Slashdot Mirror
Starting an Education in IT?
AriaStar asks: "It's overwhelming to start trying to learn all the different technologies needed to go into programming. It seems that every type of technology assumes knowledge of a different one, which in turn requires knowledge of another, until it's gone full circle. I am interested in everything from Unix to AJAX to Perl. Things like HTML, Javascript, and SQL are like English, but then again, they're basic. Where is the best place to start? What is a good path for someone who learns quickly and easily, but who is simply too overwhelmed, to take?"
There's a lot out there. You might be able to get a sample of a bit of everything, but ultimately, you should probably pick a field and stick to it. If you have wide-ranging interests, pick a place to start, and let your journey wander from there. Pick a project, learn what you need to do it, and then move on to the next. But there's simply too much out there to "learn it all"
Take a course in basic programming structure first. You need the foundation of how to properly setup the code. By that I mean declaring variables, setting up functions, properly commenting (gasp) your code so that others can follow what you've done.
Once you understand the proper structure it is much easier to branch out to the nuiances of the various languages.
Studying IT is waay too general.
Here's a question you might first want to answer: do you want to be a specialist or a generalist? to be a generalist, you probably want to be a specialist at one field, security, database, web design, web programming..etc. The fact of the matter is that there is so much out there, that you have to be a bit disciplines to really get anywhere. Or you can find an IT job doing support or so, then on the side start learning the underpinnings of a certain vertical market. Perl + Unix + C would be very good fundamentals to look at. IT is not a field like medicine, it's hard to draw a path, since there's so much to know and so many options. I would consult with someone who is an IT pro and possibly follow his/her footsteps. But yes it's a challening feat to 'master'.
Insinct is stronger than Upbringing - Irish Proverb
Well my point was not to get overwhelmed. Just start where you can. You end up going around the hermeneutical circle several times anyway.
I agree that in IT there is nothing to fear but fear itself. So be technologically independent, get your hands dirty, get the hard stuff out of the way first then enjoy the ride back down, set yourself up an Apache server or install Gentoo. Compile "Hello World" with GCC. Try to use Emacs or Vi.
While graphical IDEs are all well and good, if you can put up with the command line, text-only way of doing things then you are a bit more future proof, know a bit more how it works and are ready when the brown stuff hits the fan.
My little Linux and tech blog
Don't just read about stuff. Do stuff.
Find a pet project, find someone with a pet project, volunteer to help a friend or community organization set up a website. It has to be something you're interested in doing or you'll avoid it and not enjoy it. Pick a target and shoot for it, O'Reilly and Google at your side.
The only way to really learn anything is to do it. I read a bazillion books about DNS/Bind but none of it sank in until I had to set it up myself. Same with SQL, Perl, Python, Template Toolkit, C++, etc.
It's annoying, it's painful, it's frustrating, but the only way to really learn anything is to get in there and get your hands dirty.
Knowing what to focus on can be taken as a research problem. Before actually trying to learn how to go about using the individual technologies, you need to research what each one is, why it is important, and how it will fit in with other technologies you would like to learn.
In some cases, the answer may be to learn a little about each technology progressively. Learn a little HTML, then some Javascript, then more HTML, then more javascript, more HTML, the DOM, XML, ...
When technologies are interrelated you sometimes need to learn a little about all of them before you can really start to master any of them.
Or just start trying to learn the most about what you're most interested, and when you discover another technology is involved -- start learning a little about that other technology too.
To keep from being overwhelmed, start with an idea you have for something that you have a passion for. Think of some idea and lay it out on paper (or visio). I wanted to mess around with Python and we recently had a baby, so I wrote an "I/O" and sleep tracking system for our infant daughter in Python.
.NET, Ruby... Pick one and stick with it.
