Coding Horror

Sparks · 21-01-2011 12:34pm

Civilian_Target wrote: »

Top tip - you can install KDiff3 as your CC Merge tool and configure it to automatically manage whitespace merges.

Oh, please let that work in here...

Calina · 21-01-2011 2:28pm

AlmightyCushion wrote: »

Had that in a job I worked at before. It was an old program written in vb6 and there was barely a comment in sight. It was used by a boat load of people and it had loads of bugs which meant every month without fail there was about half a dozen bugs reported for it. Guess who had the joy of sorting most of those bug fixes. I tried commenting as much as I could as I went along but it was like pissing on a house fire. There was talk of replacing it but it kept getting put off for various reasons. I hope for the sake of everyone there that they eventually done it.

Doubt it. Things like that tend not to be gotten rid of until it's far too late.

bad and all as no comments are are badly copied and pasted comments. That's all I will say.

BostonB · 21-01-2011 3:51pm

AlmightyCushion wrote: »

Had that in a job I worked at before. It was an old program written in vb6 and there was barely a comment in sight. It was used by a boat load of people and it had loads of bugs which meant every month without fail there was about half a dozen bugs reported for it. Guess who had the joy of sorting most of those bug fixes. I tried commenting as much as I could as I went along but it was like pissing on a house fire. There was talk of replacing it but it kept getting put off for various reasons. I hope for the sake of everyone there that they eventually done it.

As a side project, I've inherited a VBA Excel/Access monstrosity thats been growing and constantly modified for 15yrs by about 5 different developers. No comments what so ever. 6GB of data, and I'd have to write all the Access VBA and Excel code to move it to SQL. Not that I have time to do that since its not my main job. Its just cracks me up when I have to go near it.

Legacy VB/VBA apps seem to be in every place I've worked. Usually in the same state as the above.

Calina · 21-01-2011 4:33pm

I suspect some of them are user written applications that become business critical and then, as userwriters leave...land elsewhere for TLC.

conor.hogan.2 · 22-02-2011 10:59pm

Teaching sql92 in late 2010. Hey why dont you teach me old english too?

Korvanica · 03-03-2011 3:35pm

AlmightyCushion wrote: »

Had that in a job I worked at before. It was an old program written in vb6 and there was barely a comment in sight. It was used by a boat load of people and it had loads of bugs which meant every month without fail there was about half a dozen bugs reported for it. Guess who had the joy of sorting most of those bug fixes. I tried commenting as much as I could as I went along but it was like pissing on a house fire. There was talk of replacing it but it kept getting put off for various reasons. I hope for the sake of everyone there that they eventually done it.

Its a similiar story here.... initial program was written in asp and then transferred to .NET - badly..

the program just evolved, additions to it weren't planned and were takd on in places.. so so so so much redundancy aswell... ive found the same method in about 15 classes....

krd · 05-03-2011 6:27pm

Korvanica wrote: »

WHY CANT PEOPLE COMMENT CODE!

F**K SAKE...

started a new job a while ago, NONE of the code is commented.... 80k lines of code? Lovely thanks for that...

There are several reasons.

1. By not commenting the code, you become the only person who understands it. You're less likely to get fired.
2. The coder is one of these ad hoc boys who start with a clean sheet and proceed to cover it in miles of ****.
3. The coder deliberately does it to make other coders look bad.
4. It depends on what you're doing, but there are companies who will sell you source code but they remove all comments. They'll even construct the code in ways that makes it impossible to figure out what they've done.

If the comments and the code is bad, it's worse than no comments.

fasty · 05-03-2011 7:06pm

Part of my job is maintaining a few legacy apps that have been dragged from Win16 to Win32 through a spaghetti code bush and the only comments in that code is commented out code. Why would you leave commented out code in something that's kept in source control?

krd · 06-03-2011 12:16am

fasty wrote: »

Part of my job is maintaining a few legacy apps that have been dragged from Win16 to Win32 through a spaghetti code bush and the only comments in that code is commented out code. Why would you leave commented out code in something that's kept in source control?

Programmers get sacked all the time. Pulling your comments is a good way of teaching the sackers a lesson. People obtusificate code for the same reason.

