Better than not having the functionality at all.

Computing is so personalized these days that no single company making a general purpose tool (such as a browser, or an audio production studio, or an image manipulation program) can possibly implement every feature every user could ever want.

I think the concept (and in most cases, even the execution) of plug-ins is great. It is the future.

It's the next level, the next "layer" of of all general-purpose computing. In the old days, you had a computer that did one task only, such as calculating or census tallying. Then we had a general-purpose operating system with generalized I/O, and we could then run individual programs, each written by different people, each of which did different things in varying ways and we could pick and choose which ones we wanted to run. Now we've got the next layer: Those programs are like little platforms of their own, creating a generalized framework under which other little specialized programs can run. I think it's wonderful refinement of the overall computing experience, and I love it.

It also opens up the market more. Can't compete with Photoshop? That's OK, make your living by writing a great plug-in. Or, write another image-editing application that accepts photoshop-style plug-ins, and then suddenly Photoshop isn't as important as it used to be any more, and people might just use your application instead. (That's what I do, actually: I'm still using Paint Shop Pro.)

I agree that a lot of plug-ins suck. So there's some time I spend trying out a new plug-in, and sometimes deleting it if it sucks. But in general, I've found that plug-in architectures enhance my productivity more than they detract from it.