The goal of the kata is to separate the generic solution to an interview question and make brain space for the actual business logic.
Kata:
<ul>
<li>write the template for a sliding window solution
<ul>
<li>pay attention to the inclusive and exclusive values, if you don't <code>+1</code> will eat your lunch
</li>
<li>loop is 0 indexed
</li>
<li>add an inner loop which looping depends on the values created by the outer loop
</li>
<li>start from different values than <code>0</code>
</li>
<li>increase the step higher than <code>1</code>
</li>
</ul>
</li>
<li>explain the O() notation of the solution
</li>
<li>Testing:
<ul>
<li>be sure the sliding happens according to the expectations
</li>
<li>use boundary value analysis to test it
</li>
</ul>
</li>
</ul>
Considerations:
<ul>
<li>what data can be simplified?
</li>
<li>what data can be reworked to get further performance?
<ul>
<li>For example, a list looked up frequently or in every step can be reworked to HashMap for faster access.</li>
</ul>
</li>
<li>Are there partial calculations that can be re-used?
</li>
<li>Is there any opportunity to cache something?
</li>
<li>If it is possible don't do any I/O operation in the loops, if you have to emphasize its problematic nature.
</li>
</ul>

The goal of the kata is to separate the generic solution to an interview question and make brain space for the actual business logic.

**Kata**:

* write the template for a sliding window solution
    
    * pay attention to the inclusive and exclusive values, if you don't `+1` will eat your lunch
        
    * loop is 0 indexed
        
    * add an inner loop which looping depends on the values created by the outer loop
        
    * start from different values than `0`
        
    * increase the step higher than `1`
        
* explain the **O()** notation of the solution
    
* **Testing**:
    
    * be sure the sliding happens according to the expectations
        
    * use boundary value analysis to test it
        

**Considerations**:

* what data can be simplified?
    
* what data can be reworked to get further performance?
    
    * For example, a list looked up frequently or in every step can be reworked to HashMap for faster access.
        
* Are there partial calculations that can be re-used?
    
* Is there any opportunity to cache something?
    
* If it is possible don't do any I/O operation in the loops, if you have to emphasize its problematic nature.

Andras Csanyi

Andras Csanyi

Sliding window