Programming as a career

I had heard that C++ had lost out a lot to Java and Visual C++ to C#, and meanwhile C is still going strong for embedded systems, which C++ is largely unsuited to. But I have never been paid to write in any of those languages so I dunno.

When I was last job-hunting I put just about every language I had even basic knowledge of, and a specialist IT recruiter told me it was confusing and I needed to narrow it down. So don't do that.

I suppose that sounds obvious (writing 30-line programs in yr spare time or in lol college != doing useful work on a real major project), but when your core skills - and I'm not sure I have any - are kind of limited and obsolete, and your jobs to date have involved a lot of whatever nobody else wants to do that day but thinks you can pick up as you go...

I have a good front-to-back knowledge of web development, from html to java, and i do put it all on my cv, but I try to position myself as a java developer. Still most of the stuff i get from agents is for entry level html crap. Basically all agents are know-nothing cunts.

is it worth learning to use stuff like eclipse and that because i guess while most people can do java C++ fewer people will know how to work with the specfic development tools etc?

It might also be helpful to learn about recursion, as told by H.P. Lovecraft:

http://www.bobhobbs.com/files/kr_lovecraft.html

3. Make up for lack of degree w/ experience (fuck going back 2 school)

I kind of sort of have this covered at my current position, but I'm sort of an ad hoc backwater one-person band supporting a group in a hardware company, so there's not a lot of support for me but also not a lot of priority given to me vis-a-vis promotions, accumulating job skills past a rudimentary point, etc. It's good for financial security, but if I need/want a real developer's job, I feel a bit underqualified, thus thinking about doing school part time to get a BA/BS. And that goes to Sock Puppet's 3b; I'm a bit intimidated by the mathier stuff on my own, I do much better in a classroom.

I have some Java on me now, I'm taking a C++ class at jr. college this Winter Q.

People who say #2 don't know what they're talking about.

Re: outsourcing -- does it matter that I work in Silicon Valley? I tend to assume that if the general trend is towards outsourcing programming, SV would be the place where it happens first.

is it worth learning to use stuff like eclipse and that

i guess it'd be a good idea to know how how yer basic ide works but no point going out of your way to learn any particular one. once you know one you know them all, more or less.

OTM. Every company has their own specifics w.r.t. IDEs so it's probably not a good use of your time to worry about the environment outside of what you need to know to get things done.

There's no concrete answer to the get/don't get a C.S. degree question. As you might have guessed already it depends... amount of experience, depth of background, relative visibility, the type of job you're looking for and what industry it's in (something like game development is it's own Universe). I can only offer up anecdotes from my own career - I've had exactly two formal programming classes, but 18 years of job experience from IT garbage collector, to webmaster, device driver wizard, and database king. Like everyone else has been saying, have a broad familiarity with a couple areas of deep knowledge. If I was starting now, I'd make Java one of those deep knowledge areas.

Also, don't underestimate visibility. Write some code, improve on it, and then blog the results. Write comments on someone else's coding blog. I've gotten a couple of contract gig offers just because I posted some sample code to my blog. Go figure.

Lastly, get familiar with database basics. Nothing ridiculous like Oracle but just your basic MySQL/PostGreSQL fundamentals. I didn't set out to be a database guy, but I knew some stuff and now four years later I'm a full-blown DBA and pretty good at it.

> I had heard that C++ had lost out a lot to Java

i think this is due to a shift towards webapps (java servlets). i'm a dyed in the wool c programmer now doing java for a living because everything is now sat on a server being served by JBoss.

(never had a lesson in java, hadn't been invented whilst i was in education... lol, z80)

i am aghast at some of the job adverts i see, the list of disparate things they expect you to know. but a lot of this is due to agency idiots. (that said, if i think about all the disparate things i use at work...)

the thing they don't teach you at college but is fundamental when working professionally and/or in groups is version control. get comfortable with cvs or svn or maybe even git. at the very least know what they do.

i'd also recommend having something you can show people, like contributing, or starting, an open source project. or just having some flashy applets somewhere.

this stuff all takes years, 25 years in i'm still picking stuff up.

