There seems to be a lot of interest in procedural landscapes just now (can't imagine why), so I've done a simple applet
to show the basics of 2D perlin noise and supply some source code.
Share and enjoy!
Although I personally have always had a great interest in fractals and procedurally generated content(landscapes and more!)
I would take a wild shot and say that Minecraft could be a big inspiration for procedural landscapes. The ability to generate beautifully rich maps without having to handcraft every element is very enticing. Especially since most programmers are not necessarily artists or level designers.
Also, I give credit that computers are having a lot more power potential nowadays. Even the big developers struggle to fill quality content. For the most part, as a small developer, your IMHO Not going to be maxing out on potential power usage. And the possibilities of generating maps (infinite maps) is very enticing to "add" a lot more 'content' to your game.
Though I think a lot of people may have great early interest in it. They underestimate the requirements to properly populate these worlds with content worth exploring. Even Mojang doesn't have an insane growth in that area(They continue to add biodomes/biomes(sp?) and other content, but at a slow/steady pace).
Sure we can make infinite worlds with various terrains. But I still see a little too much repetition in that terrain. And terrain alone is relatively useless after you get over the initiate "that's neat" factor.
Another great example is Spore, it has a lot of procedurally generated terrains and creatures. Even despite having a wide variety of "monsters/creatures/animals" and although I didn't see the "exact" same creature over and over again. I did start to feel like I had seen them before, the repetitiveness starts to shine through. I think that some people overestimate having a ton of "unique" content is going to make the game so much richer in detail. It sorta is overlooking the point. Heck, look at chess or tetris. Billions of ways it can turn out, with a very narrow scope. "content" and "variety" does not usually translate to better games(gameplay) or increase fun factor.