Sparks · 06-03-2011 12:27am

krd wrote: »

By not commenting the code, you become the only person who understands it. You're less likely to get fired.

krd wrote: »

Programmers get sacked all the time. Pulling your comments is a good way of teaching the sackers a lesson. People obtusificate code for the same reason.

Leaving aside the point for a moment that it's an incredibly unprofessional practice to deliberately do shoddy work, there's a small practical point to consider there. Our industry is surprisingly small, with a surprisingly long memory. People who code like that get a reputation for being a monumental pain in the <InsertColloquailDescriptionOfBodypartHere>. And sooner or later, they go into an interview with, or need a recommendation from, someone who's suffered through a week of trying to figure out their uncommented, badly written code while their manager taps a metaphorical foot in every meeting...

AlmightyCushion · 06-03-2011 12:46am

Sparks wrote: »

Leaving aside the point for a moment that it's an incredibly unprofessional practice to deliberately do shoddy work, there's a small practical point to consider there. Our industry is surprisingly small, with a surprisingly long memory. People who code like that get a reputation for being a monumental pain in the <InsertColloquailDescriptionOfBodypartHere>. And sooner or later, they go into an interview with, or need a recommendation from, someone who's suffered through a week of trying to figure out their uncommented, badly written code while their manager taps a metaphorical foot in every meeting...

Plus one of your team could leave and start working for a different company. Imagine applying for a job there after writing horrible code in the previous job. I think best practise is that when leaving a company (whether you left or were sacked) try not to do anything that pisses the company or the people working there. It could bite you in the ass big time.

NeverSayDie · 06-03-2011 2:00am

krd wrote: »

Programmers get sacked all the time. Pulling your comments is a good way of teaching the sackers a lesson.

No, it really, really isn't. See Sparks' post.

krd · 06-03-2011 3:52am

AlmightyCushion wrote: »

Plus one of your team could leave and start working for a different company. Imagine applying for a job there after writing horrible code in the previous job.

People get away with things. If they couldn't get away with, they wouldn't do it. They get away with it. And they'll do things they can get away with.

In my experience people who drip feed information for power get away with it. If someone becomes the gatekeeper on some code, system or whatever they have power.

It should be a given that everyone should clearly comment their code. There is ABSOLUTELY no good reason (there are plenty of nefarious ones) why a short narrative of the code and an explanation of all variables, naming conventions etc shouldn't be put in the comments. The whole team could have an air crash over the Andes in the morning.

How much resources are wasted by "teams" unable to follow or understand each others work?

I think best practise is that when leaving a company (whether you left or were sacked) try not to do anything that pisses the company or the people working there. It could bite you in the ass big time.

If you get sacked from anywhere - for whatever reasons - even if it's just a case they needed to lower the head count and you're just not popular enough - you will get a bad reference. The way the world works, when people f you over, they want to make sure you can't come back at them. The way many people see the world is life is one big game where everything and everyone is bullsh1t.

The Irish IT world is tiny and it is choc a block full of bolloxes. And it's a shadow of what it once was.

conor.hogan.2 · 06-03-2011 4:23pm

I hope I never even have to look at your code never mind work with you.

Comment when needed and dont be a "<InsertColloquailDescriptionOfBodypartHere>", it will come back to screw you over than it did other people.

zynaps · 06-03-2011 5:14pm

conor.hogan.2 wrote: »

I hope I never even have to look at your code never mind work with you.

Comment when needed and dont be a "<InsertColloquailDescriptionOfBodypartHere>", it will come back to screw you over than it did other people.

I don't think he was advocating that people leave out comments, rather pointing out how unfortunately prevalent the practice of purposely leaving out comments as active sabotage/entrenchment is.

conor.hogan.2 · 06-03-2011 7:25pm

If they were not, my bad then.

krd · 06-03-2011 8:17pm

conor.hogan.2 wrote: »

I hope I never even have to look at your code never mind work with you.

Comment when needed and dont be a "<InsertColloquailDescriptionOfBodypartHere>", it will come back to screw you over than it did other people.

Conor, see zynaps comment.

