There are 185,000 results on Google for Mark Twain's line, “Reports of my death have been greatly exaggerated.” The same can be said for search.
In Days Gone By
Search may not be dead, but it’s certainly not the search we all knew and used as recently as 10 years ago (I really couldn’t say we “knew and loved,” could I?) Verity, Fulcrum, Thunderstone and other enterprise search products shipped with command like index utilities and proprietary user-facing applications that were intended to be used by enterprise end-users to search and find corporate content. Few platforms had programming APIs, and those that did were complex, clunky and generally licensed separately at a premium price. And often the "product" was not even written using the published API.
But then a funny thing happened: search went mainstream. People born with the ability to use Google revolted against internal search that was complex and hard to understand. IT fired up internal web servers and installed "site search." It wasn’t as good as Google -- even when it was powered by a Google Search Appliance. After all, IT doesn’t have people and analytic applications examining every user interaction to make it better -- heck, many didn’t even use synonyms and spell check.
Browser-based search is, almost by definition, "read only" and archival. You search, you find (or not) and you’re done. But most enterprise content existed on desktops or in file shares, and collaboration usually required finding a document on your intranet, downloading and editing it, and then uploading a copy and/or emailing it to your team. This kind of collaboration is hard to scale. You end up with dozens of slightly different copies, and no one knows which is the right one.
To solve these problems, content management systems started to adapt and improve. Documentum, SharePoint and Adobe all have CMS products, and many open source programs are available including Alfresco, Drupal and Joomla. All of these content repositories support various levels of collaboration -- and the key to each of these is the ability to find the right document and actually let the repository manage versions. All of this takes search, and search in these packages is more often than not based on the open source Apache Lucene/Solr project.
Click or Type?
There are two kinds of people on the web: people who like to discover content by clicking around and people who prefer to find content by searching. I call folks in the first category "browsers" and in the latter category "searchers."
Web content owners spend hours designing logical menus of options for their visitors who start at the home page and simply click around to find what they are looking for. But probably just as many site visitors are born searchers, and their first click on a site is the search box. I confess that I’m a searcher, and websites that don’t have a search box frustrate me -- especially when the site is for a company that sells enterprise search.
And then there are sites, generally in e-tailing, that let you start with search, but then become point and click. It works, but probably frustrates both browsers and searchers equally.
Sometimes browser people are searching and don’t even know it. Most of us would call pre-Oracle Endeca a search platform, but one of the things that Endeca let you do is create a user experience that appeared to be hyperlink-driven navigation. The site designer could build a site such that each apparent hyperlink actually performed a search to construct what appeared to be a static web page as a result. In fact, in any good CMS, it’s possible to build applications where both browsers and searchers are actually using search to get where they want to be. In these environments, every page is dynamic, every page is assembled just in time, and every page is based on a query.
The Common Thread
My point is that search drives most of the websites and applications that you use every day. Sometimes -- as on your corporate intranet -- search is a text box and 10 blue links. Sometimes it’s behind the scenes, delivering a customized path through your site, or retrieving the document you need. What would your SharePoint experience be without search? The customized email that comes every morning from your local newspaper with topics you’ve pre-selected? Would shopping online be as satisfying without the "people like you" suggestions?
All of these and more work because search is behind the scenes, customizing your world. So I would suggest that search is not dead -- it’s behind the scenes where we don’t need to think about it as much any more.