4 ms·

Do you guys think that closed source general-purpose programming languages have any future? In the late 90s and early 2000s both Java and C# were closed source.

Do you guys think that closed source general-purpose programming languages have any future? In the late 90s and early 2000s both Java and C# were closed source. Still, they became highly successful. Today many people wouldn't even touch a language or a framework to build a business on unless it's Open Source. It's surprising how the programming landscape has changed in the past 20 years.

pjmlp 7y ago
Yes they do. C, C++, Ada, Pascal, Basic, Oberon compilers for embedded market, in many cases, certificed compilers. Commercial UNIX, mainframe and Windows compilers (VS Community doesn't cover everything) are still a separate product. Sony, Nintendo, MS SDKs for their game consoles. HPC compilers, yeah you can try with clang and gcc, but if you are doing top research, then IBM, Intel, CodePlay, PGI compilers is what you want to be using on that hardware, and they cost pennies versus the total hardware cost. And even in what concerns Apple, I would consider that XCode is somehow included in Apple^s hardware price.
- AnimalMuppet 7y ago
  I think the standard approach is to not add extra constraints. Are you writing general purpose, portable code code? You'd rather not be locked in to a specific vendor's compiler. Are you working in embedded systems? Your code is probably hardware-specific, at least to some degree. You probably aren't going to switch CPUs very often. You probably are never going to switch CPU families. You also probably are never going to upgrade the compiler, still less switch to a different vendor's compiler. So if you aren't going to do that anyway, you don't worry much about a one-vendor-only compiler or language. I think it's a similar story in HPC. You're constrained by your hardware. You get the best compiler for that setup. That means it's hard to change compilers? It's even harder to change your hardware, so you didn't add much of a constraint.
  - pjmlp 7y ago
    Agreed.
thu2111 7y ago
.NET only became open very recently, it's been quite successful. The entire iOS/Apple developer ecosystem is closed source of course. Maybe Swift isn't but nobody uses it outside the Apple platform. You could argue Microsoft is an exception and them opening .NET shows the end of that era though. The "open source or nothing" mentality might change in future. It means there's not much market for developer tools, so everything gets hacked together out of random "frameworks" found on GitHub that are here today, gone tomorrow when the bored developer at Netflix or Google gets their desired promo and moves on. There's no real planning or design, just whatever random stuff tech firms throw in the mix as a side effect of other projects, which is why HTML has virtual reality support but no table or spreadsheet view worth a damn.
- exikyut 7y ago
  Hmmm. Agreeing with the sentiment of your argument while looking at the flip side of the coin, the fruit borne of the ad-hoc lack of coordination you describe is sadly easily overwhelmed by domain domination exerted from committees and bureaucracies with vested interests in specific fields. Which kind of information-leaks that either VR is hype noone really cares about, or that none of said committees (with attendant lumbering bureaucracy) have really noticed it yet. Wait, I just described the same thing two different ways. But yeah, no, it would not do to have The Web℠ Platformᴳ, errrrr, "excel". :P
  - thu2111 7y ago
    Yes, there's a certain robustness that comes from an evolutionary approach. And some of the web's competitors died because they couldn't (in my view) correctly analyse and match the web's strengths, because those strengths were never planned or written down anywhere. Figuring out why the web is strong is sort of like trying to figure out how our own bodies work. Sure, we can see the component parts. How they come together to make the whole a success is often less clear. On the other hand, let's not kid ourselves. The web wasn't actually built by making a melting pot of diverse ideas with an evolutionary process to select between them. The web has always been controlled by one dominant company, and the stuff that gets added (or not) reflects the agenda of their internal committees. The web went from being controlled very, very briefly by Tim BL to being controlled by Netscape who added forms, cookies, JavaScript, CSS etc. Then to Microsoft, who added a pile of other stuff and then froze development for years whilst Mozilla tried to get off the ground. Mozilla added relatively little to the web given the resources invested into Firefox; virtually all their energy went into trying to fix their own architectural mis-fires. Then WebKit came along and very briefly Apple was in charge. That era didn't last long until Google came along and took the crown, which they've had ever since. Unlike Apple and Mozilla, Google actually have added a lot of stuff to the web platform but it reflects a consumer driven agenda. Hence WebVR, WebRTC etc. To the extent the other players exerted any influence on the Chrome team during this time it's - in my view - been mostly negative. Mozilla lost to Chrome because their technical choices just weren't that good, but they've continued to exercise outsized influence on decision making despite dwindling market share. Additionally the lack of any real support for libraries or dependencies in the web platform means many of the new APIs added to HTML can't be challenged by competitors. WebAudio sucks but nobody can make a better one because of the sandbox. In the end the web isn't really the result of ad-hoc coordination so much as a dysfunctional lurching tug-of-war between different groups of engineers who all work down the road from each other at different firms in California (with a detour through Seattle).