I don't work in software anymore. I could tell you a lot of stories. In my experience and in the experience of other people I've known, there is a lot of sabotage. There is total and utter bollox behavior that goes across the entire Irish IT industry. And a lot of the bolloxes are good at climbing the ladder.

"Never ascribe to malice, that which can be explained by incompetence." Napoleon Bonaparte

Napoleon was naieve. He never worked in the Irish IT industry.

If someone strips comment from code they're doing it deliberately. If they're using a naming convention that's very confusing, it's deliberate. If every time you ask them a straightforward question, they give you a misleading one word answer, once or twice is failure in communication, every time is very deliberate.

AwayWithFaries · 08-03-2011 1:42pm

krd wrote: »

If the comments and the code is bad, it's worse than no comments.

I hate this with a passion. It so much better to just to leave out comments than put in some of the nonsensical crap I've seen at times.
Or this kinda stuff drives me mad aswell:

Long complicated code with no comments.

var c = a+b;//c = a+b

krd · 09-03-2011 12:50pm

AwayWithFaries wrote: »
I hate this with a passion. It so much better to just to leave out comments than put in some of the nonsensical crap I've seen at times.
Or this kinda stuff drives me mad aswell:
Long complicated code with no comments.

var c = a+b;//c = a+b

Well, it's why people are meant to write their code clearly and meaningful. Well written stuff is surprisingly easy to read.

This is required reading

http://codecourse.sourceforge.net/materials/How-To-Write-Unmaintainable-Code.html

50.In C++, overload +,-,*,/ to do things totally unrelated to addition, subtraction etc. After all, if the Stroustroup can use the shift operator to do I/O, why should you not be equally creative? If you overload +, make sure you do it in a way that i = i + 5; has a totally different meaning from i += 5;

And I knew people who did this as a rule.
49.In C++, overload library functions by using #define. That way it looks like you are using a familiar library function where in actuality you are using something totally different.

That can be incredibly confusing if you do it multiple times. You can effectively create a new programming language. Or a programming language where the rules keep changing.

zynaps · 09-03-2011 1:07pm

krd wrote: »

And I knew people who did this as a rule.
49.In C++, overload library functions by using #define. That way it looks like you are using a familiar library function where in actuality you are using something totally different.

That can be incredibly confusing if you do it multiple times. You can effectively create a new programming language. Or a programming language where the rules keep changing.

Ouch

A testament to the importance and value of namespaces when you can use them (and not making every function globally visible).

conor.hogan.2 · 09-03-2011 7:00pm

That link is evil

and also it runs on the premise that there is no code overviews or even anyone with half a brain looking at your code even on a weekly basis to see it is maintainable.

BostonB · 09-03-2011 10:21pm

I've seen obscurification a couple of times. Pages of code that goes all over the place fore no reason. First time was a student, that was more like just badly written. The second time, was a programmer who did it to the source code of the company he was working for, just before he left to set up in competition. They only realised after he left there was deliberate errors in the code, and in their product.

Skrynesaver · 10-03-2011 4:33pm

I saw a beautiful example of 49 in the International Obfuscated C contest many years ago, someone had used #define to effectively turn cc into a COBOL compiler. Truly shock and awe inducing.

G-Money · 14-04-2011 8:27am

This isn't a coding horror issue, more just a rant.

I'd been working on a script change and it had been through several rounds of code review where people gave me feedback on various changes to be made. All changes were made and I got the thing tested finally (our environment is very restrictive/inflexible and testing can take a long time).

Sent the thing off for the next stage of the process which is to be approved and someone pipes up with several changes they want made. Even though they were on the code review list from the start.

Some of the suggested changes are possibly valid but quite a few I think are personal preferences. Let's just say I'm not best pleased.

I find it massively frustrating that this person had numerous chances to give feedback before this and they said nothing. Then when they do, it's after I've completed testing (which was frustrating to do itself) and which may set me back a week or two due to the way things run.

Lets just say they're going to need jedi mind powers to convince me to make their changes, especially at this stage.

BostonB · 14-04-2011 10:38am

I had that problem on contract, but the person was trying to derail the project and discredit me. I assume though it would replace them, which is couldn't, just save them a bunch of time. Lucky enough the people in charge could see this and basically got the problem person to sign off the project each week at a regular meeting to keep it on track.

