This is a mix of technical and general thoughts. Maybe others have voiced them in the past -- alas. Note that when I say "Wikipedia" I mean, specifically, "English Wikipedia, but not necessarily exclusively of all language Wikipedias, because I haven't checked those." Anyway:
1. Our images don't use the ALT tag. We use the TITLE tag, but only for the anchor link to the image page. What's up with that? Aside from being noncompliant with web standards, it also reduces our accessibility. And is an easy fix.
2. Is there a reason that there is a robots NOINDEX, NOFOLLOW on the image pages themselves (but only the Wikipedia ones, not the Commons ones)? It strikes me that this is likely part of the reason that Wikipedia and Commons' images have relatively low pageranks compared to their textual content. If you do a Google Image search for, say, "Albert Einstein", only one image on Commons comes up (out of 70 total, 13 of which are linked to from the heavily-accessed Wikipedia article), and even that doesn't come up as #1 like the articles do. I don't know how Google's image ranking feature works at all but I'm betting our NOINDEX and NOFOLLOW doesn't help that. (I suspect it might also have to do with the fact that our image pages have the extensions of images but are really served as HTML, but I don't really know. That's obviously not something we want to change at this point.)
(Should we worry about Google? Not to the point of doing anything contrary to our nature, of course. But given the popularity of Google and the popularity of the image search, why shouldn't we drive traffic towards our wonderful, painstakingly built and manicured collection of images, especially since many of them are free for reuse? If it's a copyright question related to the non-free ones, I am fairly sure that the burden in that case is on the search engine that is making little thumbnails of them, and even in that there have been explicit court cases affirming that such a use is "fair use". But in any case, I don't think that's any more of a legal issue than hosting them on Wikimedia servers in the first place, no?)
Just some thoughts. Hope this hasn't come up a billion times before -- if it has, please try to contain your rage and frustration. ;-)
FF