I Do not wish to open this debate here. We did our evaluation phase and
prooved that there would be a price to pay for going toward .NET. (And i
worked with .NET and Python so i know a thing or two about Managed Code!)
Now the question is,... How can we connect our OO unmanaged code to the
relational data it uses. All that above SQL Server...
.... There has to be some O\R Mapping technologies out-there to do this job ?
We looked at EdgeXtend but it doesn't support SQL Server... We looked at
Versant OpenAccess but it's restricted to the .NET world !
Any suggestions ?
[quoted text, click to view] "Adam Machanic" wrote:
> "Acoquinar" <Acoquinar@discussions.microsoft.com> wrote in message
> news:5621CEB7-97CA-4F85-85E2-1D641343FC85@microsoft.com...
> >
> > Introducing Managed Code isn't always recommended for "Obvious" reasons.
> > (i.e. Performance)
>
>
> Many studies have been done on whether managed code is less performant
> than unmanaged code, and given the current technology of managed code, the
> assertion that managed code is slower boils down to nothing more than a
> myth. In the early days of Java, before the garbage collectors had been
> properly optimized, this may have been true. However, that is no longer the
> case and today's managed code is both cheaper to develop/maintain and will
> outperform unmanaged code in most(*) situations. You should do some
> reasearch on this topic instead of assuming that it's an obvious choice.
> There is a huge amount of research available that says the choice is
> anything but.
>
> (*) Note, unmanaged code is sometimes an appropriate choice -- which is why
> C# allows you to declare "unsafe" code blocks.
>
> --
> Adam Machanic
> SQL Server MVP
>
http://www.sqljunkies.com/weblog/amachanic > --
>
>
[quoted text, click to view] "Acoquinar" <Acoquinar@discussions.microsoft.com> wrote in message
news:5621CEB7-97CA-4F85-85E2-1D641343FC85@microsoft.com...
>
> Introducing Managed Code isn't always recommended for "Obvious" reasons.
> (i.e. Performance)
Many studies have been done on whether managed code is less performant
than unmanaged code, and given the current technology of managed code, the
assertion that managed code is slower boils down to nothing more than a
myth. In the early days of Java, before the garbage collectors had been
properly optimized, this may have been true. However, that is no longer the
case and today's managed code is both cheaper to develop/maintain and will
outperform unmanaged code in most(*) situations. You should do some
reasearch on this topic instead of assuming that it's an obvious choice.
There is a huge amount of research available that says the choice is
anything but.
(*) Note, unmanaged code is sometimes an appropriate choice -- which is why
C# allows you to declare "unsafe" code blocks.
--
Adam Machanic
SQL Server MVP
http://www.sqljunkies.com/weblog/amachanic --
[quoted text, click to view] > Our application is in VC++ and our business logic involves intensive
> calculations on huge amounts of data.
>
> For business reasons, we're forced to use MS SQL and not the obvious
> ODBMS.
>
> We were wondering if anyone would be in the same situation and if anyone
> chose to use an Object-Relational Mapper ?
using an O/R Mapper means introducing another layer and this often, but not
always, can reduce performance.
It's also depending on the O/RM implementation, of course.
[quoted text, click to view] > P.S. Going .NET is totally out of question for us,... for obvious reasons
> !
no, that's not obvious.
regards, Robert
