Best Language for Employability?

[Ehtyar]

Hi all.
I am currently nearing the end of my education, and though i am versed in various programming languages, my choices tend to be rather old fashioned. What I'd like to know is what everyone considers to be the best language to learn for the sake of employability. C and/or C++ does not get you employed very successfully anymore unless you are an expert, which i would not consider myself to be. I am not interested in Java, though i could tolerate dotnet if necessary. Particularly, i would welcome the opportunity to learn C#, were i to find an IDE that was not by Microsoft and not written in Java. Any suggestion is welcome, for the sake of others as well as myself.

Thanks, Ehtyar.

[jgpaiva]

disclaimer: i love java

were i to find an IDE that was not by Microsoft and not written in Java

Pay attention to something: if you're going to work at a big company, you'll have to use the IDE they give you, which probably will be visual studio or eclipse. Thus, you're better off learning it before you start working. If you don't like it, it's yet one more reason to learn it before starting to work with it.

If you don't like java, i'd definitelly go for .net.

[Ehtyar]

The reason I'm not worried about the IDE is because i need to learn the language before using it, and i'll have no idea what IDE they're using anyway, so i might as well learn on something that meets my requirements, thus making the learning process easier, then when I'm forced to use someone elses IDE, the experience will be unpleasant, but i will have the language down pat

Ehtyar.

[iphigenie]

Well from my perspective there seems to be (in the UK) a lot of jobs for all languages - that includes old timer languages like C and Perl, and more recent ones like python, java, c#, ruby and resurgent old style stuff like erlang (telcos) - although knowing IDEs and frameworks (and listing them) will often help your CV hit the right keywords.

According to a recruiter I spoke to last week, Perl is very hot at the moment (I am trying to recruit both junior and senior programmers) so is Java again

[f0dder]

The reason I'm not worried about the IDE is because i need to learn the language before using it, and i'll have no idea what IDE they're using anyway, so i might as well learn on something that meets my requirements, thus making the learning process easier, then when I'm forced to use someone elses IDE, the experience will be unpleasant, but i will have the language down pat

You don't have to "learn the language before learning the IDE", there's a lot more to an IDE than silly little code wizards etc. For the dotNET languages, it would seem silly not to use the IDEs btw., since the GUI designers make everything so easy (and that's the frigging point of dotNET, to be easy and fast to use!)

[Ehtyar]

For the dotNET languages, it would seem silly not to use the IDEs btw., since the GUI designers make everything so easy (and that's the frigging point of dotNET, to be easy and fast to use!)

I make it a point to learn languages the hard way before the easy way. In most cases i actually never start doing it the easy way (no radasm for asm and no vs for c). As an example, for my first 6 or so months of win32 development, i did all my dialogs from scratch by hand (i use resed now).
Another aspect is portability, which visual studio is intentionally made not to be, as much of my coding takes place on several computers, many of which i do not have administrative rights on. Reading a C# thread the other day (link) i found an interesting c# IDE (link) that im considering trying which should suit me in the portability aspect.
Thanks for the answers so far guys, your post was particularly reassuring iphigenie, there may be some hope for me yet (better download a perl interpreter huh  ).

Ehtyar.

[KyleLanser]

Elitist attitudes about using VS, or eclipse are childish in inefficient.
You use the best tool for the job, doing asp.net web development, the best tool is VS.NET. Doing java development the best tool for the job would be eclipse with 5 or 6 major plugin's installed. Doing things the 'hard way' is a waste of time and resources.

For learning "C#" and "Java" the language is just a bunch of for loops and OO relationships. To be a C# or Java code monkey, you gotta know the API, and then, to be a productive employee you gotta know the tools.

As for languages to choose for employability... I always just did what i enjoyed, and because its what i love i want to be the best, and so i've always had at least 1, usually 2 contracts going.

so figure out whats fun for you.

Web development? asp,php,flash,ajax

Server development?
ftp, ssh, apache, subversion, all open source so start playing.

