I would love to see an application, which looks at my models and provide different Admin UI --- current Admin provides all necessary basic functionality but it would be nice to extend it even more for some custom apps. As soon as I have it, I'll compare it with existing Admin application and form my opinion.
Most probably it would take a lot of time to develop it => I think it should be an independent project: there are other more pressing issues in Django to be fixed. But I am all for such parallel project. Regardless of final outcome it will generate a list of Django public interface deficiencies, which can be fixed. The problem is availability of resources for such thing. Thanks, Eugene "Ian Maurer" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED] On 9/19/05, Jacob Kaplan-Moss <[EMAIL PROTECTED]> wrote: > > > On Sep 19, 2005, at 11:39 AM, Robert Wittams wrote: > > Well, I'm not an author of the framework, but the admin code is very > > hard to understand (imo). I'm sure the authors understand it fine, but > > for open source to instill confidence clarity is key. The admin view > > should be a great example to django users, but unfortunately right now > > it isn't. > > No, it shouldn't -- the admin views are incredibly complex because > the admin interface is incredibly complex. Public views are almost > always going to be much much simpler. Even if the admin views get > cleaned up -- which they should be, of course -- they still won't be > a very good example. You wouldn't tell someone to read the Linux > kernel source to learn how to program in C; the admin views should be > thought of the same way. I think I know where people are coming from. When I first started playing with Django, I was frustrated that there were no examples on how to leverage the admin into my webapp. But, once I looked at the code and read some of the tutorials it dawned on me that the admin stuff was NOT to leveraged by my webapps. Once I accepted this, I became very content with letting the admin code stay behind the curtain and concentrated on building my site using the standard set of tools. I think most new developers to Django will go through this thought process and either accept things the way they are or get frustrated and move on to another framework. So, I offer 3 suggestions: 1. Make it more clear that the admin tool is really a separate applications and cannot really be leveraged into new apps easily. This should be done in that first admin tutorial. 2. I really think the admin tool is very slick and it really is a differentiator for Django. I also understand some serious wizardry and magic went into making it, so once 1.0 is finished and in the can, I would suggest looking at the Admin tool as the guide for future enhancements or versions of the software. I would love to see the key admin components (such as the auto data tables, filters, search, and forms) abstracted in such a way that any app could use them easily. To me, this would be another layer of Rapid App Development on top of generic views that would make this framework even more productive (especially for quick-and-dirty CRUD sites). 3. This last item is really out there... But, to me it would be nice if the admin tool could be decoupled from the ORM library (using interfaces, adapters and pyprotocols?), that way people could create adapters for SQLObject or any other ORM tool and they wouldn't lose out on using the Admin console. While I am very content with the ORM library, a lot people might not use Django if it cannot play nice with their existing models. Sorry for contributing just ideas at this time... I hope I can start contributing patches as I get comfortable with the actual code behind the framework. thanks again for the great framework, ian
