Issue

[Aeson]

I never before realized how similar Tonga's flag was to UAE's.

[DriXnaK]

Ok, that works perfect then. I've always been self-taught on math anyway so this is going to give me something nice to work towards now. So nice to have something to focus on again. I've been coming home from work feeling like I've accomplished nothing and was just wasting my life away, even though I've been learning a huge amount. Thanks for all the advice so far.

[Asher]

A lot of the research that people would think of (academically) is really pointless in CS.

Over the past few months we've been investigating a way to get realistic looking crayon inks on our multi-touch boards. We found a bunch of 50+ page multi-year research projects from PhDs and PhD candidates that went to absurd lengths -- they built a physics model to represent different kinds of paper, they had calculated how much of the wax gets transferred, etc. And then you skip to the end and look at their example output and it looks nothing like crayon.

Me and another guy (both just BSc) spent a couple weeks toying with prototypes before we implemented what we will actually be using: it's ridiculously simple compared to the "research" available, and it looks way better -- to the point that it's almost indistinguishable from real crayon. Added a couple optimizations and it draws about 100x faster than their overcomplicated ****, too. And we're applying for three patents out of this work. But boy, did they have complex models and algorithms with very high-level math to do their mathy model.

Anyway, point is -- in this field, if you're working in the real world, a PhD doesn't really do much for you most of the time. I work with a bunch PhD compsci people and at the end of the day, you have no idea if someone has a BSc, MSc, or PhD except if you ask them. People are either good at it or they're not. A PhD just indicates you spent more time in academia in this field.

If anything, my experience has been PhD types tend to overly complicate almost everything, which is the kiss of death in software.

[DriXnaK]

Well I don't want to be in academia, I still want to be in the private sector, but I want to work for a company that is creating products that have heavy R&D involved like pharmaceuticals or something. Basically I just want to be doing something that ends up creating something of meaning. Working for a freight 3PL is the furthest thing from that. I had planned to get a masters anyway at some point, and for me it would probably be a good idea anyway. Keep in mind my IT degree is from an online school, and even though I would say it beats the **** out of the USF IT degree, there is still a stigma attached to online schools. In fact we had a guy come in and work for about a month with a USF IT degree and we fired him within 6 weeks.

[loinburger]

I'd say that an MS is the sweet spot for CS R&D - you have enough background to know when to apply what where, without being so over-specialized that you've got your hammer and treat every problem as a nail.

[loinburger]

At my present job I was in charge of writing a box-packing algorithm (given a bunch of items with various dimensions, pack the items in a bunch of boxes with various dimensions using the minimum box volume). The old algorithm was obviously broken - it said "if the item's dimensions fit in the box's dimensions, and there is sufficient volume available for the item, then the item fits!", meaning that you could pack two 8x8x8 items in a 12x12x12 box, which is impossible. I looked into the research on box-packing, and every paper on the subject either used a brute-force approach (which would take friggin forever on an NP-Hard problem) or else used a ridiculously complicated heuristic. In the end I spent about three days trying to decipher a half dozen research papers that were intentionally obfuscating their algorithms because they wanted to sell them to industry, after which I spent a day ignoring 99% of the research on box-packing and writing my own algorithm that was Good Enough.

[Hauldren Collider]

This book is awesome if you're going to be looking into an MS - I used it (in its various editions) in 2002 (senior year), 2004 (last year of my master's program), and 2006 (near the end, ha ha, of my phd program).

QFT QFT. I got this as a graduation present from the TJ Sysadmins program. It's ****ing awesome

It covers the material we're doing in my second-level cs class right now (and quite a bit more), although the class doesn't have a book. I can't say I'm learning much in the course though, sorta covered a lot of this in high school. I'm trying to talk to the professor about making this book the book for the course.

[DriXnaK]

I guess my next step then is to start reading those books and then talk to an adviser at USF to see what all they're going to need me to do before I'd be able to be accepted into their program. I'm going to shoot for summer of 2012. I'll have more than a full year of experience at that point and hopefully either a job that pays better while I get my degree, or my current job will somehow surprise me and offer me a sizable raise.

Btw do I need to know C or any other low level languages in depth for this you think? It's not an issue for me to learn them, but I want to make sure I have everything I need for when it comes time. Thanks again for all the advice.

[loinburger]

For the phd exam I needed to know the theory behind semaphores and other low-level C-type constructs, but I didn't need to know any C syntax.

In all of my grad school coursework, I needed C for only one class (pattern recognition). The only weird thing there was memory management - the rest of the syntax was pretty much the same as it would have been for Java or C#

Btw, if you're ever in a job interview and you're told that Java is better than C++ or that C++ is better than Java, then agree with them. In both cases they're wrong, but there's no reason to let that fact sabotage your job interview.

[Kuciwalker]

Wait, y'all actually used textbooks in your CS courses?

[DriXnaK]

I'll keep that in mind. I hate Java though. What are your thoughts on C# vs Java btw?

[loinburger]

C# is infinitely better, if for no other reason than that it makes parallel programming very easy. With Java you have to do a bunch of dicking around with threads and synchronization etc. With C# you just need to create a parallel loop.

Java suffers from the "design by committee" phenomenon. C# does not.

The box-packing program I recently wrote runs in parallel. If I had had to write in Java then it would have run sequentially - even if I'd been able to write it to run in parallel, nobody else would have been able to maintain the code.

[DriXnaK]

Yeah, we had toyed with the idea of Java for a little bit in the beginning, but it became very obvious that C# along with Visual studio and the other stack of microsoft products/technologies beat the living **** out of Java/Netbeans and its other technologies. At this point I just can't think of a reason to use anything other than C#, at least for what we do.

[loinburger]

Java is better if you're running on Linux. That's about it. (You can still run C# on Linux using Mono, but you don't have anywhere near 100% compatibility)

[DriXnaK]

One last thing. Any need for functional programming capabilities? Like F# or something?