#goodcode

Poor coding is a very toxic habit that can be detrimental to even the most established and successful programmers. The most important thing to understand is that coding is a science. You need to follow the rules that are established in the code to make sure that you are creating a successful outcome. You need to follow the correct order, do it properly, and use the correct syntax. There are only so many ways to do it. If you follow a good code of practice that is well established, you are likely to produce successful results. If you don’t, you are doomed to failure.

There are all sorts of international coding standards that are accepted around the world. Every country has its little quirks, but the overwhelming majority of programmers

Poor coding is a very toxic habit that can be detrimental to even the most established and successful programmers. The most important thing to understand is that coding is a science. You need to follow the rules that are established in the code to make sure that you are creating a successful outcome. You need to follow the correct order, do it properly, and use the correct syntax. There are only so many ways to do it. If you follow a good code of practice that is well established, you are likely to produce successful results. If you don’t, you are doomed to failure. There are all sorts of international coding standards that are accepted around the world. Every country has its little quirks.

More on good and bad code in our new article https://duecode.io/blog/good-code-vs-bad-code/

Here is a great description of different forms of code coverage. Thanks to Florian for the recap!!

Statement Coverage:

how many portions of the CFG are executed (nodes & edges)

#executed / #total

-> still possible to miss bugs

Branch Coverage:

test all possible branches in control flow (edges)

complete branch coverage implies complete statement coverage

-> still possible to miss bugs

Path Coverage:

test all possible paths (sequence of branches)

complete path coverage implies complete branch coverage

-> not feasible with loops (arbitrary # of paths)

Loop Coverage:

for each loop, test 0, 1, and 1+ iterations

coverage = #loops with 0,1,1+ iterations / #loops * 3s

Data Flow Coverage:

evaluated with DU pairs

coverage = #DU-pairs / used DU-pairs

 Glossary

CFG = control-flow graph

nodes with statements immediately and always executed after each other

nodes are connected by edges if reachable

for example, an if condition statement produces three nodes A (statements before if), B (statements if condition true),C (statements if condition false) and edges between A->B and A->C

DU: definition use pair

defines where variable used at the moment was defined at; each variable has typically many definition use pairs

for example, (2,3) means the variable a used at line 3 was assigned to (hence defined) at line 2

The chart above is the histogram of average values of measured absolute coverage across Google. The median (50th percentile) code coverage is 78%, the 75th percentile 85% and 90th percentile 90%. We believe that these numbers represent a very healthy codebase.

Great quote