It's easy to get caught up in the announcements of new and exciting stuff -- and even easier to have forgotten about them when they quietly leave the stage. Case in point: Cuil and Clusty.

You may remember Cuil (pronounced "cool"), the would-be Google-killer that launched about two years ago. It was going to have a bigger index and a better interface. Some of the innovations they tried out in their interface were interesting -- the rest of it, not so much. And after Cuil's fifteen minutes of fame, visitors largely forgot about the site. As it turns out, a fresh UI and lots of results aren't enough. Searchers actually like to see relevant results, more than anything. And now, it's curtains for Cuil: the site's offline.

Then there's Clusty, Vivisimo's showcase for their query-side clustering technology. Vivisimo has moved on to producing a full enterprise search product in Velocity, which doesn't just cluster but collects and indexes, as well. Clusty never became a huge hit. However, it has served Vivisimo well: first as a nice live example, and then, when it was quietly sold off a few months ago, it also made the company a tidy $5.6m. Clusty is still there, but now it's called Yippy. (Seriously, who makes up these names? Maybe Clusty and Yippy could cheer up Clippy, who's been out of a job for a while.)

Of course, both were public web search engines, which is an entirely different scenario than enterprise search. But if you're working on enterprise search projects, you may still find value in looking at the web for examples. In the case of Cuil, because they might give you some ideas for improving the UI of your own implementation. The bland Google-like copies most vendors ship as standard with their products could usually do with some improvements.

Clusty was interesting as a public example is a large-scale, working proof of concept. Many of the vendors we cover in our Search & Information Access Evaluation Report have started out in, or showcased their technology with, a public web search engine. A non-exhaustive list would include:

• Exalead
• FAST, who started out with alltheweb.com
• Google and the Google Appliance
• Autonomy and the Blinkx video search
• Vivisimo with Clusty
• Other smaller players...

These can demonstrate that the technology does, in fact, perform at some scale. You just have to be careful with assumptions there, since the public search versions are often highly customized, or even completely different software. They can also display some of the quirks of the specific vendors. Just throw some odd queries at them and see what they make of it.

So, let's say goodbye to Cuil and Clusty, and hello to Yippy. And let's keep public web search engines in the corner of our eye when watching enterprise search evolve.