The release of Django 1.0 comes with a promise of API stability and forwards-compatibility. In a nutshell, this means that code you develop against Django 1.0 will continue to work against 1.1 unchanged, and you should need to make only minor changes for any 1.X release.
In this context, stable means:
All the public APIs – everything documented in the linked documents below, and all methods that don’t begin with an underscore – will not be moved or renamed without providing backwards-compatible aliases.
If new features are added to these APIs – which is quite possible – they will not break or change the meaning of existing methods. In other words, “stable” does not (necessarily) mean “complete.”
If, for some reason, an API declared stable must be removed or replaced, it will be declared deprecated but will remain in the API for at least two minor version releases. Warnings will be issued when the deprecated method is called.
See Official releases for more details on how Django’s version numbering scheme works, and how features will be deprecated.
We’ll only break backwards compatibility of these APIs if a bug or security hole makes it completely unavoidable.
In general, everything covered in the documentation – with the exception of anything in the internals area is considered stable as of 1.0. This includes these APIs:
Most of the modules in django.utils are designed for internal use. Only the following parts of django.utils can be considered stable:
There are a few exceptions to this stability and backwards-compatibility promise.
If we become aware of a security problem – hopefully by someone following our security reporting policy – we’ll do everything necessary to fix it. This might mean breaking backwards compatibility; security trumps the compatibility guarantee.
While we’ll make every effort to keep these APIs stable – and have no plans to break any contrib apps – this is an area that will have more flux between releases. As the Web evolves, Django must evolve with it.
However, any changes to contrib apps will come with an important guarantee: we’ll make sure it’s always possible to use an older version of a contrib app if we need to make changes. Thus, if Django 1.5 ships with a backwards-incompatible django.contrib.flatpages, we’ll make sure you can still use the Django 1.4 version alongside Django 1.5. This will continue to allow for easy upgrades.
Historically, apps in django.contrib have been more stable than the core, so in practice we probably won’t have to ever make this exception. However, it’s worth noting if you’re building apps that depend on django.contrib.
Certain APIs are explicitly marked as “internal” in a couple of ways:
django.contrib.localflavor contains assorted pieces of code that are useful for particular countries or cultures. This data is local in nature, and is subject to change on timelines that will almost never correlate with Django’s own release schedules. For example, a common change is to split a province into two new provinces, or to rename an existing province.
These changes present two competing compatibility issues. Moving forward, displaying the names of deprecated, renamed and dissolved provinces in a selection widget is bad from a user interface perspective. However, maintaining full backwards compatibility requires that we support historical values that may be stored in a database – including values that may no longer be valid.
Therefore, Django has the following policy with respect to changes in local flavor:
For example, Django 1.2 contains an Indonesian localflavor. It has a province list that includes “Nanggroe Aceh Darussalam (NAD)” as a province. The Indonesian government has changed the official name of the province to “Aceh (ACE)”. As a result, Django 1.3 does not contain “Nanggroe Aceh Darussalam (NAD)” in the province list, but does contain “Aceh (ACE)”.
May 10, 2012