[Techtalk] [x-post] Go

[Wim De Smet]

Hi,

On Mon, Nov 23, 2009 at 9:29 AM, Rudy Zijlstra
<rudy at grumpydevil.homelinux.org> wrote:
> Op maandag 23-11-2009 om 08:52 uur [tijdzone +0100], schreef Wim De
> Smet:
>> Hey,
>>
>> Not posting this to list?
>
> accidentally... adding the list back
>
>>
>> On Sun, Nov 22, 2009 at 4:22 PM, Rudy Zijlstra
>> <rudy at grumpydevil.homelinux.org> wrote:
>> > Op zondag 22-11-2009 om 16:01 uur [tijdzone +0100], schreef Wim De Smet:
>> >> (I'm restricting this post to techtalk, feel free to repost)
>> >
>> >>
>> >> I'm not a big fan of the ':=' style for assignment, but it definitely
>> >> looks like a solid language. Unfortunately coming from a java world
>> >> I'm much more likely to write something in Scala than Go. I definitely
>> >> see this replacing C/C++ for many people though. Not to start a
>> >> flamewar but I think there's two languages whose time has come.
>> >>
>> >
>> > Hmmm... This makes me think of COBOL (Completely Obsolete Bad Old
>> > Language). It was that in 1983 when i was learning programming. It still
>> > is, and is still in widespread use...
>>
>> I know people that program in COBOL and in many ways it still is a
>> useful language. But I think its popularity has more to do with the
>> hardware/software environment it runs on than vice versa. IOW, COBOL
>> is still around because the mainframes it was used on are still
>> around.
>
> Actually from what i know the mainframes have been replaced many times.
> The SW environment is the problem.

The environment as a whole, then. The thing is, if you're replacing a
mainframe with a bunch of inexpensive but crappy servers, you're
suddenly in an entirely different environment with entirely different
failure modes and software requirements. Banks are conservative
institutions and bank programmers are as much set in their ways as
anybody. Mainframes have nice upgrade paths which are definitely more
expensive, but It Works(TM).

>
>>
>> And also because we suck at big software projects, so replacing
>> something that grew organically over so many years gets really hard to
>> replace since the replacement projects keep failing badly.
>>
>> Anyway, I don't see many people starting new projects in COBOL (I hope
>> they're not doing that, anyway) so arguably it's already a dead
>> language. I'd be more than happy if that were to happen to, say, C++
>> but I think Bjarne Stroustrup would disagree with me. :-)
>>
>
> I am not active in the financial sector, but my understanding is that
> COBOL is still actively used. You can still find development on the
> language itself, which points to active use and likely new projects as
> well.

My impression was that it's mostly code integrated or associated with
existing stuff. But I don't know, really. Some companies sell tools
for COBOL (and make a lot of money on them), I assume they see no
reason to stop developing features.

Apparently there's a COBOL version for .Net too. I thought that was kinda scary.

> I have my own gripes with C and C++ (coming from a Pascal / Ada
> background), but i expect them to be around for a very long time. They
> clearly have their uses and a lot of code exists.

They're not going anywhere anytime soon of course. I can only hope. ;-)

More seriously, I'm sure C++ at least will stay around for a very long
time, because there are still many situations where you need that
extra bit of control over how everything moves around in memory and
elsewhere. I personally don't think it's a very good language to build
user interfaces in though, but then that's just me.

regards,
Wim