Time goes bye, but that doesn’t mean I’m not thinking about the new version of P4A almost every day…The point is, P4A 3.8.x (use the github master version, it’s ok and much more updated compared to 3.8.4) it’s working fine also if it’s some years old, it does everything needed and it does is well.
To go forward I want to be really forward, this is how I think about innovation.
So, in the past month I’ve been rewriting all P4A’s core using autoload, composer, skeleton applications, namespaces and all the modern fashion programming :) It’s all on github if you want to take a look. At the moment there are no plan about removing the “builder” pattern or the “I live in the session” feature of P4A, it would become a completely different framework and simply it wouldn’t be P4A anymore.
Now we come to the GUI, this is always been a really tough task, it’s been years since the first time we spoke about it and we’re still there… I’ve trying everything (bootstrap, dojo, qooxdoo, extjs) more than one time wasting countless hours. Last but not least I thought I decided to go with a nice bootstrap admin theme called adminlte, I’ve also contributed to their repository fixing a few things but… just when I was about to finish… simply I didn’t like it anymore, simply it was not the leap forward I’d want.
It the past weeks I’ve been working around 20 hours to integrate Polymer in P4A, results are good, it was quite easy (and that’s a good thing) due to it’s declarative nature and thanks to the wise shadowdom usage, but not perfect. Polymer it’s still very, very young. A lot of things are missing (dropdowns don’t work with value/description or do not support label or they can’t be submitted) a lot of components are lacking (datepicker, colorpicker, table/grids, consistent builtin file upload and surely many more).
I think it could be worth the wait, P4A could be the first framework to deliver Polymer applications, we would be again on top of the edge and this would be great for the live of P4A itself. I think I don’t want to look back at css/js frameworks, they seem so old compared to webcomponents. With Polymer (which by the way it’s under BSD license) it would also be much easier to create new widgets, there’s already a themes website which looks quite good and could be used by you to customize your applications.
If Polymer was complete (and fully working) P4A 4 would be already out, cause the porting was surprisingly easy, but it’s not so… we’ve to wait for it to be ready.
What do you think? Please let me know.