DjangoGears: Django and TurboGears to Merge

Please note that this entry was created for April Fool's Day (April 1). It is not a factual story. Thanks for visiting :)

The core developers for both Django and TurboGears have been holding discussions on the IRC channel #railsenvy for the past several weeks about plans to merge Django and TurboGears into a single code base. Everyone is in agreement that having too much choice in the community for web frameworks is dangerous for users and detracts from the Python community. The main cause for concern is the continuing growth of Rails and the worry that Erlang and/or Smalltalk frameworks will hypnotize developers into dropping Python altogether. Not to mention that Perl 6 is due to ship anytime in the next five years. Coincidentally, Larry Wall also joined into the discussions during the final phases and practically begged the group to get the new uber-framework standardized on the Parrot virtual machine, which they agreed to.

They are still working out the final details but here is what is known so far:

• Much of the core functionality will be transfered to the client-side in order to alleviate performance problems. They are still debating over whether any of the current AJAX libraries will be up to the task or if they should just roll their own.
  
• The ORM will be redesigned from the ground up. Neither Django's ORM nor SQLAlchemy represent the quality of what developers will want going forward. The queryset-refactor branch has always had a hush-hush underlying motive of being a practice round for this work. One thing that was learned is that ORMs are just too difficult to write and even more difficult to use. This is largely due to inherent flaws in the expressiveness of SQL and the group has already worked out a new and improved way for dealing with databases that should send shockwaves through the industry. The new language is sytantically based on COBOL and in fact they have found that much of the COBOL implementation provides the majority of what is needed.
• The current template system in Django falls far short of being usable. The Django developers have long admitted that not having access to the full Python interpreter from the scope of the template is a serious deficiency. PHP developers assigned to Django projects have been known to suddenly become afflicted with all sorts of medical problems when they discover this weakness. TurboGears has already solved this problem by utilizing up to 6 or 7 independent templating engines simultaneously. However, none of these other engines will match the innovation of the custom engine they are creating just for this project. The syntax resembles Lisp and will consequently be heavily functional based. One of the coolest features in my opinion is that you can also create blocks of LOGO code (the turtle style) for drawing boxes, lines, and other shapes on your page. This should be a tremendous benefit since doing styling with CSS is difficult and non-intuitive.

This is all seriously exciting news. They should be making a formal announcement soon but they are waiting until the announcement page can be built on the working version of the new framework.

<<     Back to Unbracketed Home Post a Comment Now