pjmlp 7y ago

Yes they do. C, C++, Ada, Pascal, Basic, Oberon compilers for embedded market, in many cases, certificed compilers. Commercial UNIX, mainframe and Windows compilers (VS Community doesn't cover everything) are still a separate product. Sony, Nintendo, MS SDKs for their game consoles. HPC compilers, yeah you can try with clang and gcc, but if you are doing top research, then IBM, Intel, CodePlay, PGI compilers is what you want to be using on that hardware, and they cost pennies versus the total hardware cost. And even in what concerns Apple, I would consider that XCode is somehow included in Apple^s hardware price.

AnimalMuppet 7y ago
I think the standard approach is to not add extra constraints. Are you writing general purpose, portable code code? You'd rather not be locked in to a specific vendor's compiler. Are you working in embedded systems? Your code is probably hardware-specific, at least to some degree. You probably aren't going to switch CPUs very often. You probably are never going to switch CPU families. You also probably are never going to upgrade the compiler, still less switch to a different vendor's compiler. So if you aren't going to do that anyway, you don't worry much about a one-vendor-only compiler or language. I think it's a similar story in HPC. You're constrained by your hardware. You get the best compiler for that setup. That means it's hard to change compilers? It's even harder to change your hardware, so you didn't add much of a constraint.
- pjmlp 7y ago
  Agreed.

thu2111 7y ago

.NET only became open very recently, it's been quite successful. The entire iOS/Apple developer ecosystem is closed source of course. Maybe Swift isn't but nobody uses it outside the Apple platform. You could argue Microsoft is an exception and them opening .NET shows the end of that era though. The "open source or nothing" mentality might change in future. It means there's not much market for developer tools, so everything gets hacked together out of random "frameworks" found on GitHub that are here today, gone tomorrow when the bored developer at Netflix or Google gets their desired promo and moves on. There's no real planning or design, just whatever random stuff tech firms throw in the mix as a side effect of other projects, which is why HTML has virtual reality support but no table or spreadsheet view worth a damn.

exikyut 7y ago
Hmmm. Agreeing with the sentiment of your argument while looking at the flip side of the coin, the fruit borne of the ad-hoc lack of coordination you describe is sadly easily overwhelmed by domain domination exerted from committees and bureaucracies with vested interests in specific fields. Which kind of information-leaks that either VR is hype noone really cares about, or that none of said committees (with attendant lumbering bureaucracy) have really noticed it yet. Wait, I just described the same thing two different ways. But yeah, no, it would not do to have The Web℠ Platformᴳ, errrrr, "excel". :P
- thu2111 7y ago
  Yes, there's a certain robustness that comes from an evolutionary approach. And some of the web's competitors died because they couldn't (in my view) correctly analyse and match the web's strengths, because those strengths were never planned or written down anywhere. Figuring out why the web is strong is sort of like trying to figure out how our own bodies work. Sure, we can see the component parts. How they come together to make the whole a success is often less clear. On the other hand, let's not kid ourselves. The web wasn't actually built by making a melting pot of diverse ideas with an evolutionary process to select between them. The web has always been controlled by one dominant company, and the stuff that gets added (or not) reflects the agenda of their internal committees. The web went from being controlled very, very briefly by Tim BL to being controlled by Netscape who added forms, cookies, JavaScript, CSS etc. Then to Microsoft, who added a pile of other stuff and then froze development for years whilst Mozilla tried to get off the ground. Mozilla added relatively little to the web given the resources invested into Firefox; virtually all their energy went into trying to fix their own architectural mis-fires. Then WebKit came along and very briefly Apple was in charge. That era didn't last long until Google came along and took the crown, which they've had ever since. Unlike Apple and Mozilla, Google actually have added a lot of stuff to the web platform but it reflects a consumer driven agenda. Hence WebVR, WebRTC etc. To the extent the other players exerted any influence on the Chrome team during this time it's - in my view - been mostly negative. Mozilla lost to Chrome because their technical choices just weren't that good, but they've continued to exercise outsized influence on decision making despite dwindling market share. Additionally the lack of any real support for libraries or dependencies in the web platform means many of the new APIs added to HTML can't be challenged by competitors. WebAudio sucks but nobody can make a better one because of the sandbox. In the end the web isn't really the result of ad-hoc coordination so much as a dysfunctional lurching tug-of-war between different groups of engineers who all work down the road from each other at different firms in California (with a detour through Seattle).