[Info-vax] Fortran

[Arne Vajhøj]

On 12/3/2018 3:41 PM, Bill Gunshannon wrote:
> On 12/3/18 3:24 PM, Arne Vajhøj wrote:
>> On 12/3/2018 9:13 AM, John Reagan wrote:
>>> On Monday, December 3, 2018 at 12:44:45 AM UTC-5, Phillip Helbig
>>> (undress to reply) wrote:
>>>> I remember a statement from VSI that one of their goals was to
>>>> have up-to-date compilers.
>>>
>>> To be specific, the roadmap has mentions improved standards
>>> compliance for C, C++, and Fortran.  I just want to shut down any
>>> discussion of object-oriented COBOL from newer COBOL standards.
>>
>> Interest for that seems to be low. If somebody want OO they
>> pick another language.
> 
> You got that backwards, kinda.  OO was added to COBOL even though the
> actual user base said they neither needed or wanted it.  The standards
> body, in its infinite wisdom, added it anyway.  And then they all got
> POed when the COBOL user community refused to waste time and effort
> using it.   :-)

I don't think that is much different from what I said.

>>> I've been in touch with the flang/f18 folks.  I've been code
>>> reviewing some of their recent changes to get DWARF generation for
>>> Fortran specific features like COMMON blocks, etc.
>>>
>>> What I imagine is that once we are fully native with a newer
>>> clang/LLVM, we'll look at providing flang/f18 on OpenVMS x86 along
>>> side of the traditional Fortran compiler so you can choose.  Probably
>>> not at the very first release, but I'd hope very soon afterwards.
>>
>> The benefits of using an industry standard CPU and an industry standard
>> compiler backend.
> 
> Standards are only of value when they reflect the needs of the user
> community.  When their main purpose is to fluff up the egos of the
> standards body members, what's the point?

Both x86-64 and LLVM are industry standards or de facto standards.

Your comment seems to target formal standards (ISO, ECMA, ANSI, JCP etc.).

Arne