Windows Desktop development? all the cool stuff on this site

Graphics programming? directx, opengl, sdl,

embedded systems?
hobbyengineering.com - BoEBot great fun.
"javalin" - embedded java programming
http://bigelowaerospace.com/includes/jobs.php?jt=Software%20Engineer

pocketpc development?
write apps that interact with gps modules
alternative data entry - "dasher" (again open source)
write apps that interact with rfid tag's, and write your own inventory management app for moving companies...

there are just so many ways to go...

maybe you wanna do some computer science Research, go back to school, work on computer vision, 3d visualization, Voice recognition, voice sythesizers, Darpa great race kinda stuff...

so more than languages or api's, both of which are just tools to get a job done, decide what kinda JOB you wanna do. There are so many jobs out there that if you love what you do, you'll be good at it, and if your good at what you do, you'll probably find yourself a job doing what you love.

Kyle

[Eóin]

Elitist attitudes about using VS, or eclipse are childish in inefficient.
You use the best tool for the job, doing asp.net web development, the best tool is VS.NET. Doing java development the best tool for the job would be eclipse with 5 or 6 major plugin's installed. Doing things the 'hard way' is a waste of time and resources.

I really have to agree with Kyle here. Sure knowing how to do things the hard way is an advantage, but persisting with it is childish.

[cnewtonne]

Instead of asking yourself or others what do I need to learn to get a job? You should, instead, ask yourself how can I become so good (at whatever you like doing) that employers beg me to work for them to do what I loving doing and pay me for it.

It may sound idealistic and hard to think this way. Yes, it may not be a short-term approach. But long term, this is the only way to survive, make a good living, and retire doing it.

Focus on become a guru in something not simply finding a job. There are a lot of real-life examples all over around us. There is a guy called Mark Blank. Mark developed a nitch Palm OS application called ChatterEmail. His abilities and skills showed very good in his creation that this app became the best man-known email software for the Palm OS. Even better than what Palm itself gives out natively on their Treos. He was so good that Palm offered him a position and he took it, they bought his software, and I'm sure he's pretty happy with it.

It works this way ...

To make a good living (over 6 figures to start with), you need to be a guru in something. You won't become one till you spend a hell a lot of time learning/doing it. You won't be able to do so unless you love it. This were you need to start.

Good luck.

[f0dder]

I make it a point to learn languages the hard way before the easy way.

The language is easily learned, you probably mean "the environment" which also includes IDE and standard libraries. It's usually a good idea to learn what's going on under the hood, but imho it's not necessarily best to start doing this.

Car analogies are ever so popular (and ever so stupid), here's my attempt at one: first you learn how to drive the car well, and the various traffic rules - then you can start tuning it.

If you're going to learn dotNET, you really should familiarize yourself with the RAD GUI tools, otherwise there isn't really any point in using dotNET at all, imho. Not saying that you should reduce yourself to a click-and-play GUI builder, there's obviously a lot of other stuff than that - but do use the ease-of-use GUI stuff so you can concentrate on those other more interesting subjects.

[iphigenie]

Once you have learned 2 languages it is easy to pick up new ones. I'm a language bulimic, I find it so interesting

I am also a framework bulimic, again I find it so interesting - the key dilemnas are always the same, and seeing what choices and compromises people choose to make and what the resulting strenghts and weaknesses are is fascinating.

I'm hoping to go back to more purist stuff, and learn lisp next - I doubt it will improve my employability but with about 12 languages under my belt i don't think any other language would. Besides I kind of moved out of development for now, except as a hobby - i just keep up to date so I can still jump in and help and keep my developers on their toes.