A paper trail and a few stats, kept them from causing trouble anymore.

Seth Brundle · 14-04-2011 11:32am

G-Money wrote: »

This isn't a coding horror issue, more just a rant.

I'd been working on a script change and it had been through several rounds of code review where people gave me feedback on various changes to be made. All changes were made and I got the thing tested finally (our environment is very restrictive/inflexible and testing can take a long time).

Sent the thing off for the next stage of the process which is to be approved and someone pipes up with several changes they want made. Even though they were on the code review list from the start.

Some of the suggested changes are possibly valid but quite a few I think are personal preferences. Let's just say I'm not best pleased.

I find it massively frustrating that this person had numerous chances to give feedback before this and they said nothing. Then when they do, it's after I've completed testing (which was frustrating to do itself) and which may set me back a week or two due to the way things run.

Lets just say they're going to need jedi mind powers to convince me to make their changes, especially at this stage.

Four words: Change Control Request Form
Let the buggers know the amount of work that it would take now along with the cost of doing them!

oneweb · 04-05-2011 4:25pm

Someone forgot to comment on where the author tag is, sheer laziness...
(!)

<head>
<link href="/favicon.ico" type="image/x-icon" />
<!--This is the Document CHARSET -->

<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-15" />



<!-- This is the Title Element -->

<title>StatCentral.ie - The Portal to Ireland's Official Statistics</title>



<!-- This is the META Description Tag -->

<meta name="description" content="The home page of StatCentral.ie" />



<!-- This is the META Keywords Tag -->

<meta name="keywords" content="Home, Statcentral.ie, CSO, Central Statistics office, Statistics Ireland" />



<!-- This is the beginning of the DCMI Element Set -->

<link rel="schema.DC" href="http://www.purl.org/dc/elements/1.1/" />


<!-- Element Name: Title -->

<meta name="DC.title" content="StatCentral.ie - The Portal to Ireland's Official Statistics" />



<!-- Element Name: Creator -->

<meta name="DC.creator" content="Central Statistics Office; Statcentral" />



<!-- Element Name: Subject -->

<meta name="DC.subject" content="Home, Statcentral.ie, CSO, Central Statistics office, Statistics Ireland" />


<!-- Element Name: Description -->

<meta name="DC.description" content="The home page of StatCentral.ie" />



<!-- Element Name: Publisher-->

<meta name="DC.publisher" content="Central Statistics Office; Statcentral" />



<!-- Element Name: Contributor -->

<meta name="DC.contributor" content="Central Statistics Office; Statcentral" />



<!-- Element Name: Date -->

<meta name="DC.date" scheme="W3CDTF" content="2008-01-31" />



<!-- Element Name: Type -->

<meta name="DC.type" scheme="DCMIType" content="Text" />



<!-- Element Name: Format -->

<meta name="DC.format" scheme="IMT" content="text/html" />



<!-- Element Name: Identifier -->

<meta name="DC.identifier" content="/meta-tags/dublin/" />



<!-- Element Name: Source -->

<meta name="DC.source" content="/meta-tags/" />



<!-- Element Name: Language -->

<meta name="DC.language" scheme="RFC1766" content="en" />



<!-- Element Name: Relation -->

<meta name="DC.relation" content="/meta-tags/" />



<!-- Element Name: Coverage -->

<meta name="DC.coverage" content="World" />



<!-- Element Name: Rights -->

<meta name="DC.rights" content="Copyright Government of Ireland" />



<meta name="author" content="Central Statistics Office; Statcentral" />

zynaps · 05-05-2011 3:46pm

oneweb wrote: »

Someone forgot to comment on where the author tag is, sheer laziness...
(!)

 <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-15" />  <title>StatCentral.ie - The Portal to Ireland's Official Statistics</title>  <meta name="description" content="The home page of StatCentral.ie" />  <meta name="keywords" content="Home, Statcentral.ie, CSO, Central Statistics office, Statistics Ireland" />
[...etc...]

Man, I'd never have expected mind-numbing comment bureaucracy from them!

