thanks for speaking up.

for me, adding the discriminator column is fundamentally more sound
and supports a wider set of use cases. It's too bad such a simple idea
met with such resistance when it can improve NHibernate. Other
products don't take the perspective, since integrating with legacy
systems is often a requirement.

It's nice to know I'm not the only one who thinks the idea has merit.

peter

On Sep 27, 3:13 am, Stefan Nobis <[EMAIL PROTECTED]> wrote:
> "Jon Palmer" <[EMAIL PROTECTED]> writes:
> > As I described in an earlier email, it's entirely right, indeed
> > preferable, that NHibernate does not add the where clause when you
> > query for the base class (as is always the case if there is only one
> > class).
>
> It depends on your viewpoint. You are right from a rather narrow,
> NHibernate centric view, where your hbm.xml files are all you have to
> consider.
>
> But from a broader and more abstract view, it's not really preferable,
> because it breaks generalization. This generalization is important, as
> was said before, if you have a larger toolchain than just NHibernate
> and e.g. you want to do model driven architecture with tools that
> generate hbm.xml and other code for you. These tools don't appreciate
> special cases (but they could handle it), but what's more important:
> Your big abstract model is no longer technology neutral, NHibernate
> special cases leak into the design.
>
> And the big abstract design may be: we want be able to support
> subclasses, but not in this iteration and maybe not be this team. We
> want a design that handles legacy data and at the same time is
> prepared to support subclasses later designed/developed/added by
> another team without having to change our master design.
>
> This abstract goal is not supported with NHibernate (yes, you can make
> workarounds for legacy data when there are no subclasses and yes you
> can support superclass and subclasses in different assemblies -- but
> supporting both with the subclasses only added at a later point in
> time without changing the mappings for the superclass is not
> possible).
>
> So, from an abstract point of view where NHibernate is not the one and
> only technology/tool to consider I think, indeed, it's not prefereable
> that NHibernat does not add the where clause.
>
> It's quite another question if the NHibernate team ever want's to
> support this broader, non NHibernate centric point of view, because it
> may be a cause of many headaches... but in this single case a simple
> runtime configuration option may solve the problem and thus serve both
> camps.
>
> --
> Until the next mail...,
> Stefan.
>
>  application_pgp-signature_part
> < 1KViewDownload
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"nhusers" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/nhusers?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to