I might have taken the content of this article more seriously if it had been titled something like "Why Go Is Not Good for Embedded System Programming" since the original author of the article seems to have more experience in embedded system's domain[1].
Also, the article say "Go is a regression from other modern programming languages" and I find it amusing that a significant numbers of people here taking the article and its claim seriously when it is coming from a student[1] who just finished his 'Computer Programming' university course in Fall 2014[1] (one of the reason, other reason is mentioned in the next paragraph). Don't get me wrong here: I'm not saying that you can't say anything significant when you're a student; what I mean here is that one needs to come up with more detailed explanations and with many more examples which are valid for a wide range of scenarios and use-cases when you make a general statement about a programming language which has been created by some of the highly-respected experts in the field of programming language. The list of problems mentioned in the article are important but they're not very critical for the kind of the system development that Go language has been designed/developed for[2].
Since Go language has been developed for system programming, the term 'system programming' is not restricted to embedded system only (as few people have already mentioned it in different threads here) which are mostly limited to one component (or small number of related components working together). With the advent of internet and IoT, we are forced to develop very large software systems (read, software systems as infrastructures) in order to make next generation of internet and IoT applications possible and usable (talking from business perspective). Development of these new large scale systems bring different kinds of theoretical and practical problems like complexity, concurrency and inefficiencies in system development process (for example; code compilation of large codebase and running regression test suits). And, Go has been specifically build for this new kind of large scale system softwares[2] (competing/working along with some other programming languages in this area).
Here, I would read the initial set of high-level problems which forced Rob Pike and his team to create a whole new language[2], instead of only considering issues/problems which one face while developing a single machine software/hardware program (as I've already said that they're important but they're also not everything). Once I know the strength and weakness of a programming language, I know when I should (or should not) use it, under what circumstances it's the right tool and what advantages/dis-advantages I've to trade off when I use it.
Also, the article say "Go is a regression from other modern programming languages" and I find it amusing that a significant numbers of people here taking the article and its claim seriously when it is coming from a student[1] who just finished his 'Computer Programming' university course in Fall 2014[1] (one of the reason, other reason is mentioned in the next paragraph). Don't get me wrong here: I'm not saying that you can't say anything significant when you're a student; what I mean here is that one needs to come up with more detailed explanations and with many more examples which are valid for a wide range of scenarios and use-cases when you make a general statement about a programming language which has been created by some of the highly-respected experts in the field of programming language. The list of problems mentioned in the article are important but they're not very critical for the kind of the system development that Go language has been designed/developed for[2].
Since Go language has been developed for system programming, the term 'system programming' is not restricted to embedded system only (as few people have already mentioned it in different threads here) which are mostly limited to one component (or small number of related components working together). With the advent of internet and IoT, we are forced to develop very large software systems (read, software systems as infrastructures) in order to make next generation of internet and IoT applications possible and usable (talking from business perspective). Development of these new large scale systems bring different kinds of theoretical and practical problems like complexity, concurrency and inefficiencies in system development process (for example; code compilation of large codebase and running regression test suits). And, Go has been specifically build for this new kind of large scale system softwares[2] (competing/working along with some other programming languages in this area).
Here, I would read the initial set of high-level problems which forced Rob Pike and his team to create a whole new language[2], instead of only considering issues/problems which one face while developing a single machine software/hardware program (as I've already said that they're important but they're also not everything). Once I know the strength and weakness of a programming language, I know when I should (or should not) use it, under what circumstances it's the right tool and what advantages/dis-advantages I've to trade off when I use it.
---
[1]- http://yager.io/resume.pdf
[2]- http://commandcenter.blogspot.de/2012/06/less-is-exponential...