[Grok-dev] Grok and chameleon

[Sebastian Ware]

8 feb 2010 kl. 16.42 skrev Martin Aspeli:

> Souheil CHELFOUH wrote:
>
>> That's probably a bit early to start and argue about back  
>> compatibility.
>> Let's try and do the best option for Grok and see how we can get some
>> backward compatibility.
>
> I don't mean to harp on about this, but statements like that are
> incredibly dangerous. If you want people to use your software, you  
> can't
> leave backwards compatibility as an after-thought. Unfortunately,
> perfect is quite often the enemy of good when it comes to releasing
> software that is actually used in the wild, as is grokcore.view.
>
> Martin
>

To me it is all about cost/benefit. If the pain of upgrading is  
greater than the benefit of doing so, then it is a no go.

However, if there are serious benefits, and the upgrade is pretty  
straightforward, then it is no problem. Evolution if you like.

In other words, if it was part of a release of bugfixes I would be  
really pissed. But if it was a transition to the next level I would be  
happy to spend half a day making trivial changes such as dumping path  
expressions for python notation.

In this case it would seem that you could easily write a compatibility  
layer by means of extending the TAL-plugin. Then there is even less of  
a problem.

I understand the gut reaction to be conservative, but I have full  
confidence in Martijn and that he takes the introduction of breaking  
changes very seriously. What I would like to see though is slightly  
better deprecation warnings than when we changed the default  
permissions on views...

Mvh Sebastian