Any one in southeast london/kent want a relaxed perl job, let me know

[Grorgy]

A quote from a manager/supervisor type in a development i was involved in once
'I don’t care if it works or not, I want it on time and on budget' so depending as to how prevalent that view is, it may not matter a lot what you know 

[CWuestefeld]

Ehtyar, your attitude is convincing me not to hire you 

For an entry-level developer I want you to be familiar with the platform and able to work enthusiastically with the team. Your argument here violates both -- you are refusing to work with the team (viz the normal dev environments, most likely VS.Net by far) in order to avoid using important parts of the platform.

I really don't care if you can do from memory what IntelliSense will prompt you for anyway, especially since with IntelliSense you can save time by not having to type the whole word. I don't care if you know the names of every little Property that the GUI designer will generate into your code; you'll (almost) never need to do it by hand, and you're just wasting time typing it yourself.

For an experienced developer, where you've shown the ability to function in many different environments, I don't care a lot about the specific language. As long as you've got a good track record and can demonstrate mastery of something in the same language family, I'll be satisfied. That is, I'm really looking for C#, but any C++-derived language is sufficient to show that you've got the mindset for OO in a statically-typed structured language.

Another bit of advice: a single programming language doesn't go very far anymore. A non-trivial application is going to include elements from many different areas. You're likely to need to understand other things like SQL, XSL, etc.

I also want a lot more than just programming. I want to know that you're good with design and documentation. I want you to show me that you'll be good through the whole lifecycle (requirements gathering including customer communication; design; implementation; testing; deployment; support).

[Ehtyar]

I know the SDLC, thank you, in fact i know most of what all of you have told me. Seeing as you're all far more interested in criticizing my current approach rather than making suggestions as was asked, i will have to assume I'm better off continuing down the Systems Administration path as i have been the past year and keep software development strictly as a hobby, to be used in the workplace only when necessary. Thank you to all those who made suggestions.

Ehtyar.

[edit]
You may also be interested to know i was a junior mechanic from age 11, and could replace a fan belt among other things before i received my learners license, perhaps this is just habit for me
[/edit]

[mouser]

CWuestefeld response sounded harsh to me, i don't think he was meaning it to be.  And some of the other posts did as well, but if you look deeper you will find some reasonable statements.

First it should be said, that you are in the very small minority of programmers who actually jump into the deep end and are capable of learning the nitty gritty stuff that can be so important to learn.  Many people will avoid such stuff at all costs, and lean on their tools like crutches.  So that's a good thing.

But some of the other comments people have made are also quite true, which is that in terms of getting a job, there are a lot of other skills that are probably as important or more important that skill at a particular language.  There are a few books i would really recommend for this kind of stuff, books like:

• Code Complete
• The Pragmatic Programmer

The last comment that really bears repeating is to find a niche that you love and get good at it.  Programming is too much work, and has too much of a potential to be so fulfilling and enjoyable, to focus on an area you don't enjoy just because there are a few more jobs there.  There are enough jobs in almost every language and area that if you are good and enjoy it you can find work.

Oh and let me also agree with KyleLanser, that there is no reason to limit yourself to just one tool or language -- they get easier to pick up as you go along, and you are learning key skills of programming that don't have to do with any particular language.  The more you program, regardless of what language, the better you get at coding, no matter what job you take.

[CWuestefeld]

CWuestefeld response sounded harsh to me, i don't think he was meaning it to be.

I'll consider myself duly rebuked  To clarify, I didn't intend to be mean, but I did intend to make a firm statement, because I think it's important.

I've just come from a couple of months trying to hire 3 mid-to-senior level developers. The quantity of people available (at least that we could get through 2 headhunters and various web sites) is extremely low. And the vast majority of candidates I just can't see being any kind of asset to my team.

In Ehtyar's defense, many of the candidates I can reject because a 20-minute phone conversation -- or even just their resume -- reveals that they don't have the depth of knowledge that I think is necessary for a developer grow to the point of making correct decisions independently. The number of CompSci graduates with a couple of year's experience who can't describe a virtual function is appalling. To the degree that Ehtyar's plan fills that requirement, I owe him an apology.

