediately come to mind though there are a great many projects currently underway to extend Haskell into new frontiers. I'm currently working on a Django codebase and if I had my way, I would replace it all with Servant or Yesod in a heartbeat. Out of all the web frameworks I've used, my experience with Haskell has consistently out-performed the competition.
@@allseeingeye93 The Haskell hype peaked at around 2014, its pretty stagnant now. The only real places where there is some growth is Blockchains and Crypto Currency e.g Cardano etc. But outside of that no one cares for Haskell. Using Reddit members as a measure it utterly weak! that doesn't mean you have 66K Haskell developers LOL even so if you want to play that game Python sub as 800K.
If you find Haskell useful more power to you, just realise out there in the real world Haskell is so niche even VBA is more desired then Haskell.
It is a common misconception that adding more and more cool features makes a language better. Security means that certain things are not possible and that you get some warranties about code you see. In this sense it can be better to miss some features.
The fact that Simon hasn't yet convinced any big corporations to adopt Haskell for 30 years, compared to other opensource language such as Rust, shows it's a failure.
It depends on what failure means for you. If you go to 17:36, you might learn a bit about Haskell's philosophy and why Haskell hasn't failed in what it wants to achieve.
Additionally, Haskell has, of course, been used (and is being used) by Facebook, GitHub, and others.
@@gintsd2 I wouldn't keep boasting of Facebook's trivial use of Haskell...and if it was so great then why they haven't expanded to more projects other than the spammer? I'm a big fan of ML languages and I really hate to see the slow death of the only living ML language which might quicken its pace with the success of Rust... even FP complete turned to Rust..The Haskell community should have a sense of urgency now..
Rust is just a really good language. It doesn't have to be "this or that". The majority of the haskell community really likes rust. Haskell's goal has always been more about influence than anything else. Rust is such a beautiful testament to said influence. I'd be surprised if SPJ ever actually tried to convince big corporations to adopt haskell - but the influence and impact of haskell is felt everywhere. Everything on the mainstream is picking up ideas straight from haskell. Not just mainstream either, I'm sure you're aware already. Haskell has been the foundation of almost every lang born from the functional community within the last couple of decades. What is this, if not success, at its truest meaning?
No, he always said we must avoid success at all costs.
@@amigalemming He could at least convince Microsoft to model F# after Haskell instead of OCaml