Something is Pareto optimal or Pareto efficient if one variable cannot be increased without decreasing another variable.
To be Pareto optimal is to be on an "efficient frontier."
Figure A is a production possibilities curve (PPC). Figure A measures the production of goods vs services, but the production possibilities frontier (PPF) can include every and any number of goods. The job of a PPC is to depict the possible combinations of goods that can be produced with the resources available.
Point D falls inside the PPF (below the curve). It is possible to produce that combination of goods and services, but it is not Pareto optimal. The economy would not be working at its full capacity.
Points outside the PPF (above the curve, not shown) represent combinations that are not possible with the resources available.
Points A, B, and C fall on the curve, representing possible combinations of goods and services that can be produced with the resources available. Points on the curve represent trade-offs, you cannot increase one output without decreasing another output.
The reason that the curve is curved and not strait is that not all resources are equally useful in the production of all goods. Some resources are specialized (land is good for only certain crops, people are trained for only certain jobs, machines are made for only certain functions, etc.), so decreasing the production of one good will not equivalently increase production in other areas.
When the available resources increase (due to technological improvements, for example), the curve will shift. This is not to be confused with movements along the curve; the former changes the curve while the ladder changes the point in a given PPF.
As well as economics, Pareto optimality is a useful concept with respect to evolutionary fitness, because living things have scarce resources and will thus need to optimize their morphology across possibilities.