| View previous topic :: View next topic |
| Author |
Message |
ressas
Joined: 15 Nov 2019
Posts: 135
|
| Is "else" always necessary? |
 Posted: Fri Oct 09, 2020 6:34 am |
 |
|
Hi.
When I use if in the program, do I have to use else after it?
While checking any button or condition checking
| Code: |
if(input(pin_b0))
output_high(pin_D0);
//else ????
if(hsEnable>3)
hsOut = 0;
//else ???
|
|
|
 |
temtronic
Joined: 01 Jul 2010
Posts: 9226
Location: Greensville,Ontario
|
|
| Posted: Fri Oct 09, 2020 6:42 am |
|
|
Usually yes...
Consider your code example:
| Quote: |
if(input(pin_b0))
output_high(pin_D0);
|
The first time this test is made, and if true, the pin goes high
it will stay high forever unless your code 'elsewhere' resets the pin.
If you want the output pin to 'follow' or 'mimic' the input pin, you'll need the 'else' half of the IF statement.
Others can probably explain it better...
Jay |
|
|
ressas
Joined: 15 Nov 2019
Posts: 135
|
|
| Posted: Fri Oct 09, 2020 10:29 am |
|
|
| temtronic wrote: | usually yes...
consider your code example
...
if(input(pin_b0))
output_high(pin_D0);
...
the first time this test is made, and if true, the pin goes high
it will stay high forever unless your code 'elsewhere' resets the pin.
if you want the output pin to 'follow' or 'mimic' the input pin, you'll need the 'else' half of the IF statement.
Others can probably explain it better...
Jay |
Thanks temtronic.
What you said in the button example makes sense. However, is it also necessary for the variable? When the HSENABLE variable in the example is less than 3, I don't want to do anything. Then? |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9226
Location: Greensville,Ontario
|
|
| Posted: Fri Oct 09, 2020 10:41 am |
|
|
re:
| Quote: |
if(hsEnable>3)
hsOut = 0;
|
In this code,
the FIRST time hsEnable = 4, 5,6,....
then hsOut will be set ot 0.
AND stay there, forever.
It will not be changed to say 255 or 3 or 1, whenever hsEnable <4.
Jay |
|
|
dluu13
Joined: 28 Sep 2018
Posts: 395
Location: Toronto, ON
|
|
| Posted: Fri Oct 09, 2020 2:39 pm |
|
|
It's not strictly necessary. For example, if you want to do some kind of action where your else statement basically says "do nothing", then you can just not put anything after the "if" statement.
However, for my own sake, I usually prefer to do this in order to remind myself that I purposely want no action if my condition is not satisfied:
| Code: | if (stmt)
{
// do stuff
}
else {}
|
|
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9226
Location: Greensville,Ontario
|
|
| Posted: Fri Oct 09, 2020 3:17 pm |
|
|
The problem with not specifying an 'else', action is that once the if (condition) is met and 'something is set or done, it will NOT be reset or undone.
In his first test, when the input pin B0 is high, the D0 pin gets sets high and stays high forever, even when the input pin goes low. |
|
|
dluu13
Joined: 28 Sep 2018
Posts: 395
Location: Toronto, ON
|
|
| Posted: Fri Oct 09, 2020 3:51 pm |
|
|
| temtronic wrote: | the problem with not specifying an 'else', action is that once the if (condition) is met and 'something is set or done, it will NOT be reset or undone.
In his first test, when the input pin B0 is high, the D0 pin gets sets high and stays high forever, even when the input pin goes low. |
Yes in his case it's probably a good idea. But that's why I said it's not strictly necessary. In my example maybe you'd just increment a counter or something. |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19513
|
|
| Posted: Fri Oct 09, 2020 11:56 pm |
|
|
Exactly.
It is perfectly reasonable to have a set of operations that you want to
occur 'if' a set of conditions are met, with no need for an 'else'.
However in probably 90% of situations, there will be 'other things' that
have to happen if the conditions are not met, and in this case an 'else'
is the way to go.
