Ars reviews the Steam boxes; the controller stuff is on
page 4.
No real surprises; you can tell what the controller is good for and what it's not good for just by looking at it. Ars says it's excellent for FPS's (the next best thing to keyboard and mouse), decent for point-and-click games that don't require precision timing, but not so great for point-and-click games that do, like Torchlight. Its lack of a D-pad means it's exactly as poor a fit for D-pad games as you would expect, and apparently it's absolutely terrible for games designed to use a dual-stick control scheme.
A
second article where they chat with the hardware team suggests that revisions and alternate models may be coming.
[...] another Valve design-team veteran, Robin Walker, informed us that the company's "always in beta, always in development" ethos was indeed extending to the Steam Controller, if not other products like the Steam Link.
"The likelihood that any one particular controller design is right for all of our users is extremely unlikely," Walker said during our meeting. "Our goal was to get to a point where hardware could be as flexible as software. We’ve learned over the years, when we build software and it’s in hands of customers, they help us improve it. Customers using a product always improve a product."
What does that mean exactly for customers? Is the first wave of Steam Controllers going to become eclipsed by a nicer model within months? Neither Valve rep we spoke to was clear on that point. Johnson repeatedly referred to "a specific approach to how we're manufacturing the hardware," which includes "a really flexible automated assembly line" for the sake of pushing newer designs to customers on the fly.
I'm pretty happy with my PS4 controller. But there are definitely cases where I could see having a Steam Controller around as an alternate. Like the other day when I fired up Shadowrun Returns in my living room and found that it doesn't have gamepad support.