I’ve become convinced that the next big leap in software won’t be built by programmers. It will be built by software users.

I had this epiphany during a meeting several days ago with Adobe Fellow Martin Newell and imaging director Dave Story. Newell was talking about the pace of innovation in facial recognition technology, and how Adobe Systems (ADBE) is working toward the day when its Photoshop image editing software will be able to “understand” images not simply as a collection of pixels, but as unique objects – when it will be able to search a batch of files and find not just file names, but pictures of

cats, or pictures with women wearing blue hats.

The best and fastest way for Adobe to accomplish this, it seems to me, is to get its giant installed base of software users involved.

Take facial identification as an example. Adobe already has software that can find a face in a photograph. What if it let Photoshop users tag the faces in their photos, and share that data with the rest of the world? For example, a person could add metadata specifically to the part of a photo that contained the face, and identify that face as belonging to “Angelina Jolie”; the person could choose either to share that information just within his trusted group, or to send that information to Adobe’s central database.

A savvy move by Adobe would be to devise a system for giving software users credit for their contributions; taking a cue from Digg, it could award them points based the number of valid face identifications they submit.

Adobe could then use this information to build an amazingly valuable tool that could identify photos of Angelina Jolie, even if they haven’t been tagged; or, just as valuable, it could identify photos of people who look like Angelina Jolie. The more people contributed to this database, the more capable it would become. It could include cars, clothing items, collectibles. Eventually it could even include video.

Adobe is in a unique position to start building such a user-enriched database since it owns Photoshop, the ideal platform on which people could do the tagging. But if Adobe doesn’t do this, it would also be a smart disruptive move for Google (GOOG), Yahoo (YHOO) or Microsoft (MSFT). Adobe is in the best position to successfully build such a community, but Google, Yahoo and Microsoft have more to gain from doing it. Not only could they sell ads around user image searches, they could sell photos and video related to the searches. (Google bought an image search company a few months ago; Yahoo owns Flickr; and Microsoft Chairman Bill Gates, who owns a stock image company, has plenty of them to sell.)

I have to note that if a company were to build such a user-powered database, it would have to be very careful about the inherent trust and privacy issues. It’s one thing if people can do online searches to find text about you; it’s quite another if they can search for photos of you that a random stranger might have snapped at a concert over the weekend.

Still, I think the right company can answer those concerns and bring forth a product like this. And I should note, the concept of user-refined software works far beyond image recognition. Almost any application you can think of could benefit from it; Intuit (INTU) could use it to let users tag and organize merchants, so Quicken users can better track and organize spending by category. (Imagine a charity’s community had tagged it not only “charity” but also “health,” or “children” or “religious.”) Restaurants could be tagged by the type of food they serve, so people could see how much they’re spending on fast food versus candlelit dinners.

The idea is to get the community to do the work of improving the product, and letting them share in the benefits as a reward. It’s already working on the Web, as we’re learning from Digg, Yelp, Amazon (AMZN), eBay (EBAY) and other sites where people gain status in a community by contributing valuable data.

It’s time to bring that revolution to mainstream software.