To expand on your point, if one was to sustainably grow all vegetables required for a healthy, sustainable diet, then currently this can be achieved on around 100-200m2 per person (on earth, requiring no extra fertilisers / soil etc) based on research by John Jeavons. This hasn't taken into account external rainfall or air movement, but the basic premise is fairly well established. Throw in water treatment / recycling and humanure composting to close the circle completely, then all you need is a light source (assuming a closed atmospheric situation....)
Assuming that we can compress that down into 50m2 per person, the spatial requirements for a sustainable colony blow out quickly to large proportions, and we don't have information over the long term of just how well such a closed environment can nourish a small population - there's literature and research suggesting that we must have animal products in our diet for long term health, so that quickly adds up.
Given all these factors, supply runs from a known good source (earth / the mother country / home) are a sensible choice - but once you start roaming much further, it isn't an option any more. This is the stage we'll find ourselves at in 10-20 years time, as space travel will be economical to the point of sending out pioneer crews to space.
The closest long-term information we have comes from Biosphere 2. They spent a lot of time managing their ecosystem, and I think that would be true of any similar attempt for the next few decades. They had about 1500m2 per person, which included "water treatment / recycling and humanure." By comparison, ISS is 837 cubic meters, so assuming 2m for z gives a bit over 400m2 total.
We have a long way to go before getting that sort of volume. (Which, yes, is precisely what you said. I just wanted to work out the details for myself, out loud. :)
BTW, there's many people who have gone their lives without animal products in their diet, so I don't know what literature and research you are talking about.
I strongly doubt it will be economically viable to think of building (near) self-sufficient colonies for many decades. Brin's essay was very influential on me; it's much easier to build a self-sufficient colony in the Gobi desert (or the Sahara) than in space, so I would expect to see those first, if there's an economic need for the space.
If asteroid mining, or He3 mining of the moon, gives the economic impetus for a long-term off-earth location, then I look to oil platforms or McMurdo base as more relevant example.
Assuming that we can compress that down into 50m2 per person, the spatial requirements for a sustainable colony blow out quickly to large proportions, and we don't have information over the long term of just how well such a closed environment can nourish a small population - there's literature and research suggesting that we must have animal products in our diet for long term health, so that quickly adds up.
Given all these factors, supply runs from a known good source (earth / the mother country / home) are a sensible choice - but once you start roaming much further, it isn't an option any more. This is the stage we'll find ourselves at in 10-20 years time, as space travel will be economical to the point of sending out pioneer crews to space.