For the most part, all the high-level languages are the same. There are differences and strengths, but at the point you're at most of these would be lost on you anyway. You already know Javascript, so pick another language -- and try to build your project in it. PERL, Python,
About 5 years ago I wrote a project for my brother so I could sink my teeth into PHP/MySQL. It's turned into a product we now sell and make (a small amount of) money on each year. The hardest part for me is sticking to the project with no real client to be beholden to -- hence MY need to build something I really want to make, not some example out of a textbook.
When I decide it's time to learn something new, I generally see what is "up and coming" and learn that so by the time I'm fluent, I'm much more marketable. Right now, Ruby looks like a pretty decent choice.
Good luck!
1. Discrete Mathematics (recommended textbook: Discrete Mathematics by Kenneth A. Ross and Charles R. Wright)
2. Finite Automata and Computability (recommended textbook: Introduction to Automata Theory, Languages, and Computation by John E. Hopcroft, Rajeev Motwani and Jeffrey D. Ullman)
Technologists who understand the fundamental theory can generally write more elegant, more efficient computer programs than pseudo-technologists who are ignorant of the science in computer science.
Once you have trained your mind on the fundamental theory, you will discover that most information technologies are quite simplistic.
Finally, one often overlooked subject is English. Learn to write and speak well in English. It is the fundamental mode of communication in the world of advanced science and technology. You may have great ideas, but if you cannot them to your English-speaking peers, then you are no better than a pseudo-technologist.
I know it might sound a little strange but if you can develop mainframe skills in 5 years you will be worth your weight in gold. So many of the current people that are mainframe system programmers will be retired by then and there are not many people learning the skills to fill their places. The one problem is it is hard to learn the skills because you can't just play around with the latest versions of mainframe software on your PC like you can with most windows and unix tools and applications. IBM has a lot of information and free learning modules on their site and several schools are starting programs to teach these skills to fill the knowledge gap. Plus you can learn how to run linux on a mainframe too so that you don't feel totally in the dark ages :)
The general progression for a CS student(with some variation) is: Basic Programming -> Data Structures -> intermediate topics: declarative programming, software engineering, assembly -> more specialized stuff: compilers, concurrency, AI.
I'd suggest that this is a good path.
I'd also suggest you avoid php like the plague. It is a language that does not encourage good programming practice among novices. Also, it is terribly designed: you should see the best that IT has to offer, rather than the worst.
I am so tired of people saying that installing linux is a good way of learning a *insert language/technogloy*. Read a book. Work on a project. Linux is not the greatest tutorial invented to teach - it is an open source source operating system that offers flexible solutions to many problems for people that want to learn the applications and environment.
/.
The best way to learn a certain language or technology is to work in an environment that you understand and build on your current knwoledge. Jumping into a different environment and leanring it can get you frustrated and distracted from the true goal and I am sure has lead more than one perosn to be discouraged enough to stop pursuit of what they set out to accomplish.
I am drunk, sorry if I am rude. But being a linux fanboy is not the answer to every f'ing question posted to
And there are computer curicula that do the same thing. And one possible strategy is to work your way through such a curiculum. But most computer geeks would rather work on their own, and they'd rather study technology they see being used, not what some textbook says are "basics". That's actually the way I learned.
And if you take that approach, you will constantly run into references to technology you know nothing about. So deal with it. It isn't a mortal sin to skim over the parts you don't understand well. So you're reading that PHP manual and it starts talking about SQL. That doesn't mean you have to drop PHP and start studying SQL. It probably means that SQL is something that you'll have to learn eventually. But for now, the thing to do is just skim the SQL parts and get a vague sense of what it is and how it relates to PHP. Maybe take a little time to Google or Wikipedia SQL — but don't lose your focus on PHP.
Of course, sometimes you will discover a new topc that's a prerequesite for stuff you don't understand. (Every programmer know aboutbinary encodings and two's complements.) But mostly not. I've been working with computers longer than most Slashdotters have been alive — and I still encounter places where the subject I'm trying to study seems to intersect subject I know little or nothing about.
