Proposed by Vitalii Melnyk (profile, biography)
How will I do that project
I will have summer holidays at the time of GSoC, so I will be able to spend 25-35 hours a week for this project.
What methodologies will I use
I will use TTD(Test-driven Development).
I will try to split my project work to many smaller tasks. After finishing every task I will show result to my mentors and decide all pros and cons with them.
Suggested timeline and milestones
I should have 4 phases:
1) Learning Smalltalk and get hands-on with the topic from a UI Perspective
Here the idea would be to implement a UI for searching the image.
This UI should take into account what we have and be inspired by what is there outside of Smalltalk (e.g. SpotLight).
This allows me to learn Smalltalk by using it and to think deeply about what is needed from an image indexer from a client perspective.
Results
- User Interface
- Smalltalk knowledge
- Criteria for what we need for the backend
2) Backend
- What exists already?
- Smalltalk solutions
- External Libraries
- Services (like cloud based indexing)
- Write some tests, explore the solutions. Then hook the most promising to the UI
3) Developing full text indexer
4) Porting project to different Smalltalk dialects
After all works on one system i should work with project portability to other Smalltalk dialects.
Where I see the risks
First two weeks I will have exams in my university, so I will spend less time on this project. But it is not a problem, because I will spend more time after exams.
Also there may be a problem with porting project to Smalltalk dialects because I don't know the differences between them. I think my mentors will help me with it just giving me some good literature about it.
How the results will look like
An intelligent search feature with GUI, which will use search technology to rank all the packages, classes, methods and comments in the image. Ideally it would be portable to other Smalltalk dialects.
|