What kind of development process leads to such variety of sliders? Don't they have one GUI library team? Did they buy all the components from other companies that happened to have their own slider implementation?
I mean, sure, some specialized implementations are useful (e.g., picking a color/hue), but this is a bit excessive. Somebody must have written all these implementations. Somebody needs to maintain them. This costs real money... why would you want to do that?
Nope. They don't even have one string object or math library.
When I talked to a guy who worked on Mac stuff for Adobe, he said that pretty much anything that could be shared, isn't. Every flagship product is independent, and most were the result of at least one acquisition (some circular!). Pretty much every effort they've made to rationalize things into a coherent suite has just resulted in regressions, as none of the object models line up.
The engineers at Adobe know full well how shitty their flagship products are. Lightroom is what they can produce in green-field development, and it is awesome both in implementation (mostly Lua) and usability (no bloat). Apple's Aperture had a significant public first start, and was still quickly beaten with a superior product. Quite unexpected.
My experience with Photoshop leads me to believe that either its internals are a mess or they don't have a UI team at all or that that they're morbidly afraid of changing anything. Or all of those things.
Several different styles of sliders are the least of Photoshop's UI's problems. It still doesn't have some features we've been taking for granted even back in the 90s. I hope customizable toolbars and palettes finally made it into CS4 because in CS3 my screen is full of buttons that serve as nothing but a visual distraction. Also, looking my current PS layout, I can count FIVE different places where I can change the selected color and not all of them behave the same.
It looked like things were getting better when they finally implemented Corel-style context sensitive toolbars a while back, but of course they got it wrong by hardcoding them and only including certain commands based on who knows what criteria. But my (least) favorite f*ckup was when they introduced customizable shortcuts but couldn't be bothered to remove all the hardcoded ones first. This meant that some of your custom shortcuts would do something completely unexpected, like launch the scripting window.
I really don't understand people who go on and on about how amazing Phostoshop's UI is. For me, using PS has always been a highly frustrating experience.
Many of these sliders have special behavior which is quite useful. For instance, the "blend if" sliders have two controllers, each one of which can be broken in half by option-clicking it, the threshold slider shows a histogram above it, and many of them have keyboard control of one sort or another (which is not always reasonable to completely homogenize).
It is true that having such a diversity of styles is a bit absurd. They could do certainly do with no more than 4 or 5 styles, all mostly visually similar. Still, the actual behavior of all these sliders is pretty solid, in my experience, and I'm not really bothered too much, as a long-time user, that consolidating widget appearance isn't Adobe's top priority.
I would guess it is a combination of too much other work, too much legacy, and no one really caring about the aesthetics.
Marketing guys pressuring devs for new features. Enough late nights and you just stop caring about those details. Unless you have an ethos of good design like Apple.
I'm thinking different parts might be owned by different teams, or they were written by the same team, but over a long period of time. We have a lot of UI issues like this, in a 100 person company. I imagine the bureaucracy of a place like Adobe could be hard to move.
I mean, sure, some specialized implementations are useful (e.g., picking a color/hue), but this is a bit excessive. Somebody must have written all these implementations. Somebody needs to maintain them. This costs real money... why would you want to do that?