b*****y
发帖数: 44 | 1 If you can build systems, you will be able to find a good job anywhere.
Merely publishing papers will get you to nowhere although you are in
the system area.
Networking is already satuated. Anway, hard to say in 3 to 4 years' time. | d*********w
发帖数: 5 | 2 Selecting a PHD area based on whether the area is hot or not is a bad idea
since CS in general is moving too fast for one to predict what WILL BE hot in
3-4 years. The only sensible way is to select a PHD area based on one's own
interests. I will list some points to support this statement.
1) You are going to work in this area for a long time (and in your best
years). You better like what you are doing (it won't last long otherwise).
2) If one does not like an area, it is hard for him to dedicat
【在 b*****y 的大作中提到】
: If you can build systems, you will be able to find a good job anywhere.
: Merely publishing papers will get you to nowhere although you are in
: the system area.
: Networking is already satuated. Anway, hard to say in 3 to 4 years' time.
| b*****y
发帖数: 44 | 3 Not application level programming, but kernel hacking, etc. | f*****r
发帖数: 229 | 4 From my understanind, kernel programming is only basic requirement. In fact, you
is like an architect, know what does a system needs, and how do you implement
with a cost-effective way. If you have some idea, make sure it is practicle
and really useful.
Theoretically, a system has at least 5 kinds of requirements: functionality,
performance, dependibility, useability, and cost. You don't need to
be an expert for every catagory, at least you need to know what is the common
practice for each of t | u*****n
发帖数: 160 | 5 quite good points and well-said!
you
implement
common
time.
【在 f*****r 的大作中提到】
: From my understanind, kernel programming is only basic requirement. In fact, you
: is like an architect, know what does a system needs, and how do you implement
: with a cost-effective way. If you have some idea, make sure it is practicle
: and really useful.
: Theoretically, a system has at least 5 kinds of requirements: functionality,
: performance, dependibility, useability, and cost. You don't need to
: be an expert for every catagory, at least you need to know what is the common
: practice for each of t
|
|
