I have a different perspective

On Thursday, April 5, 2018 at 10:26:19 AM UTC-7, bingj...@gmail.com wrote:
>
> Almost 10 years golang appears in the world. 10 years is not a short 
> duration. I think if it is not popular until 2020, it will never be popular.
>
> Golang is designed for cloud and internet areas. Really?
>
> The creators of golang have a lot of experience in C and C++. And golang 
> borrows features from C and C++. But C and C++ do not fit the requirements 
> of cloud and internet areas.
>
> Let's look at two popular programming languages java and php. What is the 
> most important features of these two languages? Simple, ugly but 
> practical... I find one feather: they are both not just programming 
> languages but also platforms. They are almost the same in Windows and 
> Linux. That's why java and php are very popular in recent days.
>
> C and C++ are just pure programming languages, not platforms. On Unix and 
> Windows, C and C++ are very different. A developer of windows C++ is not a 
> developer of UNIX C++, and a Linux C developer is not a Windows C developer.
>
> If golang wants to be widely used by developer all over the world before 
> 2020, it must learn some thing from java and php, must be a 
> programming-language-is-a-platform.
>
> Until now, programs written in golang still does not have binary 
> distribution format like jar, dll or so. People have to share libraries by 
> source code. It is so foolish.
>

As my company is a potential vendor of commercial Go libraries, I've been 
interested in this very question. I believe your assertion is incorrect. It 
is possible, but not pleasant, to share pre-built Go libraries. The 
challenge is that you have to include stub code to make it work, and make 
sure the timestamps on the stubbed-out code don't get updated. Of course, 
you also have to ship the binaries specific to the platform that the target 
developer is going to use, but that is possible. The advantage of shipping 
the stub code is that you can get all the benefits of existing IDE tooling 
that work off of the source to generate documentation, and understand the 
meaning of your code and how it is used.
 

> Yes, Golang is very like C and C++, which are only pure programming 
> language, But this times, we need "language as/is platform" technologies, 
> just like php and java.
>

What is meant by "language as/is platform"? Just about any library I need, 
I can find for Go. And they happen to work across multiple platforms, just 
like with Java. As with Java before it, Go delivers on database APIs, web 
APIs, and protocols. There's a robust ecosystem of third-party libraries to 
make coding in Go much better. Numerous IDEs and editors include excellent 
support for Go.

Tons of community, including Meetups around the world.

One piece that has been missing relates to versions on libraries. That has 
been addressed in various ways, and a proposal that will most likely join 
the toolchain is in the queue for Go 1.11 (vgo).
 

> I have watched golang for many years, but never turn to it. Why? I think 
> it is still semi-finished product. Creators of golang are researchers, not 
> engineers, they worked too slow.
>

What would meet the definition of finished? Java still isn't finished, all 
these years later, because they're still adding language features (not just 
libraries!). Is Java slow because of this?

Go has felt to me like it moves very rapidly, as incredible enhancements 
come very quickly to the ecosystem. The compiler is itself now written in 
Go, and most of the runtime, which means that it is more possible for Go 
developers to help tweak the future direction of the language than it is 
for say Java developers, for whom the JVM will mostly be a black-box 
written in C++ that they probably never understand.

Eric.

-- 
You received this message because you are subscribed to the Google Groups 
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to golang-nuts+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to