Re: GSWeb/GDL2 Status

[David Ayers]

Nicola Pero schrieb:
>>Eg something I consider more "duplicate efforts" is libSQLClient  
>>because this is already covered by the GDL1 *and* GDL2 adaptor level
> 
> Yes ... but it's also nice to have different packages for different types of 
> users. ;-)
> 
> I suppose if you want to remove duplication, at this point why not re-base 
> GDL 
> on SQLClient :-)

FWIW, I think you right about it being good to have different packages
for different users.  In my view current GDL2 users have a requirement
aka constraint that that SQLClient users do not have.  That is WO45
compatibility.  This includes the nasty KVC and EOObserver-dealloc
method-swizzling and hacks in GDL2.

This gives SQLClient (and EOF-like layers using SQLClient) the freedom
to make it's own design decisions and develop that design in the future.
 Of course we could add new features as optional extensions but certain
design decisions will constrain you.  SQLClient doesn't have this
constraint.  It can go wherever experience leads it.

If the GDL2/GSWeb designs weren't so powerful for certain type of
applications, I would recommend building on top of SQLClient and a newer
independent XML/XHTML templating library someone still comes up with in
favor of them.

In this respect I must admit that both GDL2 and GSWeb could be
considered dead end streets.  OTOH, I also believe that the full
potential of their concepts can produce powerful applications we haven't
seen here yet and frameworks building on SQLClient should take a good
look at GDL2 and see what concepts and code it can recycle.

But rebasing GDL2 on SQLClient does not seem feasible in my view.

Cheers,
David