Agh, Canon's making things tough for me again.
I'd love to see IS on this lens. The relatively fast aperture and the pretty modest focal length doesn't make its inclusion as important as it could be, but I think that it should be done all the same.
More importantly, but perhaps less likely to happen, I'd love to see the ability to focus past infinity. Perhaps a mechanical detent would do the trick of giving smooth operation in normal circumstances, but the inability to focus past infinity makes it hard to throw the plane of focus dramatically sideways, such as running alongside a fence running almost directly away from the lens, at least on a crop body. With these lenses I don't care so much about hitting infinity; all I care about is being able to set up the scene I have in mind, and unfortunately the 17mm (and I assume by extension all the others) simply don't go far enough. (I'm not the first person to have made this objection, though I was relieved to see that I wasn't simply doing something simple wrong and to realize what the simple problem was).
I think that near-macro is (hopefully) a given. If not it should be. The 17mm focuses to almost macro ranges - it doesn't have macro levels of magnification, only something like .17 times magnification, but given how wide it is that's incredible: You can focus as close as 250mm. The TS-E 24mm II is a marked improvement over the original TS-E 24mm in this regard - from .14x to .34x! - so a 90mm replacement should follow the general pattern set. If nothing else, Canon are aware of the use of TS-E lenses for product photography and so will put more emphasis on close focus.
The Nikon 85mm f/2.8D PC-E Micro Nikkor
looks able to focus at about 390mm. This is closer than the TS-E 90mm, though not as dramatic an improvement as from the original TS-E 24mm to the TS-E 24mm II. The jump is bigger in terms of milimeters from the 50mm minimum focus of the TS-E 90mm to the 390mm MFD of the Nikon PC-E, but it's not as big a percentage jump as across the TS-E 24mm lenses. If nothing else: Canon has to at least rise to the level of Nikon's optic.
From what I understand the 100mm f/2.8L IS macro is pretty well-regarded, but offering tilts and shifts on top of that would open things up even more for me. And for crop bodies that 90mm perspective is starting to get pretty long - about the length Sigma targets one of their macros, I think; on a FF body it would become closer to my familiar 50mm focal length on a crop body.
I could use more autofocus lenses, but the incredible image quality of some of the TS-E lenses compared to the regular versions (check out The Digital Picture guy's TS-E 24mm versus EF 24mm f/1.4) makes me less worried about lacking that autofocus in most cases.
Boiling down - a simple replacement with no new features of the TS-E 90mm would be a compelling option for me.
A TS-E 90mm II with IS and impressive macro focusing capabilities would be an unquestioned top three optic choice, and probably a reflexive purchase.
Given that Schneider-Kreuznach presented the PC-TS Super-Angulon 50mm f/2.8 HM and the PC-TS Makro-Symmar 90mm f/4 HM at Photokina, I guess Canon will have to make the new version as a reaction to those lenses. I was actually planning on buying a TSE-90 until Schneider's announcement. If Canon also present theirs, I might be tempted to wait. But, as I always say, how many picture opportunities will I loose until I actually acquire the new lens?
Shows what I know, I didn't know about these lenses until your post.
It is one of those "collar" (don't know of a proper term for it) shift lenses, where I think that I like Canon's "scale" approach better, where you can tighten up the points pretty well. Don't know price, but I also don't know if it focuses past infinity. If it does, that would be a big point in favor of the Schneider-Kreuznach. If it doesn't, I'll be looking at price and how close if focuses, and lack of IS compared to any possibility of IS (however slim) in the Canon lens.