Hm, I like a syntax-highlighting text editor with parenthesis matching and suchlike, but for the actual business of compiling and running I'm still alt-tabbing to a console window, guess I'd better stop being so afraid of letting the IDE take care of it...

Ha, agency job ads. Seen ads asking for 3 years' experience in technologies which aren't 3 years old.

This is good encouragement to actually listen to the woman who's been trying to get us all to use svn. The only version control I've used was command-line only and didn't do much beyond locking the file and adding a timestamped comment. Seems they're a lot more sophisticated now.

file locking is so '90s.

Hm, I like a syntax-highlighting text editor with parenthesis matching and suchlike, but for the actual business of compiling and running I'm still alt-tabbing to a console window, guess I'd better stop being so afraid of letting the IDE take care of it...

Yeah, I programmed in notepad etc for years, before moving to a proper IDE, mostly because I figured I should learn to program without any tools to help me out. In hindsight that wasn't a particularly good idea, it just made thinks more convoluted and slow. Rewriting code was a major pain etc.
Learning an IDE (Eclipse, for instance) is hardly any work at all, since you can start off treating it as little more than a fancified text-editor, and learn the cool tricks as you go along. I mean, hell, I'd used one for over a year before I even heard of "extract method"! Sheesh.

Actually, that reminds me that getting comfortable with Maven or Ant is fairly quickly done, and something well worth doing once you're comfortable with Subversion (SVN) or CVS.
Also, testing frameworks. JUnit if you're using Java. It's both quick to learn, and well worth it; just don't let the annoying Test-Driven Design (TDD) fanatics put you off.

there are things netbeans cannot do that easily vim will do in a heartbeat but i can't live without the code completion stuff as i don't really know the libraries (and they change). plus java projects have such a deep directory trees and our stuff is so scattered that you end up spending most of your time in vim typing directory paths to swapping between files (could use ctags i guess)

eclipse i never got to grips with - there's no 'compile' button as it's continually compiling and i like having a compile button.

svn can be used on the command line but we use tortoise.

Programmers are the magicians of the modern age

Not the ones I've known. And I've known more than a few.

buncha putzes

Programming is the worst

DevOps is the worst.

Software Engineer USA™

we like to pretend we're architects and engineers and builders but we're really more like apprentice mechanics or those dudes that assemble pre-made furniture half the time

programming is great if functions and syntax are well documented. it is the worst thing imaginable otherwise.

I do enjoy the critical thinking parts of my mind that were unlocked by learning CS theory and programming over a period of time, but it really chafes me to see software developers think that they're able to solve non-software societal problems with that toolkit

professional googlers

Oh, I forgot that one. Software Architect. Classic.

I much rather SysAdmin, coder, developer, webmonkey/webmaster, script kiddie

Mind you, my end goal is probably to be a 'Software Architect', so I should lol carefully

yeah, really smart or tricky code makes you seem like a wizard but what it really makes you is an asshole if it's ever meant to be maintained

pretty sure my coding style has gotten progressively dumber on purpose

^^ thank you

I think software/systems architect is a fine title, even if my actual designs-buildings-and-structures friend recoils in disgust. I hate when people introduce themselves as "architects" without the qualifier.

Now, the part of the business where people use "architect" as a verb... not so good.

I do enjoy the critical thinking parts of my mind that were unlocked by learning CS theory and programming over a period of time, but it really chafes me to see software developers think that they're able to solve non-software societal problems with that toolkit

― a strange man (mh), Monday, June 9, 2014 5:09 PM (1 minute ago) Bookmark Flag Post Permalink

they probably also think they can throw together that application in a matter of a few days

programmers are horrible estimators

working with computers makes people feel very powerful because computers are powerful

bring back punch cards

ya, it's just funny because in canada you're not really allowed to use "software engineer", because, well, you're not an engineer. but in the states, it's quite common

I feel like 'engineer' is fair

you are designing and building something more abstract, but you are still designing and building something

I should be a software architect because I have all the artistic pretensions that building architects have

interesting how engineer and developer have become prominent as the job has become less about programing

you should be an engineering software architect or an architecture software engineer.

my dad once purchased landscaping software from a barnes & noble. i sometimes think about the people who developed that software.

i dunno, engineering would imply a discipline that's way more predictable than software. there's no ISO manual you can check that tells you the number of tests you need per 1,000 lines of java.

