Mailinglist Archive: opensuse-project (538 mails)

< Previous Next >
[opensuse-project] [GSoC 2011] - GTK+ client for smolt
Hello!

I have already posted my ideas for this GSoC and a little bit about me (http://lists.opensuse.org/opensuse-project/2011-03/msg00459.html)

I looked at the Smolt code more closely and have submitted the proposal on GSoC site. I want to repeat my proposal to discuss it here.
I'll be very happy to your suggestions and comments!


Br, Max Usachev

-----------------------------------------------
Title

GTK+ client for smolt


Abstract

The goal of this GSoC project is to create program architecture which supports two or more independent UIs and redesign current codebase using this architecture. Implement GTK UI for Smolt and adapt current QT-based UI for new architecture. It's necessary to refactor and improve current codebase to get better perfomance and code style. I'm also going to implement some important and necessary wishes from Smolt wish list. The proposal is based on http://en.opensuse.org/openSUSE:GSOC_2011_Ideas#GTK.2B_client_for_smolt.


Detailed Description

Background

Smolt is a project for gathering hardware information from computers running Linux. Currently smolt-gui is written in PyQt and it's difficult to add this client to GNOME installation of openSUSE. So it's definitely important to have a GNOME integration to ship on the default install. It's also important to bring new features, which users proposed (Smolt wishlist). There is a separate request in openFATE for implementing GTK smolt client: https://features.opensuse.org/310490.


Use Cases

Use graphical version of Smolt client with native GNOME UI.


Benefits

The biggest advantage of this project is that openSUSE will get native Smolt client for GNOME desktop environment.Users, who use Gnome environment will get native Smolt client. Smolt client will be redesigned and will get good program architecture, clean code and new features (from Smolt wish list).


Caveats

The most difficult task from my point of view will be to design support of multiple UI toolkits. Smolt was designed and developed as QT application, so some work will be required to make it to support QT and GTK UIs. However I have skills in adapting Qt-based UI to multi UI architecture and I'm sure that I can do this task.


Technical Details

The project tasks can be devided into 4 parts:
1) Analizing current codebase and choose the best solution for new program architecture. The first ideas that come to my mind is use factory pattern to generate UI instances and maybe modified MVC pattern as program architecture. We can use factory pattern for all widgets or factory only for general UI. For example:

"""
from factory import ui_factory

ui = ui_factory(args)
ui.run()
"""

All logic about creating necessary UI placed in factory module and returns suitable UI depending on desktop environment. The good idea it to move each UI to separate module, as I did in 'Mnemosyne for Maemo': qt_ui, gtk_ui. My proposals about starting the program: smolt-gui -u gtk|qt, something like that.

2) Adapting current QT-based UI to new program architecture. It's very important task, because while implementing some new features we must not forget about old features. I think it will not take much time, because current codebase looks very clear and understandable. It's necessary to move all Qt related code to qt_ui package and some helpful UI functions in qt_ui/utils.py module.

3) Writing tests. It's necessary part of the task, because these test will help while implementing GTK UI and also cover most part of exisiting code. I'm going to refactor some modules and add new features, so, the test are vital to prevent code breaking.

4) Implementing GTK-based UI. It's very important and the biggest part of the task. We should decide, will new UI be looks the same as Qt UI or not. If we decide to change UI, we should also take time for UI design process. Anyway, I think we should take as a basis current Qt UI.

5) Refactoring the code and add new features accroding to wish list. There are many parts of the code, which we can improve. For example, the gate.py module. It't better to implement the _Gate class as singleton (there are some ideas about it: http://stackoverflow.com/questions/42558/python-and-the-singleton-pattern) and remove Gate function. It's necessary to use Pylint and pep8 checker to catch invisible errors and clean the code and make it beautiful and correct.


Why Me

I've been programming in Python for 4 years already. I have developed many applications - scripts, services, parsers, web applications, but most of them are GUI programs based on PyGTK. Also I have good skills in UI designing and vast knowledge in architecture of GUI programs. I've learnt a lot about porting Qt-based applications to PyGTK, when I was porting PyQt application to Maemo platform, during my previous GSoC project. I use Linux and know it very well. I am sure that I am the best candidate for this project, because my skills very suitable for this project and I have good theoretical and practical knowledges and strong ambitions.


Contact Information

Mail: maxusachev@xxxxxxxxx
Jabber: maxusachev@xxxxxxxxx
Skype: maxusachev
ICQ: 468893421
--
To unsubscribe, e-mail: opensuse-project+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-project+help@xxxxxxxxxxxx

< Previous Next >
List Navigation
This Thread
  • No further messages