Memoization

Take a look at the Wikipedia entry on Memoization for the nitty gritty.

Esentially, it's a term for a specific form of caching, or so it would appear. Take the following data-set:

List Families [FamilyId : Name] {1 : Smith, 2 : Jones}
List People [PersonId : Name] {1: Adam, 2 : Ben, 3 : Carol, 4 : David}
List PeopleInFamilies [FamilyId : PersonId] { 1 : 1, 1 : 2, 1 : 3, 2 : 4]

If you wanted to work out how many people were in each family, you could:

foreach(familyItem in Familes)
{
  counter = numberOfPeopleInFamily(familyItem.FamilyId)
  print("FamilyId {0} Count {1}", familyItem.FamilyId, counter);
}

function numberofPeopleInFamily(familyId)
{
  counterInFunction = 0;
  foreach(peopleInFamiliesiitem in PeopleInFamiles)
  {
    if familyId = peopleInFamiliesiitem.FamilyId
    {
      counterInFunction++;
    }
  }
  return counterInFunction;
}

However, for a set of 1000 families and 800 people in families, this would mean that there would be 800,000 iterations over the "PeopleInFamilies" dataset. The code could be re-written more sensibly as:

dictionary<int, int> familyIdByCountInFamily = numberofPeopleInFamilies;
foreach(familyItem in Familes)
{
  print("FamilyId {0} Count {1}", familyItem.FamilyId, familyIdByCountInFamily[familyItem.FamilyId]);
}

function numberofPeopleInFamilies()
{
  dictionary<int, int> familyIdByCountInFamily
  foreach(peopleInFamiliesiitem in PeopleInFamiles)
  {
    familyIdByCountInFamily[peopleInFamiliesiitem]++;
  }
  return familyIdByCountInFamily;
}

Now there are only 800 iterations over the "peopleInFamilies" dataset, much better! That's my take on memoization, anyway.