|
Bad code comes from (lack of) expectations. Most devs are capable of writing good code. But they are human beings, and humans are lazy when there's no penalty for being lazy.
It's pretty hard, as a peer, to get people to write good code. You can't just say, "Why do you write such sh*tty code?" because people will (rightfully) think you're an a**hole if you trash-talk their work. You may get slightly more traction with a helpful comment like, "Do you know about RAII? If you use RAII here, you won't have to explicitly delete all those pointers."
(Huh, codeproject auto-censors my cuss-words. I thought the authors were doing it...)
You could try to get your peers to take up a static checker tool or de-linter. Sometimes even a reformatter helps, to get the whole code base into a common-looking form.
You may have more leverage by encouraging your manager to promote the use of these tools. As I said, humans are lazy when there's no penalty. If your manager wants you to write better code, that's a pretty big penalty for laziness. The Holy Grail would be to convince your manager to adopt a code review methodology.
When you ask your manager for help improving the code base is generally when you learn if there is any hope. If your manager views any attempt to review or refactor the code base as time wasted, then you're never going to get any traction. It is my sad experience that most managers are this way, making the usual fallacious arguments about efficiency and how the company can't afford to slow down. When you hear this, all you can do is polish your resume. You'll want to get moving.
|
|
|
|
|
In some cases I find that management don't want to hear that such and such code is crappy, because they probably went to a lot of effort to find the person(s) who wrote that code, and paid them a lot of money. To later hear that the code that was delivered was "sh*tty" probably sounds like you are pointing the finger at them.
You need to find a way to turn around their viewpoint - that the code isn't necessarily bad, but there has been improvements in technology, and refactoring this code could lead to improvements in A, B and/or C (as appropriate).
Once you have a few successes with such efforts (keeping things as positive as possible), you will find it easier going forward to get approval for cleanups like this.
Not an easy task, and sometimes you just gotta vent too.... 
|
|
|
|
|
Andreas Mertens wrote: Not an easy task, and sometimes you just gotta vent too..
Amen to that brother!
Jeremy Falcon
|
|
|
|
|
"...must I accept... if people don't care about the quality of their work then you can't force them to?"
yes, you must.
|
|
|
|
|
Sad but true.
Jeremy Falcon
|
|
|
|
|
This is what's worked for me in a similar situation: lead by example.
Rather than say "this is cr@p, we need to change it all", instead work on what's assigned and make changes around the code you touch in order to improve it. Don't talk to anyone about refactoring, just do it in the course of doing the work at hand. That is what refactoring is all about anyway. It's not a separate task. Eventually, hopefully they can see the changes and recognize for themselves the improvement. Then they'll have a template to follow.
Improving an existing code base is a marathon, not a sprint.
|
|
|
|
|
Bad code will have bad side-effects because you will underestimate how bad it really is and that little bit of re-factoring will cause you to dig yourself a huge hole because "production" will blow up some time in the future for no apparent reason...
There will be a few times in your career when you will get to work on a new, shiny product; most of the time, you will be fixing what the "team" built before going off on a new adventure.
... Unless you happen to be the owner of the company (which should be your goal IMO).
|
|
|
|
|
I'm kind of in a similar situation. It doesn't have to do with code, but my company can't keep developers very long because they treat them like crap; mules to get stuff done but are given almost 0 respect from the people asking for their help.
I started bringing up the problems, being really vocal about them. Management started to listen and I thought it was sincere, until they had a company meeting where they basically said "stop complaining and get back to work". I ignored them and kept at it until I was threatened with disciplinary action because my negative attitude was affecting the company.
So, here's what I've done to survive while I look for another job.
1. Pick the most important battles. I think I drowned them with so many problems they didn't know what to do with them. So I decided to ignore some things and tried to focus on a few of the most important things.
2. Talk in detailed specifics. When I told them about widespread general problems, they heard "Everything here sucks"; which was true, but apparently offensive. So I've started talking about specific people or specific conflicts and suggesting a change to this one, specific event that might have a broader effect for the future.
