Using orthogonal matrix to design test spec.

  Before yesterday,I wrote an article named the balance of the number of specs and the coverage of testing.In that article,we talk about a testing method--orthogonal matrix.This time I'll talk about the detail for it.

  As I said in the last article,we always in front of inputing multiple values into different textboxes.For instance,there have two value to test five textboxes,if execute total testing,the collection is same with the knowledge of scatter math.There have 2 power 5 different results.Suppose,there have more value and more textboxes.What is the result?It will increaseing by exponential.That will a really big number.
  So let us see the orthogonal method.It choose some representative gather to test base on orthogonal from total testing.Firstly,it has several basic concptions of it.Factor,it's the pre-test textbox.Level,the value need to input into the factor.Runs,the number of line,namely,testing times,the number of collections.Factors,the number of columns.Levels,the biggest number of any single factor,namely,how many values the textbox can be inputted.The format of it is L(Runs)[Levels power Factors].Secondly,the step of using orthogonal to design test spec.Study how many factor;how many level for each factor;choose a fit orthogonal table;mapping the value of factor into table;every line instead of a spec,which a collection of factor;at last,you can add some spec which you think need test.
  Now,in above these question,I believe the key step is choosing a orthogonal table.There are three points need to focus on.A table is consists of factor,levle,and runs.On second thoughts,there are three instances.The number of factor and the number of levle for each factor are same;The number of factor isn't same;the number of level isn't same.For the first point,have same number of factor and level,this is fit for an orthogonal table.The seconde point,level is same in table,but the factor has different value,in this case,the number of factor should bigger than the actual value.At last,the number of level of each factor isn't same.
  As I haven't understand the whole principle of it,I must stop here today.Above these are some primary knowledge of it,just for your infomation.I'll go on studying it,so you'll see some new in recently.Thanks your attention!