Looking for PhD Research topic

[gaurav.parashar]

Hii guys!! I am gaurav from India. I love this device and would really like to contribute to raspberry pi and for that i am looking for a PhD research topic, so that i can make a product around it, so if anyone has any idea from where i can start?

[tufty]

PhD topic based around the raspberry pi? I'd probably go for something more generic, myself, but that's me. Here's some ideas, though, mainly based on what I'm doing myself.

"A low-footprint OS for teaching computer science"
"RAM as writable cache for persistent memory"
"Real time non-moving garbage collection strategies for a flash-backed persistent store"
"Ubiquitous persistence in the presence of extremely slow storage"
"Transactional storage subsystem for a system with no safe shutdown mode"
"Just in time compilation onto heterogenous cores"

If you want more, let me know.

[jackokring]

* The required contents of hardware grab bags such as to force co-operation between bag owners to construct projects.
* Common code subroutine nesting, to extend the idea of common libraries to a minimal memory occupation for a magnificent code base.
* Elegant real-time editor systems for dynamic programming with continuous execution.
* Implementation of a BWTS sector or block loader for low touched files, improving read bus bandwidth on SD media.
* Error information strategies for turning stack traces and error messages into informative content.
* Power budgets, and the creation of the realizable personal computer.

[johnbeetem]

My general advice: if you don't already know what you want to do for PhD research, get out and work in the real world for a while. You will discover all sorts of things that people are doing badly. Ideas for doing them better form a good basis for PhD topics.

As Kettering said (IIRC): "Research is deciding what to do when you can't go on doing what you're doing now". So look for things that aren't scaling well using current approaches, and invent new approaches. Above all, don't assume that current solutions to problems are the right solutions for the future. Current solutions are based on past constraints, and those constraints may no longer be present. My favorite example of this is how virtually every programming language you see today is compatible with teletypes and punched cards. Seen one of those lately?

[gaurav.parashar]

Thank you guys for the help extended. I am grateful. I will ponder upon your thoughts and would revert after brainstorming for my comfort on the topics. I agree to johnbeetem that i must do some work and then will be able to come to some conclusion.
Again thanks guys for the help extended.
Regards
GP