On Jul 30, 2010, at 7:22 PM, Iceberg wrote: > On Jul 31, 1:15 am, Jonathan Lundell <jlund...@pobox.com> wrote: >> On Jul 30, 2010, at 9:19 AM, VP wrote: >>> On Jul 30, 9:35 am, mdipierro <mdipie...@cs.depaul.edu> wrote: >>>> http://gluonframework.wordpress.com/2010/07/30/web2py-and-metaclasses/ >> >>> This is really nice. Please do more of this. >> >> My initial reaction is the opposite. The result might be more readable, but >> it doesn't strike me as more writable. >> >> What would be most helpful for me would be a deeper explanation (in the >> book) of what's going on behind the existing DAL "magic" syntax, rather than >> adding yet another layer of magic. > > You make a good point, Jonathan. And I think there is a underlying > question here. Which kind of audience is web2py targeting to? If for > developers, the existing DAL syntax is already powerful and magical > enough (the document is also good, here it is. > http://web2py.com/book/default/chapter/06 > ). Developers don't need another layer which is more fancy but not > more powerful.
I'm not satisfied with the treatment in the book. I'd like to see each of the DAL objects more completely described, especially as to the underlying Python types and the operations that they implicitly support. Several of them IIRC are polymorphic wrt their argument types, and you either have to divine this telepathically or read the source in detail. Likewise operator overloading. I'm sure it's second nature to Massimo, but for most of us, we have to hunt around for an example that matches our situation, and blindly copy & paste. Either that or experiment until it stops raising exceptions....
