I'm not in agreement with Denny on the starter gravity. In talking with Chris White (White Labs) many years ago, he said they culture their yeast using a much lower gravity wort. He indicated about 1.020, but I see that in his new yeast book, he mentions 1.030 as a prefered starter gravity. The issue of the Crabtree Effect is more pronounced when you increase the wort gravity, which diverts some yeast metabolic effort into alcohol production instead of yeast mass growth. I use a starter gravity in the low 1.020s and typically add about 700 ml of this wort twice during the starter propagation (I'm mostly an ale brewer). Most of the spent wort is decanted after the yeast has been dropped by chilling leaving me with a nice yeast slurry for pitching.
I don't think there is much difference between the use of a foam stopper or foil cover since I don't believe that there is enough Brownian Diffusion of oxygen into the starter vessel when there is a net exhalation of CO2 during the yeast growth process. I prefer to pump filtered air into the vessel to assure that the headspace maintains a normal oxygen level. Over the years, I've found that using an airstone is not really necessary since I use a 6L Erlenmeyer flask with a 3-inch stirring magnet and there is a significant wort surface to transfer oxygen.
The other problem I found with using an airstone was that it caused significant foaming problems. Even with the 6L flask, I would sometimes push foam out of the vessel. Wheat yeasts are particularly prone to creating a dense foam. This foaming was a problem, no matter how I adjusted the air flow rate. So, no airstone needed!
The need for oxygen during yeast growth is well known. But as evidenced here, pure oxygen is not needed during starter preparation. With beer wort, a good shot of pure oxygen at pitching time will bring the wort's dissolved oxygen content up much higher than can be accomplished with pumped air or splashing. Oxygen is fine when you have only one time when you'll be adding it. But for the extended yeast growth cycle during starter preparation, constant air supply is a cheaper and equally effective alternative. The primary concern is the cleanliness of that air.
I've found that 0.45 micron disposable filters that are intended for groundwater sampling are great for filtering air. Even though they are touted as disposable, they can be used for many years if you are diligent in keeping them sanitary. They have huge filtering capacity and should handle millions of cubic feet of air flow through them. There are 5 and 10 micron filters available too, but you need 0.45 micron or smaller in order to filter bacteria and other airborne debris from the air flow. You'll see an example of the filter I use here: http://www.geotechenv.com/disposable_filter_capsules.html
These are not cheap at about $15 each, so you will want to make sure it lasts a long time. I would still be using my original filter except that I allowed sanitizer to backflow into the filter once. I found out that you have to keep those filters dry!! Fortunately, my engineering company performs groundwater sampling and I was able to sweet talk another filter out of our stock. I use a regular aquarium air pump and vinyl tubing. The tubing will actually fit tightly into the filter housing, so no special fittings are needed. I am very particular about keeping the output side of the filter clean and sanitary. I remove the outlet tubing and place tape over the filter outlet while its not in use. The other end of the filter stays hooked up to the tubing and pump.
This has proven to be a robust and reliable yeast ranching method. If you're looking to improve your starter preparation, this will get you there.