They might as well claim to be 'magicians of our modern age' and finish the job.

pretty sure metrics for unit test coverage in large companies are getting there

Yeah but are they as useful as "minimum cross-section for structs on suspension bridge given expected load" etc etc, I mean physical engineers actually know things

Programmers - the next job to become demand-weakened by too many people who enjoy it ad thus do it for free

That will never happen, I don't believe there are enough psychos out there who also love programming that'll fit the demand

as a cs student, I'm really not feeling the hyperactive "hackathon/build a startup in a day/be the next zuckerberg" horseshit mentality that i fear dominates. I mean, it's obvious that I'm doing this in order to get a job, why do i have to care about forming a startup

it's difficult, because you do need to keep up on new technologies and development practices, but to do so you end up viewing/reading/attending material that has an overlap with people who are WE ALL MUST MAKE NEW BUSINESSES and it's irritating

I guess the larger tech-oriented gatherings don't have that problem, but then you're at Microsoft's Build conference or JavaOne or whatever the fuck people go to these days

Every year my dad never fails to ask me to go to this one Oracle conference with him

on the bright side, the huge corporate ones usually have really cheesy entertainment

I think I went to one yeeeears ago with Mini Kiss, Battlebots, and a Rolling Stones cover band.

Silicon Valley Techno-Utopianism

programmers are horrible estimators

Guilty as charged.

Is it worth (re)learning front end development in 2023 or will it all be replaced AI soon?

Take home challenges have always been a part of some interviews as long as I've been in the field.

Sorry to hear that Vinnie!

I had an actual post apocalyptic nightmare last night where Javascript stopped working.

Front End development involves knowing what users and business analysts want when they can't tell themselves - it'll be the last thing to go!

This dropped on the AI thread, but certainly pertains to this one. If anything, more so. Worth watching even if just the first 15 min.

https://www.youtube.com/watch?v=JhCl-GeT4jw

i'm glad he likes nick bostrom because it means i don't need to take him 100% seriously.

I think he's clearly right about the general trend and mostly right about the speed at which programming norms will change to have AI do the heavy lifting. The offsetting trend will be how many more lines of code will be incorporated into every day real world processes.

I thought it nice of him to start (with uncommented code samples) by making sure that we know he's a bad programmer - the "the tests are easier to write than the programs" is a similarly brave declaration that he should be let nowhere near CS students.

It sounds like his vision is that software development will go through the following phases:

1) AI writes the code, humans still read it

2) AI writes the code, humans can no longer understand it

3) There is no more coding as such. The AI is the only program you need. It does it all.

who codes the coder?

Yes, that's a simplification. There will still be people coding the AI. But will be a tiny, very specialized workforce. Much smaller than the current legions of software developers. Roughly the size of the workforce that designs CPUs nowadays.

In line with that, a lot of this handwringing and prognostication over AI seems to be an inversion of the enthusiasm people had for increasingly higher-level languages allowing civilians to not deal with lower-level code but... C and assembler are still being taught decades later--we still use Fortran!

but that'll free us to use our time for art etc.

https://www.bbc.co.uk/news/uk-67302048

"Elon Musk tells Rishi Sunak AI will put an end to work"

future be like the Federation in Star Trek

yeah, I tend to think of this as just another iteration of making an even higher-level language, albeit this one is probably going to end up being a higher step than in previous instances. I think there will continue to be cases where a very specific and detailed logic will be required and this will still have to be written in detail one way or another by a human, even if it's just very specific English language sentences, but that's still programming.

xp

Are C and Assembler still being taught, or is it just that there's still jobs in those skills because of legacy codebases? I mean, you can still learn C on Pluralsight, I suppose...

I spoke with someone who graduated university a couple years ago and they did an assembly class, but it was more to learn how the processing works at the most basic level rather than anything intended for some kind of practical application (which was already the case 20+ years ago when I was in school).

