"Psst--wanna see Photoshop 15?"

John Nack was the product manager on Photoshop who worried the most about the encrusted menus and overload of expert features in Photoshop, as it went into it's millionth release cycle. It's a software industry problem, as he points out in more angst in this blog article.

From a design and usability perspective, it's a nightmare problem with few solutions, and the problem just gets worse. You can't remove features, you can't do major surgery (see the angry comments referenced in John's post), you can't NOT add features because that's what keeps you selling! And you want to add things that help your customers, that they've asked for.

You reduce the product usability with every new item you add after a certain point: it becomes increasingly hard for your customer to sort out what the task flow for any particular problem is (especially tasks involving multiple commands/palettes/menus), and it's harder for them to discover the new and interesting or even refind the old and true in all that built up cruft. John says: "And the sheer volume of options can be overwhelming. At one point I counted 494 top-level menu items in Photoshop CS. In CS2 we've added roughly 60 more, and that's not counting the new Adobe Bridge application. So, back to the hypothetical Photoshop 15: at our present course and speed, we'd add at least 350 more menu commands. We'll need to raise the minimum screen spec just to hold the menus!"

Splitting out a new app, like breaking it apart into different consumer market products (like Photoshop Elements vs. Photoshop for Gurus) are one way to go. I think that's not bad, actually -- your customers then have 2 things to buy, and the usability of each is a little better. Assuming there's a good place to cut them apart. Allowing customized menus is another (John's fix for CS2 release--he's not sure anyone cared); patching your workflow problem by adding "wizardy" walkthroughs aimed at common tasks is yet another way to go. You still need access points for the wizardy walkthroughs, which is yet more functionality on menus/palettes/toolbars...

This is a truly hard UI problem, which every mature software product faces eventually. I'm just pointing at the pain, I don't have deep solutions; it's at heart a business problem that in turn points to harder economics problems about why people buy things that aren't good for them.