One of our biggest sources for entry-level applicants is Rutgers University, as they're just 10 minutes from here. I was surprised to learn that the only language that a CompSci student there must learn is Java. I've got nothing against that language. But how can you understand how a computer works without having experienced assembly language, even if you'll never actually use it? How can you understand the benefits of encapsulation if you've never been burned using global variables in C or Fortran or something? The list goes on; the point is that to advance to real expertise, you need a broad foundation underlying your experience. But CompSci curricula seem to be building more academics, while what we need in the real world is really software engineers.

I don't need rogue programmers, not even prodigies. The scale of the systems we're building is such that no one can do it alone. And no matter how good you are on your own, if you can't do shoulder-to-shoulder with others, you'll never be any more use to me than writing backend batch import and cleanup scripts.

So when I'm interviewing someone, I look for:

• Correct answers to my questions
• The way you answer. Your choice of terminology, e.g., reveals much of your underlying understanding. And if you don't know the answer, trying to figure it out orally reveals your thought processes as well as understanding, which can be just as good as the correct answer.
• Ability to accept constructive criticism. When you don't know an answer and I explain it to you, I want to see you demonstrate that you now understand the concept. This shows that you have the foundation to learn new stuff as well skills to work and play well with others. Ehtyar failed this dismally by refusing to accept the suggestions of several posters, insisting that his own approach was sacrosanct.
• Enthusiasm for the problem domain. I want someone who is interested in the things we're going to do. If you're just dreaming about being in, say, graphics development, then go out on a limb and tell that to such a company.

• Communication skills. I don't care how smart you are, if you can't make your ideas understood by the rest of the team, it's all wasted.

[iphigenie]

It's the same in the UK nowadays - people can get a computer science degree having learned just one language and UML and side, easy subjects (multimedia???). And when you talk to them they haven't ever learned anything on their own. These people might be useful in a large corporate machine, but they certainly are no use to someone with a team of a dozen where all need to pull some of their weight)

I have a degree from switzerland, theoretical physics... and we had more programming & computer science on our degree than many computer science graduates I interview...  (including assembly, but then that makes sense from an experimental set up point of view)

I think it comes in countries where there's this strange idea that 60% or more of people ought to have a degree, which again seem to happens where you have had a hiring culture ruled by HR departments (i.e. the people interviewing have not enough clue to be able to evaluate a candidate's experience and abilities....)

Also weird in the UK is that even people with 20 years experience and a degree feel that they have to put their A levels (high school grades!!!) on their CV/resume. And that many HR people care ("ooh, he only had a B in maths" at 17 for someone aged 40...)

rant over

[Ehtyar]

I was comfortable with these two replies until that third last point from CWuestefeld. What you all fail to understand (despite the fact that i've said it several times in not as many words) is that I am completely aware that i will not capable of functioning in a business environment until I'm using the proper tools. You also fail to take into account that most of the computers I currently use do not have visual studio available to me, which entirely nullifies the point.
I am, however, capable of learning the language my way (which can be done on a machine without visual studio) until i am able to use visual studio regularly. I'm afraid constructive criticism does little when when the person you're dictating to is not capable of what you suggest.
So once again, if anyone has any further suggestions, please bring them up, and if anyone wishes to reply to this post they may, but I'm not particularly interested in hearing more about using the proper IDE. Constructive criticism on other aspects of my learning is welcome.

Ehtyar.

[mouser]

I want to underline one thing especially that CW mentioned, because it's a really important for very smart people to absorb (emphasis mine):

I don't need rogue programmers, not even prodigies.

One of the things that can be particularly hard for really smart people to learn and accept, is that brilliance can only get you so far.  In fact, there are lots of fields where being really smart can be a very real impediment to actually getting things done.  My father always quotes his thesis advisor who said (paraphrasing): "Be a genius on your own time, while you're here just get the work done."  I wish i had been able to absorb that lesson.