I don't know if C is still taught by itself, but C++ is still taught and you kind of get the basics of C with it.

I assume that you'd need C if you wanted to get into OS/Unix stuff? I also heard that NASA uses C? And lots of embedded devices?

As for assembly, doesn't that get used in heavy duty graphics programming (i.e. video games)?

Not as much anymore. It's all fairly abstracted down to frameworks and licensed engines. And a lot of the GPU stuff is via both frameworks and drivers.

So there's some low-level programming being done but it's being broadly reused because very few studios will build their own engine from scratch rather than licensing Unreal or Unity.

i don't think many large engineering orgs are writing new userland code in C proper.

os/device level is another matter.

but C++ will never die, partly through legacy systems, but partly because it's what e.g. pytorch and tensorflow are implemented in. and C is a prereq for most C++ dev work.

the closer i've got to ml platform and infra, the more jobs i see that i can't apply to because i don't know c++

modern c++ is actually pretty decent. i picked it up on the side over the past year or so since qt in c++ is the only option for desktop linux applications (lol) that doesn't make me want to walk off a cliff.

i think everyone should give it another look, esp if it's preventing you from applying for cool shit

I've got back into c++ recently, and loving it. sadly not for any career aspect, i don't think I could handle the pressure of applying it to in a job.

So, uh, do I need to start thinking about a third career?

I've been thinking the same lately, or maybe going back to my second career. My job search is lasting much longer than I hoped

But at least I got a story from it. I recently applied for a position where they had me take this homegrown multiple-choice online assessment which was straight up hilarious - couple questions flat out had no correct answer, couple more ambiguous, and several arcane code questions where they ask what the output would be but you can just execute the code in a browser and see the correct response. They scheduled me for an interview and then I saw they have literally the lowest rating on Glassdoor I've seen: 1.6 out of 5. Thought about cancelling but deciding to just do it for the practice. Place looked awful and all work is onsite. Company owner asked me a few things illegal to ask during an interview. Backend is a software I've never heard of that was apparently discontinued in 2007. They were even hesitant to tell me the benefits when I asked - turns out they have no PTO or sick days for one year, no insurance for three months! They did send me an offer that's at the very bottom of the range I was asking but I decided not to bother going further, I feel like there's no guarantee they'd even pay me

smdh JS:

-1 % -1
-0

Fortunately -0 === 0 but still: lol

unary+ converts an array or boolean to a number
double negation converts an array to a boolean

+[] == 0
+!![] == 1

the video silby linked eons ago that's mostly about javascript still pops into my head
https://www.destroyallsoftware.com/talks/wat

saw somebody yesterday claiming

"2" + "2" - "2" = 20

which i guess works if + on strings is append but - is subtract

I've started looking again (thankfully a voluntary search this time), but I've been whiffing pretty badly so far. Just 3 phone screens, and nothing beyond that (though I'm not applying every day), and one thing that I've noticed is that I hate the "walk me through a complex project you've worked on" prompt. Maybe this is confirmation bias but I feel like every time I get asked that, the interviewer is at best unimpressed with my answer. What exactly do they want to hear? (Obviously that depends on whether it's an HR person vs. someone on the tech side, but I don't think I've success with either.) If I'm being honest with myself, I don't think I have any impressive projects in my career, but maybe I'm not being generous to myself?

Imo the point of that question is to demonstrate some combination of: you are good at explaining something, you have done actual work, you made technical decisions, you worked effectively in a team, you worked effectively with limited direct supervision, you worked effectively with legacy code. It’s not to dazzle the interviewer with like “I wrote full self driving for Tesla!” Just pick something from the most recent gig that you can explain what you did and what the impact was.

For a phone screen, I don't really know what they'd be looking for. Probably some combo of impact and your importance to the team? I've never been asked that question on a screen

But in a technical interview, I've struggled a bit with that question. The only project I've worked on that truly felt complex was in a 16-developer team but I only worked on a couple parts of it in depth and I struggle to explain the project in full

I should mention I did finally get a senior SWE job late last year which I'm still working in now. So far so good: place is a bit behind in tech stack but the people are good to work with and open to improvements. Very relieved to be out of the search for now