Re: Microsoft's bridge between OO and relational

erk wrote:
> Marshall wrote:
> > JXStern wrote:
> > > What do y'all think of this stuff?
> > >
> > > J.
> > >
> > > http://www.ftponline.com/vsm/2006_04/magazine/features/rjennings/
> > >
> > > Test Drive VB9 and DLinq
> > > The January 2006 Language Integrated Query (LINQ) preview for the next
> > > ("Orcas") version of Visual Basic enables automating SQL Server
> > > object-relational mapping for DLinq and enhances XLinq syntax for
> > > literal XML and late binding.
> > > Roger Jennings
> > >
> > > March 27, 2006
> > >
> > > Technology Toolbox: VB.NET, SQL Server 2005, XML, Visual Studio 2005,
> > > or Visual Basic or Visual Web Developer Express editions, Visual Basic
> > > 9.0 LINQ Technology Preview (January 2006), SQL Server 2005 Express
> > > Edition or higher, Northwind sample database
> > >
> > > The forthcoming Visual Studio "Orcas" release promises major upgrades
> > > to data-management programming with Visual Basic 9.0 and C# 3.0.
> > >
> > > Language Integrated Query (LINQ) and its data (DLinq) and XML (XLinq)
> > > libraries transform relational data and XML documents into
> > > first-class, interoperable CLR objects
> > >
> > > ...
> >
> > Fascinating, but ultimately not compelling. They get all the right
> > functionality, but they don't manage the complexity at all.
> > It's not elegant; it's backwards-compatible. Still, one of the
> > few real attempts to do something ambitious. A noble failure.
> >
> > It is worth reading some of the papers that started out their
> > whole effort.
>
> Erik Meijer and some of the others working on this are decent
> researchers, with pretty solid pedigrees. They seem to be adding some
> useful type inferencing to Visual Basic, but I agree with Marshall:
> fascinating but not compelling. They omit constraints completely, for
> one.
Yep and worse they did not address in a satisfactory manner domain definitions. That I can tell you for sure. You can not have a good RM implementation without a proper domain support.
> It's a nice evolutionary step for Visual Basic, but like AOP
> (aspect-oriented programming), it's a re-invention of something done
> better long ago, albeit in a slightly more "hip" context.
Are you suggesting it's a regression. I do personally believe so.

> If Java and other alternatives weren't so completely stagnant and
> rotting in the clutches of their masters, this wouldn't look so good. I
> think I'm impressed mainly because it's a stark contrast to the last
> decade of Microsoft "innovation." But given the researchers they've
> bought over the last few years, we should expect much better.
>
> - erk
Can we really speak of innovation? I think things are getting worse everyday and I tell you that from experience. The whole idea about implementing XML structure in dat management is a mess and brings us back 30 years in past (hierarchical model), All what RM was supposed to cure. Received on Mon Jun 05 2006 - 20:20:41 CEST