Draw out the situations:
Condition met, needs to do xxxxx.
Condition not met needs to do yyyyy.
If there is nothing in the 'yyyyy' section, then no 'else' is wanted or needed. |
|
|
asmallri
Joined: 12 Aug 2004
Posts: 1634
Location: Perth, Australia
|
|
| Posted: Mon Oct 12, 2020 1:33 am |
|
|
| Ttelmah wrote: | Exactly.
It is perfectly reasonable to have a set of operations that you want to
occur 'if' a set of conditions are met, with no need for an 'else'.
However in probably 90% of situations, there will be 'other things' that
have to happen if the conditions are not met, and in this case an 'else'
is the way to go.
Draw out the situations:
Condition met, needs to do xxxxx.
Condition not met needs to do yyyyy.
If there is nothing in the 'yyyyy' section, then no 'else' is wanted or needed. |
It's a personal programming style thing. In my own case, the 90% is the other way around. Most of the time there are no else statements.
For example:
| Code: |
// msg is a pointer to some prebuilt string.
if (RadioActive)
SendMessage(msg, Radio);
if (ConsoleActive)
SendMessage(msg, Console);
if (LoggingActive)
LogData(msg);
|
There is no "else" execution required in this scenario
_________________
Regards, Andrew
http://www.brushelectronics.com/software
Home of Ethernet, SD card and Encrypted Serial Bootloaders for PICs!! |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19513
|
|
| Posted: Mon Oct 12, 2020 3:38 am |
|
|
I must admit for my customers, there would have to be handling to
ensure that if none of the outputs were 'active', some form of emergency
reporting was done. |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9226
Location: Greensville,Ontario
|
|
| Posted: Mon Oct 12, 2020 4:57 am |
|
|
re:
| Quote: | if (RadioActive)
SendMessage(msg, Radio);
|
Question. Do you reset 'RadioActive' within the called function 'SendMessage(msg, Radio); ' ?
I'm assuming you do, otherwise once 'RadioActive' was set, you'd always call the 'sendmsg' function. I've been caught in that 'trap' a few times...
Now I always reset the 'flag' on the FIRST line of the called function, well mostly....
Jay |
|
|
asmallri
Joined: 12 Aug 2004
Posts: 1634
Location: Perth, Australia
|
|
| Posted: Mon Oct 12, 2020 6:04 am |
|
|
No. These values are set in the initialisation logic during the hardware and configuration phase. Is there are Radio plugged in? If so use it etc
_________________
Regards, Andrew
http://www.brushelectronics.com/software
Home of Ethernet, SD card and Encrypted Serial Bootloaders for PICs!! |
|
|
jeremiah
Joined: 20 Jul 2010
Posts: 1349
|
|
| Posted: Mon Oct 12, 2020 6:32 am |
|
|
| asmallri wrote: |
It's a personal programming style thing. In my own case, the 90% is the other way around. Most of the time there are no else statements.
For example:
| Code: |
// msg is a pointer to some prebuilt string.
if (RadioActive)
SendMessage(msg, Radio);
if (ConsoleActive)
SendMessage(msg, Console);
if (LoggingActive)
LogData(msg);
|
There is no "else" execution required in this scenario |
I am in a similar boat (though not 90/10...more like 70/30). Most of my code ends up being validation, and using if's without else's is a much cleaner method for me as it sets up a type invariant pattern to work with. I'm often times working through a list of invariants at the beginning of a function so that when I get to the bottom of my function and start doing math or whatever logic, I know that the data types I am using are safe to use within that context:
| Code: |
if(error_condition1){
return error_code1;
}
if(error_condition2){
return error_code2;
}
if(error_condition3){
return error_code3;
}
if(error_condition4){
return error_code4;
}
// now do logic knowing values won't overflow or be in an invalid state
|
It's really useful for message validation before acting on a command for example. If/else chaining quickly becomes unreadable and a huge maintenance hazard due to how many levels deep it starts to get. Some of it is also personal preference for me as well. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