Q_Ball · 05-05-2011 3:58pm

conor.hogan.2 wrote: »

That link is evil and also it runs on the premise that there is no code overviews or even anyone with half a brain looking at your code even on a weekly basis to see it is maintainable.

This is the case where I work. Peer reviews are supposed to be done, but they're never factored into estimates. Logic behind this is that we've never done them before so why add additional cost and time to the estimate. After all, we've "built" similar for the client before, they are expecting a cost that is less or in line with.

We propose improvements to the underlying codebase, "under the hood" so to speak, but why would the client pay for that when they are not getting any new functionality.

I won't tell you what we were actually told to do. I died a little on the inside.

Funny thing is, if the client actually had a technical person who could evaluate our code, I'm sure we'd have been fired long ago. The problem is they don't. So they see the figures coming in, thinking we're great and management are giving themselves a pat on the back for a "job well done".

Meanwhile, the application (d)evolves quite rapidly into an unmaintainable mess. It really is enough to break a person.

New job time methinks.

km991148 · 06-05-2011 11:18am

Q_Ball wrote: »

This is the case where I work. Peer reviews are supposed to be done, but they're never factored into estimates. Logic behind this is that we've never done them before so why add additional cost and time to the estimate. After all, we've "built" similar for the client before, they are expecting a cost that is less or in line with.

We propose improvements to the underlying codebase, "under the hood" so to speak, but why would the client pay for that when they are not getting any new functionality.

I won't tell you what we were actually told to do. I died a little on the inside.

Funny thing is, if the client actually had a technical person who could evaluate our code, I'm sure we'd have been fired long ago. The problem is they don't. So they see the figures coming in, thinking we're great and management are giving themselves a pat on the back for a "job well done".

Meanwhile, the application (d)evolves quite rapidly into an unmaintainable mess. It really is enough to break a person.

New job time methinks.

I (Briefly) worked in a place like this. Soul destroying - and as a result had very high turnover. Good developers left quickly, not so good ones stayed a while longer and polluted code further making it increasingly harder for new comers to figure out what half the products were doing/ supposed to be doing. Management couldn't understand the problem as products were still being shipped and were not interested in investing time in improving quality even although it would altos certainly decrease overall development time (It even took ridiculous amounts of time to find code, never mind work out what the fu(k it was doing due to half ars3d practices of previous 'developers').

Genghiz Cohen · 13-05-2011 11:02am

"Hey, this didn't save my input."
"Oh yeah?"
"Yeah, I put in the names then clicked 'Save For Later' and when I came back it didn't load them. Any ideas?"
"Oh, that 'Save For Later' button doesn't actually do anything."

BostonB · 13-05-2011 11:24am

How about the secret button.

"Hey these calculations don't work"

"Hit ALT Z T H"

"What the .... does that do"

"Fixes the numbers"

Some coder decided to code hidden buttons to do things, so he could fix things in his apps, and others wouldn't be able to.

BostonB · 13-05-2011 11:29am

Another one I don't get is coding error handlers, but then not putting any plain text explanation so users don't understand the errors, or coding to UI to to provide feedback to the user . IE "Truncated String Error" you've entered too much text. Then repeat that for every type of user input that requires validation. Or not checking a record/data already exists on input, so users enter duplicated data constantly, requiring records to be merged and fixed.

Just ARRRTGGGGGHHH. :mad:

fasty · 14-05-2011 4:06pm

A friend of mine recently did some development on a game I've been working on...

Seeing people use boost::function and boost::bind when a C style callback will do makes me cry.

Boost is the business and the above has it's place in an event driven Windows app, but not sprinkled liberally in code where performance matters inside loops processing hundreds of elements in some other data structure.

Plus he really does not give a crap about thread safety which is why things are crashing on him and thanks to his liberal use of the above event system, I'd have to slap mutexes all over the place to make it thread safe.

Thank **** for source control...

Seachmall · 19-05-2011 7:20pm

Figured this was suited to here (if not just delete it),

[Too wide to embed, NSFW language]

"That's The F'ing Business"

Thomas_S_Hunterson · 01-06-2011 3:36pm

Spotted on StackOverflow: link

Closely related to thread title.

srsly78 · 01-06-2011 5:40pm

Thomas_S_Hunterson wrote: »

Spotted on StackOverflow: link

Closely related to thread title.

