[TYPO3-commerce] commerce development - next steps

[Christian Kuhn]

Hey.

On 06/13/2011 07:41 PM, Rik Willems wrote:
> There are a lot of issues in Forge that are either very old, not well
> described or just ignored. In my thought it would be good to clean up
> Forge to get a good view on what are true problems with commerce
> currently. These problems can be tackled so we end up with a short issue
> list what makes it much easier to maintain the extension.

Some of the current issues will become obsolete / fixed together with 
the switch to fluid.


> To be honest I'm unsure wether I'm in favour of the switch to Fluid.
> Probably it will be a lot of work. This work pays with the introduction
> of a flexible templating system. On the other hand it doesn't fix any
> real problems in commerce. I'd much prefer to freeze the features of
> commerce and put all the effort into fixing the major current problems.
> Implementing Fluid in Commerce at this point to me feels like putting
> new rims on an already great car that still has some engine problems to
> sort out.

pi2 and pi6 are already done (but not committed yet). The patches make 
the pi classes much smaller and cleaner and help a lot in cleaning up 
the code base, eg. pi2 is now ~1/3 of previous lines of code, together 
with better comments and a straighter approach. Furthermore the 
templates are much easier to read (for example there is no more even/odd 
subpart snafu and so on).


> I saw various comments on breaking changes. I'm very much in favour not
> to implement these. Running a commerce implementation requires much
> attention to details already. If you are not able to update a commerce
> implemtation because of breaking changes I think many users will be very
> much annoyed. Breaking changes is not in line with TYPO3 development in
> general, too.

Yeah. We thought about backwards compatibility for quite some time. But 
in fact it is impossible to change a template engine without breaking 
something. I still think it is a good idea to release a last version 
that already cleans up some things but aims not to break existing 
installations. We can still fix minor things in this version then, and 
we can go ahead with trunk and more radical changes after that.


> Perhaps it would be good to dive into all the Commerce backend scripts
> as well. But a recent look into that found that is even more complex
> stuff than the frondend. Perhaps that is for later but good to keep in
> mind.

Yep, the backend modules needs some love as much as everything else. For 
example the orders module could be easily rewritten based on an extJS 
grid as soon as the object API (in lib/) is in a usable state. For usual 
developers the pi classes and templates are more important, that is the 
primary area where customization is done. Imho the project gains much 
more if we improve the frontend area first.


Regards
Christian