2008/5/14 jeyram jadenthradevan <j.jeyram@xxxxxxxxx>: > hi every one, > > i am a Bsc computer science students. we have to do a research project next > year (2009). > and research time duration is one year. it's a group project. Hi! There are many interesting subsystems on the kernel for research. Virtual memory, process scheduling, block device layer are examples of some (allways) hot areas. At first, things depends on some factors: your fluency on the subject (the OS subject), your skills writing (and reading) C code, and your motivation, your inclination toward some topic. At this point, it looks better to acquire a general overview of the Linux kernel (reading Robert Love's Linux Kernel Development is a good start - http://rlove.org, or the famous -- and free as in free beer! -- Linux Device Drivers, aka LDD3, by Corbet Rubini and Kroah-Hartman http://lwn.net/Kernel/LDD3/). After a global view of the kernel by you and your colleagues, thigns will become clear and a topic (or some) will emerge, narrowing your options from hundreds to a handful. Since that you said on a post that you like things related to device drivers (and are confident about some subtopics), why not to implement a sample driver (the same one) using more than one strategy for acquiring device data, and scientifically justifying (with theories, measures, graphics, extrapolations et cetera) which one is better? You can research over the zillion existent drivers, changing stuff and measuring thousand of useful metrics. Go ahead. Anyway, keep in mind that research projects are not meant to be started as Nobel prize nominees, and sometimes simplicity at the surface can mean hardcore tracks deep down, and a lot of fun :) Asking why that data structure is used there, or why that algorithm is used somewhere are good questions for planing an attack to some research topic. Cheers Mauad -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