Ok I just spat my coffee all over monitor

Flojo · 02-06-2011 5:57am

How come on this forum the BBcode is turned off in sigs? Surely it would prevent a lot of "sig too big" crap if allow alignment BB code NoYes?

AlmightyCushion · 02-06-2011 6:06am

Flojo wrote: »

How come on this forum the BBcode is turned off in sigs? Surely it would prevent a lot of "sig too big" crap if allow alignment BB code NoYes?

I don't know about anyone else but I find right or centre alligned sigs very distracting.

Flojo · 02-06-2011 6:13am

AlmightyCushion wrote: »

I don't know about anyone else but I find right or centre alligned sigs very distracting.

Really hehe, how so? Surely it's better than looking at a car crash sig!

fatlog · 09-06-2011 4:46pm

Thomas_S_Hunterson wrote: »

Spotted on StackOverflow: link

Closely related to thread title.

love the comment...

#include "fritzl.h".

Korvanica · 09-06-2011 5:00pm

if (newConfirm("<%=CO_Corrections_p%>", checkMessage, 2, 1, 0)) 
{
     document.CorDetails.ApplyToAllDays.value = 1;
}

document.CorDetails.ApplyToAllDays.value = 1;

ffs....

spider guardian · 17-06-2011 1:02pm

https://github.com/MrMEEE/bumblebee/commit/a047be85247755cdbe0acce6

687474703a2f2f692e696d6775722e636f6d2f7937486d392e6a7067

zynaps · 17-06-2011 3:25pm

spider guardian wrote: »

https://github.com/MrMEEE/bumblebee/commit/a047be85247755cdbe0acce6

Oh my christ.

@@ -348,7 +348,7 @@ case "$DISTRO" in
   ln -s /usr/lib/mesa/ld.so.conf /etc/alternatives/gl_conf
   rm -rf /etc/alternatives/xorg_extra_modules
   rm -rf /etc/alternatives/xorg_extra_modules-bumblebee
-  rm -rf /usr /lib/nvidia-current/xorg/xorg

:eek: :mad:

Pygmalion · 18-06-2011 9:17pm

Been looking over some of my old code, found this gem (was scrolling through the file and the 7 levels of indentation immediately piqued my interest).

Not something I'm particularly proud of

.

In fairness it was something I wrote as part of a competition and I was in "get something compiled before time runs out mode" (with a tiny bit of "oh god my brain stopped working an hour ago mode").

spider guardian · 18-06-2011 10:26pm

Pygmalion wrote: »

Been looking over some of my old code, found this gem (was scrolling through the file and the 7 levels of indentation immediately piqued my interest).

Not something I'm particularly proud of .

In fairness it was something I wrote as part of a competition and I was in "get something compiled before time runs out mode" (with a tiny bit of "oh god my brain stopped working an hour ago mode").

Good lord man. Have you entered that into the obfuscated java contest?

Odaise Gaelach · 21-06-2011 12:00pm

This is probably a stupid question, but... what does it do?

conor.hogan.2 · 21-06-2011 1:59pm

Pygmalion wrote: »

...7 levels of indentation...

I used to do that…..

Well, the gist of it, I carriage return the opening braces.

Not coding horror, but I love random/funny comments sometimes. obviously they can be abused as comments are supposed to be helpful.

I saw "hacked from some f'in place" in some code the other day, needless to say it was bad code….

fasty · 21-06-2011 2:24pm

But what's the fix for such deep indentation?

zynaps · 21-06-2011 3:22pm

fasty wrote: »

But what's the fix for such deep indentation?

Maybe to extract a method or two, or maybe some of the logic can be "flattened" somehow to mitigate the real problem (apart from a lack of comments); the nested loops 4-levels deep.
It might be faster to walk the list of frames and then do whatever "bubbling up" process happens in the deepest loop, rather than testing frame.ontile(x,y) for every position on the grid.

Still, it's a polynomial runtime at worst, which isn't a big deal

Coding Horror

Comments