The bottom line is that when working in the real world and with other people is that there are a lot of important skills, and it's often less important to be the smartest person in the room who knows the language best, and much more important to be the person who knows how to learn what they need and get the job done well and on time while working well with their co-workers, etc.

Having said that, there are lots of more independent ways to live if you are the super smart type who doesn't play well with others, just be aware the path you are taking.. 

[mouser]

Another thought:
It is impossible not to benefit from practicing coding and learning new things.  Do those two things and it's hard to go wrong.

[mouser]

This is good timing, the author of Code Complete (Steve McConnel) has a blog, and his latest entry is:

"How to Self-Study for a Computer Programming Job"

http://blogs.construx.com/blogs/stevemcc/archive/2007/08/12/how-to-self-study-for-a-computer-programming-job.aspx

[Lashiec]

One of our biggest sources for entry-level applicants is Rutgers University, as they're just 10 minutes from here. I was surprised to learn that the only language that a CompSci student there must learn is Java.

It's the same in the UK nowadays - people can get a computer science degree having learned just one language and UML and side, easy subjects (multimedia???). And when you talk to them they haven't ever learned anything on their own. These people might be useful in a large corporate machine, but they certainly are no use to someone with a team of a dozen where all need to pull some of their weight)

What the... Where are those universities? I want to finish my degree there! UML, Java and multimedia... easy! </sarcasm>

[iphigenie]

IMO UML and multimedia sure are easy, and just one language is just not enough to be a software professional at degree level. People in all sorts of other majors learn one programming language and a bit of software development! You need multiple languages, architecture, modeling (and not just the use of UML), problem solving, algorithms, etc. You need to learn all the basics, solve certain problems several times in different languages, pick up algorithms and data structures, so you understand how it all fits together and why some of the current patterns and practices are good (and where they might be weak)... You need to pick up knowledge, but also experience and craftsmanship...

Well when we interviewed students for their year out, I wasnt impressed. These were students from places like Leeds and DeMontfort, so not totally lame unis from what I hear. To me, there was far to much application and product specific courses, and many superficial topics (and not just in the 1st year, where they are justified).

Or maybe I have the wrong expectations of a "computer science" or "software engineering" degree. But to me it does imply a strong base and a lot of craftsmanship, not having spent maybe 10% of the whole 3 years actually writing software...

I think it is a necessary result of policies that say that x% of people must get to go to university and we need x% more technical graduates... The only way to achieve this kind of stuff quickly (considering that you would have to wait 20-30 years if you did it by improving the school system and make sure more people end school capable for university) is to create more subjects and make things easier.

The side effect is that you used to be able to assume that about 60% of people with a certain degree actually had acquired a pretty good understanding of the subject field. Now I fear it's more like 30%.

Nah, i'm not a cynic 

[Ehtyar]

I'm afraid i can't disagree on anything you've said iphigenie, except for one point. When you suggested creating *more* subjects i almost yelled at my monitor. It is my belief that this is precisely our problem. When i went for my Software Design diploma last year, it infuriated me that out of an entire week, *TWO* periods only were spent in actual software design. Now i understand that there is much material to cover in addition to the software design itself, but to have only a fifth of the course dedicated to the design of software is absolutely outrageous. In my opinion the number of subjects should be cut in half at least with more consolidation and less redundancy in the content of the off-topic (for lack of a better word) subjects and a firmer focus on the core subject of a course.
Aaaaaanyway, getting on to what i intended this post to be...I wish to announce i have officially found my portable C# IDE.

SharpDevelop appears to have covered every single one of my requirements (we can forgive them for coding it in C# ) and then some. It is an open source IDE aimed specifically at C#, but does support various other languages (VB.NET, Boo etc). It has full support for WinForms and the soon to be released version 3 will have full support for dotnet 3.5, though version 2 does support 3.0 in a far lesser capacity. It is fully portable after a couple of modifications to the source and a recompile (by default it saves to %appdata% and %temp%). You may find a list of its features here.

Ehtyar.