Win2Linux-Chooser: An app to help migrate from Windows to Linux

Project Abstract

The project aims to help users migrate from Windows to GNU/Linux by analyzing their system specifications, asking relevant questions, and recommending a ranked list of distributions based on their preferences.

Deliverables

A MVP (Minimum Viable Product) that recommends a ranked list of distributions to the users switching from windows to linux

Mentors

Joseph P. De Veaugh-Geiss ([@joseph:kde.org])
Snehit Sah ([@flyingcakes:kde.org])
Aakarsh MJ ([@aakarshmj:matrix.org])

Merge Requests

BLog Post

https://drowsywings.github.io/sok25.html

Weekly Progress

Week 1

Discussed criteria for recommending distributions.
Decided to ask 4-5 questions without excessive technical details.
Explored distributions suitable for beginners, including Kubuntu, Fedora KDE, Linux Mint, and Lubuntu.
Started working on the recommendation engine logic.

Week 2

Decided on parameters to rate distributions:

 # Ease of installation
 # Old hardware support
 # Good documentation
 # Community support
 # Updates

Finalized JSON format for questions and CSV format for distribution weights.
Chose local processing over server-side for privacy and efficiency.
Decided to use PySide6 + Kirigami for the project.

Week 3

Set up the basic repository on KDE Invent.
Implemented the basic UI.
Worked on hardware detection using WMI and psutil.
Discussed filtering methods for the recommendation engine:

 # Agreed on using a cosine similarity model.
 # Treating each distribution as a vector.
 # Calculating similarity with the user preference vector.

Week 4

Basic recommendation engine (more detailed breakdown in the blog post) operational using dot product; switching to cosine similarity.
Majority of the desktop client UI completed.
Improved the questionnaire and refined the algorithm by adding a penalty criterion.
Researched driver and external hardware support.

Week 5-6

Narrowed the scope to Win102KDE-Chooser to KDE-based distributions.
Discussed the driver support, and decided to start with Nvidia driver support and extend it later.
Improved questionnaire so that questions directly map to parameters which are predefined in distribution vector
Merged the script which extracts hardware info (uses wmi) from the devices

Week 7

Implemented the recommendation logic
Made adjustments to the parameters to refine the recommendation system.
Used a penalty mechanism for categorical questions, to push certain distributions lower in the list which doesn't match user preference
Added a mock-ui which shows ranking of distros with there respective scores.

Week 8-9

Implemented and integrated hardware detection in the recommendation logic.
Implemented a UI for Hardaware Info Page.
Begin testing the application on a Windows machine
Also added an option to skip hardware detection if needed.
Decided to take Blue Angel Specifications as reference

Blue Angel Specifications (Reference System)
* Model: 2015 Esprimo P920  
* CPU: Intel i5-4590 (4 cores, 3.3 GHz)  
* RAM: 4 GB DDR3 (1600 MHz)  
* Storage: 500 GB SATA III HDD (6 Gbit/s)  
* Graphics: Intel HD Graphics 4600