3. Be careful how it's directed. The management team here is thin skinned. They took the accusations personally and heard it as "You're bad at your jobs". They're in top management at a mildly successful firm... so they've done something right. So I'm really careful that what I say to them can't be interpreted as a personal attack, but is more a suggestion on a way to improve this 1 thing about this 1 small event.
4. Support the hell out of what I control. Since management won't support a company change, I made changes where I could. Treats, lunches, drinks, etc are regularly brought in with a message of appreciation and understanding how much they work. I talk about the developers, the kind of crap they go through, how important they are, and how talented they are to anyone that will listen. If management won't make a change, maybe I can convince individual people to change how they work with them.
I hate politics and have a hard time with people who get offended by open discussion about differing opinions, so I don't expect to be here much longer. But this has seemed to help shift the perception of me from a negative complainer to a "team player". Maybe we'll even be able to keep a developer around long enough to actually finish a project.
|
|
|
|
|
I think you've hit the nail on the head. I know a lot of devs don't really stand up for themselves, so I guess people can get used to pushing them around. Which is pretty sad if you ask me considering... you know... someone has to do the actual work.
Jeremy Falcon
|
|
|
|
|
It's very frustrating that, while having the company's interest at heart, you feel like you have to 'fight' for doing things right. To me, this is a clear sign that it's time to move on.
I always try to do things to the best of my abilities / time-constraints / etc., and while I'm not above delivering code not 100% up to my own personal standards due to good reason(s), getting your efforts squashed every time because of people's idiocies and/or politics will drive me out of any gig very, very quickly.
This is all personal, but my 2¢ would be: do your best with what you're given, honestly. If you get shut down for stupid sh*t, leave. Others will be glad to have you on board.
|
|
|
|
|
Thanks for your insights.
Jeremy Falcon
|
|
|
|
|
The only solution I've ever found for that problem is to lead by example. It sounds very New Age-y, but it seems to be true. Implement solutions to the code quality problems in your area of responsibility, let your boss know what you've done, and why. Offer to help with the worst problem areas. The difficult part of all this is doing it without looking like a grandstander and pissing off your coworkers.
Software Zen: delete this;
|
|
|
|
|
Well said. 
Jeremy Falcon
|
|
|
|
|
Thanks, Jeremy - it's the benefit of learning how not to be an asshat of a coworker  .
Software Zen: delete this;
|
|
|
|
|
Yeah, I have been in a similar boat to you in the past about other people's code. Usually, a lot of code has been turned quickly out by programmers transient to the project and I'm the guy left to pick up the pieces. This used to be really frustrating until I was reminded that it is precisely because I'm the 'better' developer that I'm the one doing all the refactoring.
So, it is important for you to highlight the deficiencies in the code base but unfortunately it is there and everyone is stuck with it. You could argue for a complete re-write, but the is unlikely to happen just because you say the code is 'bad'. If it is working and/or being sold, your management won't give a fig about your problems; and, after all it is your problem so no complaining and get on with your job 
In this situation there is one thing I say fairly early on and it goes along the lines of: 'this code isn't where I want it to be. All the while I'm working on the code I'll seek ways to improve it incrementally until it is. Or until I'm reassigned'. On that basis you'll have to accept the process could take years and make sure your management know this is a normal part of your day to day activities of fixing faults and CRS.
Then I will look for opportunities for refactoring, on the justification of a fault to fix or a CR. Always ensure you've got a good testing regime in place to attempt refactoring (that doesn't necessarily mean automation) and you are able to mitigate the risks of the change. e.g. rollback, incremental changes.
It is important that your management build confidence in your ability to refactor code, so pick your refactoring targets carefully to begin with so you don't mess up. (You may not know what past history your firm has had along these lines).
Sometimes refactoring is simply not appropriate and a direct rewrite would serve you better. Again choose your target carefully, but in a rewrite it will allow you to develop and test in parallel and to establish a core code base to build out, from within.
