Glass windows

>>The biggest weakness of glass is that they don't make very good radiation
>>shields. O'neill's design called for mirrors shining through glass...

The Ames Summer Studies used reflected light; As far as I know, they didn't
give artificial lighting a serious look. Apparently they decided to cut out
the losses due to conversion (not that such losses mean much when it's only
a little more mirror area) and the extra complexity.

>>but in our everyday lives, we wouldn't need "real" sunlight.

Actually there are many healthful qualities to filtered sunlight. That needs
to be factored into the cost/benefit analysis.

>>In fact, someday there will be interstellar colonies for whom there can be
>>no "real" light, only "artificial" light.
>
>Well, to be fair, even in interstellar space you could produce an image
>of a star as big across as our view of the Sun, with a sufficiently
>large mirror.

I'm a fan of the idea of nomadic interstellar colonies. You start out
assuming that Space Colonies will work, and be made to keep working long
term, provided a source of materials. Provided a transportation
infrastructure or source of resupply resources, it's stipulated that no
closed-cycle systems are perfected.
In O'Neill's book "the High Frontier" it establishes an arbitrary
"continental shelf" for colonies around the Sun, where the mirror mass would
be no more than twice the model 1 habitat. It came out to about 4 light-days
(around 720 AU). Out so far, the Sun is only a particularly bright star, and
there may be others as bright. At this point, -especially if you raise the
acceptable mass of the mirror), there's effectively no limit to how far away
from a star you can go.
Another source for reading of interstellar colonies calculated that for a
population in interstellar space, using bright stars for power, the mirror
could be forseen to be at least 3500km across to give each inhabitant an
energy budget better than all but the richest population of today (it being
assumed that energy formed the base of what must certainly be a
high-technology, high-energy standard of living. A mirror array like a
fair-sized continent, for a village of 500 people. ("The Prospects for
Interstellar Travel", a good read).
Such a huge mirror might seem ludicrous, but it's only metal film, and a
small industry to maintain it. Would an artificial light system -even if a
power source is assumed to be provided- need less? What about that power
source?

A true nomadic interstellar cometary habitat would need only a little
velocity change to take them to match course with another cometary body at
typical distances by the time they're running out of the present victim's
ices. A series of more changes and more comets eaten up by their factories,
and they make a little delta-V addition to take them to a comet that's
technically part of another star's comet family. They'll eat a few more
before they really enter another star's environs, and become the first
"interstellar colonists" even though little if any of the original
inhabitants or the colony itself remain.
The Zen way to intetellar travel: living the journey is more important than
any destination.

Acceptance of the possibilitiy of this kind of technical adaptation, and no
fundamental new inventions needed, was behind Fermi's estimate that there's
no reason why ET life forms, once having reached spaceflight, couldn't
spread across the galaxy (and they should already be here). It formed the
basis of Freeman Dyson's assertion that once we achieve space factories and
habitation, there's no reason why we couldn't outlive the Sun.

I almost like to assume that colonists -especially in the future, when
they're not assumed to come from Earth- won't particularly care if there's a
convincingly fake erzatz "Sky" or a Sun to move across it.
I'll accept that it's hardwired into us to like green growing things around
us, rather than a sterile, controlled urban environment only, but I fail to
see how anyone will prefer to have their eyes maybe occasionally decieved by
a painted ceiling with projected clouds and a moving sun spot. Accept that
there's probably going to be an overhead structure of some kind, even if
it's the other side of the inner spinning hab, kilometers above. Either
provide diffusion of the apparent source, or narrow it to a pinpoint source
if that's convenient (maybe we want to minimize the amount of livable hab
floor area the "windows" take up). Maybe even admit light from different
directions during the inner hab's "daylight hours".
Nobody will ever be fooled into feeling they're on a planet. And why should
they need to? They know it's filtered starlight coming from light
conduit/"windows" at different points on the "ground" on the other side of
the hab.
People adapt to living most of their lives in similarly weird-seeming
conditions on Earth. Constant clouds or rarely clouds. An ocean sky or one
narrowly framed by mountains. Surrounded daily by natural, wild greenery or
only rarely/never being in greenery beyond what their eyes can take in, in a
glance.