What you're trying to do is somewhat akin to trying to understand paragraphs before you know how to construct a sentence or before you have an adequate vocabulary. An intro to CS class will provide the grammar and basic concepts that you can then apply to all these high level things.
I'm sure some anti-school posters will tell you just the opposite, but you should remember that schools are only as good as the teachers and your willingness to learn. CS classes developed to provide structure to what seemed, even in the 80s, like a vast quantity of unregulated knowledge. A year of classes at your local community college may provide more help than three years of wading through newsgroup postings and online tutorials you don't entirely understand.
Also, write at least one non-trivial assembly program. You may never use it ever again for anything real, but it will change forever the way you approach programming.
I think it's a good idea to start by focusing on one area and then branching out. Once you have a good understanding of one area you can safely move on to other areas of computing to help you understand how this works.
If you want to become a good programmer I would suggest doing something like this:
Start with C.
A) Write a program to open a file and display it to your screen.
B) Expand that to make a copy of the file backwards.
C) Make a new program that gets 2 numbers from a file (A and B) and prints A + B.
Move on to C++ I would read Ivan Horton's beginning C++
A) Start with a program that opens a file, reads a list of numbers, and sorts them.
B) Extend that so you can work with 10+GB files quickly. (This would be a good time to pick up a book on computer algorithms and learn how your operating system handles memory.)
Learn enough ASM to replace 2 functions in the above project with ASM
I would stick with java for a while.
A) All the above projects
B) A simple web server.
C) Extend the web server so it returns the contents of a small Database.
Learn enough LISP and Pearl to parse a java file and find all the variable names for each function.
At this point you should pick complex project you want to work on and using the best language for that task get it to work.
PS: Have fun.
It is IMHO a bad idea to learn C directly before C++. Good style C programs are usually bad style C++ programs, so you'll have to "unlearn" a lot of habits from C. OTOH, when learning C++ first, then it will be easier to go to C, because the compiler will usually complain if you use C++-typical idioms.
If you insist on learning C first, it's probably a good idea to learn Java before C++, because that way I think it's easier to get into OO habits.
It may even be a good idea to have some LISP experience before going to C++, because some advanced template techniques are basically functional style (I guess that's why many people shy away from those).
BTW, I guess you meant Perl, not Pearl.
The Tao of math: The numbers you can count are not the real numbers.
Here's my tip. Ignore anyone that suggests that you learn assembly.
It's like telling an architect that he needs to learn how to make bricks to be a good architect.
You can learn only so much in a given period of time. Don't waste it on something you can't use. I understand assembly, and it is at times a comforting feeling, but in all honesty I belive the time I spent learning assembly would have been better spent learning design patterns.
The Internet is full. Go Away!!!
It may even be a good idea to have some LISP experience before going to C++, because some advanced template techniques are basically functional style (I guess that's why many people shy away from those).
Naw, it's because C++'s template syntax is crufty and stupid, and the error messages are totally opaque.
I'd say Java (or C# - same thing, really, except C#'s libraries suck more), C++, Lisp, C++.
I got my Linux laptop at System76.
>It's like telling an architect that he needs to learn how to make bricks to be a good architect.
No, it's suggesting that you have to know how strong bricks are to build a building that won't collapse. And the Dean of the school of architecture where I went to school was there because a building he designed collapsed.
Anyone that suggests that you DON't need assembler is fooling themselves.
Actually you are dead right on this. I find it is better to be a generalist (you normally earn more for a start) than a specialist because you have the ability to actually think outside the square. In my job if there is a need to learn a specific language I just learn it (may take a week or two) since as far as I am concerned most Programming languages are similar.
Basically from a programming perspective I would not be as good as a specialist, however I don't normally program for a living and actually design at a much higher level so I normally tell programmers what is required.
Not sure about the Mac. I find Linux with some good stats packages work for me and If you want a serious document preparation package use LaTeX (it's free and surprisingly easy to use).
There ain't no such thing as proprietary standards only proprietary formats. Standards are by definition open.