On 10/15/17 13:40, Laeeth Isharc wrote:
On Saturday, 14 October 2017 at 22:43:33 UTC, Adam Wilson wrote:
On 10/7/17 14:08, Laeeth Isharc wrote:
In a polyglot environment, D's code generation and introspection
abilities might be quite valuable if it allows you to write core
building blocks once and call them from other languages without too much
boilerplate and bloat. One could use SWIG, but...
Oh dear, I seem to have accidentally set off a firestorm.
Personally, I think there are better places to focus our energy than
worrying about what the users of other languages think. We like D,
that should be enough for us. The last line was somewhat
tongue-in-cheek. There is no way we're going to convert C#/Java users
either, and unlike C/C++ we cannot easily inter-operate with them.
If we can convert Pascal users, why won't some C# and Java programmers
be receptive to D? Plenty of people have approached D from Java:
Indeed. I am an example of just such a convert (from C#). :) But it is
much more difficult to inter-operate. The easiest path that I can see is
Micro-services. Hide your different languages behind a REST API, such
that your components are not talking C#<->D anymore, but HTTP<->HTTP.
Then the language matters a LOT less, then you can convert individual
services to a new language whenever business needs dictate.
https://dlang.org/blog/2017/09/06/the-evolution-of-the-accessors-library/
https://dlang.org/blog/2017/08/11/on-tilix-and-d-an-interview-with-gerald-nunn/
https://github.com/intellij-dlanguage/intellij-dlanguage/
(Kingsley came from Java).
Why can't we easily interop with Java and C#? I didn't find interop
with Java so bad for what I was doing (embedding a Java library via the
JVM with callbacks to D code), and Andy Smith found similar.
http://dconf.org/2015/talks/smith.pdf
(towards the end)
C# interop for what I am doing sees easy enough. (I will generate C#
wrappers for D structs and functions/methods).
This work wasn't open-sourced, and nor did Microsoft send out a press
release about their use of D in the COM team. But I spoke to the author
in Berlin (maybe you did too), and it wasn't so much work to make it
useful:
http://www.lunesu.com/uploads/ModernCOMProgramminginD.pdf
Very cool. I had no idea MSFT was doing that. I didn't talk to him, but
I was fighting a bad cold at DConf this year. :( I avoided talking to a
lot of people.
Even so, I was more commenting on the fact that D has built in support
for inter-operating with C/C++. It's not that it's impossible to
inter-operate with C#/Java/etc., but it is significantly more work. And
that can be a significant barrier to conversion when the ecosystem they
are coming from comfortably provides everything they need already. :)
Instead of worrying about how to get more people to come from a
specific language. People will come if they see an advantage in D so
lets try to provide as many advantages as possible. :)
Yes - agree with this.
--
Adam Wilson
IRC: LightBender
import quiet.dlang.dev;
