Could we make a thread specifically about bad practices?

[darth_bubbles]

I am learning programming and everywhere I look there's people saying "This is bad practice" and "This is not bad practice"

 

So in order to get rid of some confusion, let's post all the bad practices we can think of and why.

 

Thanks!

[Sklo]

Never use Go to's  for looping, loop exiting or instead of functions or anything.. that IS bad practice.

[xnoobftw]

Masturbating infront of your parents is a bad practice

[fizzlesticks]

Go to's never use them, they work against the logic of a structural programming language.

Goto are fine in certain circumstances, such as error handling.

[79wjd]

Using

break;

is a bad practice....or so every professor i've had likes to say--although It certainly has its merits. 

[LukeTim]

Go to's never use them, they work against the logic of a structural programming language.

 

Go to has its uses in handling errors when you need to do clean up. Don't dismiss it entirely.

Just don't use it for looping, loop exiting or instead of functions or anything.. that IS bad practice.

[Sklo]

Go to has its uses in handling errors when you need to do clean up. Don't dismiss it entirely.

Just don't use it for looping, loop exiting or instead of functions or anything.. that IS bad practice.

okey thats what I meant

[deviant88]

My fails:

Not knowing how to learn to code.

Not knowing how to write reusable code(waste of time later).

[LePawel]

Never use Go to's  for looping, loop exiting or instead of functions or anything.. that IS bad practice.

There are good uses for goto, they are simply rare.

Looping: use continue; instead (skips rest of loop, repeating loop cycle)

Loop exiting: break; unless you have nested loops. Here you can either use flags OR goto (last answer shows how to use flags)

Functions: just make a function

 

 

Using

break;

is a bad practice....or so every professor i've had likes to say--although It certainly has its merits. 

??? Interating through an array to select an object has a perfectly valid use for break; once you find it, provided it's not separated into a function.

for(int i=0;i<array.size();i++)  if(array[i].name == nameIAmLookingFor){    arrayObject = &array[i];    break; //saves unnecessary loops, much cleaner than a goto  }//work with arrayObject here

[DigitalHermit]

while(1){malloc(32767);}

is a bad practice...

[Nuluvius]

Have a look at the Bad, terrible or simply funny code examples thread.

[metal_remco]

Trusting user input

[darth_bubbles]

Using

break;

is a bad practice....or so every professor i've had likes to say--although It certainly has its merits. 

but it's good in switch statements, right?

[kingdorian]

Trusting your OOP professor who saying its good practice to use comments every one or 2 lines, using a lot of white lines, and for example when using an if statement:

if(variable<something){[some code]}

Putting the { on a next line.

Because all of the above things improve readability in his opinion....

Of course you have to use comments but come on... every 1 or 2 lines, bs. Using a lot of white lines and putting every curly bracket on a new line... Just NOPE

(btw this is mainly focused on Java)

[Bene]

In c++, using a struct of 100+ elements and totally about 800kB

[fizzlesticks]

In c++, using a struct of 100+ elements and totally about 800kB

What's wrong with that?

[Bene]

What's wrong with that?

First, it's c++, not c, it's better to use class,

the names must be very hard to remember. it's better to nest struct  grouping some elements.

then one have to using pointer to void pass large size of data to functions, and allocate in stack.

[Midnight]

First, it's c++, not c, it's better to use class,

the names must be very hard to remember. it's better to nest struct  grouping some elements.

then one have to using pointer to void pass large size of data to functions, and allocate in stack.

 

Why is it better to use class? Functionally, struct and class are the same.

[PlutoNZL]

 

There are good uses for goto, they are simply rare.

Looping: use continue; instead (skips rest of loop, repeating loop cycle)

Loop exiting: break; unless you have nested loops. Here you can either use flags OR goto (last answer shows how to use flags)

Functions: just make a function

 

 

 

??? Interating through an array to select an object has a perfectly valid use for break; once you find it, provided it's not separated into a function.

for(int i=0;i<array.size();i++)  if(array[i].name == nameIAmLookingFor){    arrayObject = &array[i];    break; //saves unnecessary loops, much cleaner than a goto  }//work with arrayObject here

Your loop would work, but it is very bad practice. Both of these solutions would be  better:

 

for (int i = 0; i < array.size() && arrayObject == null; i++) {   if (array[i].name == nameIAmLookingFor) {      arrayObject = &array[i];   }}		int i = 0;for (; i < array.size() && array[i].name != nameIAmLookFor; i++);arrayObject = &array[i];

[LePawel]

 

Your loop would work, but it is very bad practice. Both of these solutions would be  better:

for (int i = 0; i < array.size() && arrayObject == null; i++) {   if (array[i].name == nameIAmLookingFor) {      arrayObject = &array[i];   }}		int i = 0;for (; i < array.size() && array[i].name != nameIAmLookFor; i++);arrayObject = &array[i];

Let's agree to disagree,

This loop is literally 5 lines long there's nothing a coder can miss out on. I would follow your style if I was to work with a large, screen-height+ loop with multiple exit points, but here? No reason, the code looks clean regardless.

I have heard that it's apparently slower to use break than check for another flag, but I wonder how does that stack up in scenario of even just 100 objects scanned (100 checks) vs 1 break.

 

Lastly, to make people think I'm an absolutely terrible coder, you could use well-named goto to break out of a large loop with multiple exits or a nested loop! *makes scary ghost sounds*

[LukeTim]

First, it's c++, not c, it's better to use class,

the names must be very hard to remember. it's better to nest struct  grouping some elements.

then one have to using pointer to void pass large size of data to functions, and allocate in stack.

 

A C++ struct is not the same as a C struct. A C++ struct is basically the same as a class... only struct objects are copied instead of referenced when reassigned.

[Ciccioo]

A C++ struct is not the same as a C struct. A C++ struct is basically the same as a class... only struct objects are copied instead of referenced when reassigned.

What do you mean by that? What is referenced when reassinged? Everything is copied by default in both c and c++ afaik, via field-to-field copy or copy constructor (just like classes)

The only difference i knew is that structs have public members by default, while classes default to private

[LukeTim]

What do you mean by that? What is referenced when reassinged? Everything is copied by default in both c and c++ afaik, via field-to-field copy or copy constructor (just like classes)

The only difference i knew is that structs have public members by default, while classes default to private

 

Perhaps I am getting confused with C# then, sorry.

In C#, IIRC, a struct will be copied when assigned to another variable, where as with a class a reference will be assigned.

 

Therefore 

 

Struct a;Class b;Struct c = a; // c is a copy of a, it is not the same objectClass d = b; // d and b are references to the same object

[fizzlesticks]

 

Perhaps I am getting confused with C# then, sorry.

In C#, IIRC, a struct will be copied when assigned to another variable, where as with a class a reference will be assigned.

 

Therefore 

 

Struct a;Class b;Struct c = a; // c is a copy of a, it is not the same objectClass d = b; // d and b are references to the same object

Right, that's a C# and Java thing because all variables of class types (as opposed to ints and floats) are just pointers.

In C++ both cases will just call your operator= function.

 

edit: Those will call the copy constructor. If they were already declared and being reassigned it will call operator=.

[mikebald]

Some ASP.Net / General Web Programming things not to do:

 

• Do not use control adapters
• Do not set a control’s style properties in the code-behind, use CSS
• Do not use a Browser Capability list, instead use client-side feature detection such as Modernizr
• Do not depend on request validation to protect against XSS attacks, use input validation / sanitation
• Do not use <%= %>, use <%: %>
• Do not use Cookieless Forms Authentication
• Do not Set EnableViewStateMac = false.
• Do not use Medium Trust or any other level as a security boundary [http://msdn.microsoft.com/en-us/library/tkscy493%28v=VS.71%29.aspx]
• Do not use to appSettings disable security fixes [http://msdn.microsoft.com/en-us/library/hh975440.aspx]
• Do not use UrlPathEncode, use UrlEncode
• Avoid using PreSendRequestHeaders & PreSendRequestContent, they can cause issues with async requests
• Avoid writing async void methods for Page lifecycle events like Page_Load. Use Page.RegisterAsyncTask()
• Avoid using “fire and forget work” such as timers and ThreadPool.QUWI. Use a Window Service or Worker Role.
• Avoid reading Request.Form or Request.InputStream before the HandlerExecute event. If you need to, use Request.GetBufferlessInputStream() or GetBufferedInputStream().
• Try to avoid using EnableViewState, instead use ViewStateMode = “Disabled”.

Tips taken from Scott Hanselman's talk at NDC 2013: http://vimeo.com/68390507