August 17, 2016 Re: ISO D | ||||
---|---|---|---|---|
| ||||
Posted in reply to Russel Winder | On Wednesday, 17 August 2016 at 10:43:01 UTC, Russel Winder wrote: > On Wed, 2016-08-17 at 08:02 +0000, eugene via Digitalmars-d wrote: >> Hello, everyone, >> will ISO D be in future or not? > > I am not sure this would be a good thing. Given the history of Fortran, C, and C++, avoiding ISO standardization processes would seem to be a good thing for any other programming language. Even the Java process is fairly dreadful. > > Having a sensible version release programme is more important that a standard, though having a formal language specification in machine readable format is a good idea. >avoiding ISO standardization processes would seem to be a good thing for any other programming language I disagree. ISO, or not ISO, D should be standardized/stabilized. I dropped D once about a year ago because the new DMD version has broken backward compatibility. Some libraries have stopped compiling, and it caused a huge mess in dub package manager. I was trying to build Dash engine with dub, spent some days with this package hell, and ended up with removing dub from my PC. As for C++98, it didn't change since 1998, and I'm pretty sure I can compile anything written in C++98 even today. I can't say I love C++, but I like its versioning. C++98, C++11, C++14 seems to be different languages, and nobody's trying to compile C++11 code with C++98 compiler. Regards, Alexey |
August 17, 2016 Re: ISO D | ||||
---|---|---|---|---|
| ||||
Posted in reply to TencoDK | On Wednesday, 17 August 2016 at 12:57:59 UTC, TencoDK wrote:
>
> I dropped D once about a year ago because the new DMD version has broken backward compatibility. Some libraries have stopped
You could have used STABLE DMD (v1), right? Assuming that you *intentionally* used the unstable (v2) DMD, I would say it is unfair to complain...
|
August 17, 2016 Re: ISO D | ||||
---|---|---|---|---|
| ||||
Posted in reply to Dejan Lekic | On Wednesday, 17 August 2016 at 14:19:08 UTC, Dejan Lekic wrote:
> On Wednesday, 17 August 2016 at 12:57:59 UTC, TencoDK wrote:
>>
>> I dropped D once about a year ago because the new DMD version has broken backward compatibility. Some libraries have stopped
>
> You could have used STABLE DMD (v1), right? Assuming that you *intentionally* used the unstable (v2) DMD, I would say it is unfair to complain...
That make sense. Though, I think, this must be a common misunderstanding among newcomers.
D1 seems more like "an old, deprecated alpha version with support dropped a long time ago".
|
August 17, 2016 Re: ISO D | ||||
---|---|---|---|---|
| ||||
Posted in reply to TencoDK | On Wed, 17 Aug 2016 14:28:24 +0000, TencoDK wrote:
> On Wednesday, 17 August 2016 at 14:19:08 UTC, Dejan Lekic wrote:
>> On Wednesday, 17 August 2016 at 12:57:59 UTC, TencoDK wrote:
>>>
>>> I dropped D once about a year ago because the new DMD version has broken backward compatibility. Some libraries have stopped
>>
>> You could have used STABLE DMD (v1), right? Assuming that you
>> *intentionally* used the unstable (v2) DMD, I would say it is unfair to
>> complain...
>
> That make sense. Though, I think, this must be a common misunderstanding
> among newcomers.
> D1 seems more like "an old, deprecated alpha version with support
> dropped a long time ago".
Far from alpha. But it's deprecated, and support for it ended in 2012. code.dlang.org shows four projects that support D1. You can't find any information on D1 at dlang.org -- no downloads, changelog, spec, anything. There is no official, simple way to use a library from an old D version. Dub defaults to building everything with the system D compiler. New features and bugfixes provide an incentive to use the latest compiler version.
The effect of it is that, if you have a library, it needs to support the most recent D version. If you have the time, you can support parallel branches for several versions of D, but that's laborious, so most people mix library feature development with language change updates.
This is true even of incremental updates -- if I want to support a company that standardized on DMD 2.044 as well as the majority who are on DMD 2.070 and higher, that's probably going to mean two branches of development. D1 and D2 even more so.
|
August 17, 2016 Re: ISO D | ||||
---|---|---|---|---|
| ||||
Posted in reply to eugene | On Wednesday, 17 August 2016 at 11:34:01 UTC, eugene wrote:
> On Wednesday, 17 August 2016 at 10:47:35 UTC, qznc wrote:
>> On Wednesday, 17 August 2016 at 08:02:42 UTC, eugene wrote:
>>> will ISO D be in future or not?
>>
>> What would be the benefits?
>
> unified language standard?
As a former Common Lisp user, I have to disagree rather strongly with that idea. IMO the worst thing for Common Lisp is the ANSI standard. You have several issues:
- One group argues that the CLHS is all you need. Then you're using a language largely defined in the late 1970's.
- There will never be an updated standard due to cost and impossibility of getting compromise on the original standard.
- To do anything useful - or at least modern - with the language, you're in the land of non-standard extensions and people telling you to use the standard. Any attempt to modernize the language by getting rid of awful names like 'princ' are quickly beaten down because princ is in the standard.
CL is largely a 40-year old language (not since standardization was complete, but since the language itself was defined) and there is little hope that it will be modernized. If D had a base of millions of developers and tens of thousands of commercial users, it would be different. Standardization would be at least as much of a disaster for D as it was for CL. Even C++, which mostly has users because of legacy code and a lot of money on the line, has found it necessary to continually update the language.
|
Copyright © 1999-2021 by the D Language Foundation