It's a nuisance that greenhouses don't work by the greenhouse effect. Some seem to want to make it out to be some sort of catastrophe or indictment of science instead. But it is annoying nevertheless.
The greenhouse effect's existence was discovered by Fourier in 1827. Namely, the surface of the earth is too warm unless something else were going on -- as we found in our simplest climate model. Digressing a second, Fourier realized this even though it was more than 30 years before anyone knew that there were greenhouse gases. Tyndall published his experiments, that showed water vapor (H2O) and carbon dioxide (CO2) among others to be greenhouse gases in 1861. Conservation of energy is a very powerful law -- you can reach the conclusion even if you don't know how, exactly, it comes about.
In any case, the idea was that greenhouses let the sun's rays in and didn't let the air in the greenhouse radiate energy back out. The definition of 'greenhouse gas' became this -- that they were transparent to solar energy (completely or nearly so) but absorbed earthly radiation (in the infrared, rather than the visible and near-infrared of the sun).
Unfortunately for purists, greenhouses don't work that way even though 'greenhouse gases' do. The way greenhouses work is to physically trap air near the surface of the earth. The sun's rays heat the surface (by radiation) and the surface heats the air in the greenhouse (by conduction). Since the greenhouse is closed, the warm air can't rise very far and the greenhouse stays warm.
Ok, those are two word pictures. Both are nice sets of words. But ... can we do an experiment that will show us which one is right? We can, and it was published in 1909. The experiment built two different greenhouses (small, but enough to test the idea). In one, the walls were the usual glass, which is transparent to sunlight but tends to absorb 'earthlight'. In the other, the walls were halite (rock salt) which is transparent to both sunlight and 'earthlight'. If the infrared properties of the walls matter, then the two greenhouses will reach different temperatures. If it is only important that there be a wall, the halite-walled greenhouse will get just as warm as the glass-walled. (Give or take a little for how well the two conduct heat.) The result was that the halite made just as effective a greenhouse as the glass. (Wouldn't work so well as a practical matter. Why? But just as effective for what was tested.)
So greenhouses don't work by the greenhouse effect. Planetary atmospheres, on the other hand, do. It's a nuisance that the words don't really apply thoroughly. But this is English. It's not like this is the first time there has been a definition that didn't make rigorous sense. We're talking in a language where flammable and inflammable mean the same thing after all.
Project: Build your own demonstration of the greenhouse effect versus greenhouses. See
http://www.wmconnolley.org.uk/sci/wood_rw.1909.html for the original experiment and a discussion of the significance.