You could champion a set of standards and best practices within your group. Rather than take on the mammoth task of refactoring yourself, perhaps a change in team culture would help. Don't focus on variable naming conventions and the like. That's irrelevant. Focus on things like project structure, where library methods live, code templates snippets etc. Consider your standards in three tiers - must adhere to (e.g. things that break the build or can cause serious problems such as memory leaks), should be applied where appropriate (e.g. usually best followed, but there may be good reasons for deviating from standard) and advisory (e.g. there are various different ways of implementing IDispose depending upon the circumstances).
Finally consider ways to present your arguments in objective terms, because the 'right' way to code is incredibly subjective. Here is a list of ways to highlight the problems in the code base:
1. Static code analysis.
2. High incidence of faults in a given area; make sure your replacement doesn't generate as many faults as before!
3. Performance problems in a given area.
4. Obsolete technology.
5. Complicated interactions with other systems and between sub-systems.
6. Complicated build and installation process. If your code doesn't build out the box, requires 'installing' before you can compile it etc Streamlining build->release->test should see a measurable benefit in the faults database and project lead times.
Hope this helps.
|
|
|
|
|
Nothing to do with the previous thread with what some might say is a similar name (but it is clearly different, so I don't expect copyright problems).
Ten minutes ago, I used my winio tablet (the only winio device I have) to browse some pictures with the superb IrfanView.
I double-tapped the screen to go to full-screen mode, and looked at what I wanted to look at (wiring diagrams, not booty, unfortunately), and then...
And then What! I couldn't get out of full-screen picture-browsing mode! Tap and double-tap (click and double-click) are browse forward and backward!
The windows button only brought up the (hidden) taskbar one time out of five, but it would only accept one tap before going back behind the full-screen image, and there was bugger all there within one click to help me, because activating the @&#@^&@ on-screen keyboard was a complete waste of @&#@^&@ time!
It's got NO ESCAPE KEY, which is what you use to escape full-screen mode, and NO ALT KEY, so I couldn't even Alt-Tab out of it or use the three-finger salute!
TF I had a file manager in my Quick-Access toolbar, which I could use to browse to the weven OSK that I'd copied across (which has all the required keys, despite not being part of the "Best Windows Ever!") (Where's my wu mao?!?)
Fortunately, if you run into this problem yourself, the winio on-screen keyboard, which is as highly-professional as the rest of winio, does have a full page of emojis, so you can spend the rest of the device's life looking at them, then throw it away and buy an android tablet.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
|
Now .... If you had a recent system image of your Win 7 setup .... Just thinking out loud! 
Get me coffee and no one gets hurt!
|
|
|
|
|
I could probably do a System Restore, but that's not the point.
MS has no right to infect my PC. It's illegal and amounts to nothing more than hacking.
If it's not broken, fix it until it is
|
|
|
|
|
I doubt that a system restore will work ... Anyway, good luck!
Get me coffee and no one gets hurt!
|
|
|
|
|
Kevin Marois wrote: It's illegal and amounts to nothing more than hacking.
It's illegal for them to install software onto their own operating system? 
There are only 10 types of people in the world, those who understand binary and those who don't.
|
|
|
|
|
Remember the box with the license agreement where you just clicked 'I Agree' on? I'll bet the culprit is in there.
|
|
|
|
|
Precisely.
There are only 10 types of people in the world, those who understand binary and those who don't.
|
|
|
|
|
"Agreements" do not permit illegal behaviour -- e.g. you could add "You agree to murder my boss/ex-wife/mother-in-law if I ask you to", but that doesn't make murder legal. Not only that, but it could be argued that adding illegal behaviour to a contract makes the entire contract invalid/unenforceable.
I wanna be a eunuchs developer! Pass me a bread knife!
|
|
|
|
|
Thank you captain obvious.
Where does it say it is illegal for Microsoft to auto-update their own software?
There are only 10 types of people in the world, those who understand binary and those who don't.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
