GSoC QMLify plasmoids proposal

[Sebastian Kügler]

Hi Viranch,

On Monday, March 28, 2011 13:07:39 Viranch Mehta wrote:
> Please review my proposal for GSoC project QMLify Plasmoids:
> https://docs.google.com/document/d/19zZn2ObkGWFRVOjaaVu1qp54Olb7g50FN6RgyAQ
> hFm0/edit?hl=en&authkey=CKPlh6oP
> 
> I've specified the order of plasmoids that I think would be appropriate in
> increasing order of complexity.

You're referring a lot to QML/C++ Plasmoids, which is a central concept of 
your application, but unclear what it means. Most of the plasmoids you name 
right now are actually already written in C++, so not everybody reading your 
proposal might understand what it's about (though the framing in the 
introduction is quite good).

I thin the focus should be on:

- Port Plasmoids visual appearance and interaction to QML and JavaScript

- Where necessary, enhance existing dataengines, or create a new one. (this is 
  often the hairy point, there is logic in some Plasmoids which you want to, 
  or have to do in C++, in that case, try moving it into a dataengine

- For some Plasmoids, it would be a lot of work to do dataengines for all 
  their data, many Plasmoids would benefit more from a customized data model 
  they can use to display the data. Here, we'd need to find a more structural   
  solution. (Could be a "generic model->dataengine" engine, which flattens a 
  model and offers the data to the plasmoids as normal dataengine entries.)

- In some cases, you'll want a hybrid approach: Logic in the Plasma::Applet 
  C++ subclass, which then loads one (or more) QML files, depending on 
  formfactor. I think this is the least desirable option overall, since it 
  bears the least value for re-usability.

I hope my feedback is useful for your application.

Cheers,
-- 
sebas

http://www.kde.org | http://vizZzion.org | GPG Key ID: 9119 0EF9