We all use Google on a daily base and search has probably become the number one navigation tool on the web. This changes however on a site level, because good search engines are expensive and complex to integrate. The good news: in version 4.1 Kentico replaced it's rather slow on-site search with the enterprise level smart search that is based on the well-known Lucene.Net search engine.
Fast
There are two commonly acknowledged criteria for what makes a good search engine: speed and the relevance of results. I would add a third one: the representation of the search results
I integrated the smart search already on this site. Try it. It is incrediably fast. No compairison with the old search engine.
Relevance
It is difficult for me to make a call here. My site has less than 500 pages, but all searches I tried seemed to return decent search results. The results are marked with a relevancy meter. There is not much documentation available, but it seems that keywords found in the title of the page or the URL score more points.
Result representation
As always in Kentico, the site designer has full control over how the results are displayed. A nice feature is that searched keywords of the document title or the text snippet of the found page can be highlighted.
On my site I am also displaying the Meta description of the found page to give the user some more information about the found document.
The search result will highlight the searched item and display a relevance bar. In this example, the meta description and the creation date of the document is also displayed.
Controlling searchable documents
On the backend, smart search enables quite a few options to configure which documents or sections of the site can be searched: The way it works is that the system administrator has to create so called "search indexes". A search index contains keywords to all documents that can be searched. Documents can be restricted by their locations in the site (i.e. news section, product section, etc) or document type (news, event, regular page, etc.).
You can even define which fields of a document type can be searched. In essence, there are many scenarios that could be achieved with different configurations.
Enabling advanced search
For advanced search requirements, you can enable the "Full search syntax", which allows users to specifically search certain documents fields with a very precise syntax. In some cases you also may want to enable search filters to limit the range of documents that will be searched, or define how the search results will be sorted, but the scope of these details are beyond this introductory article.
Updating search results
The main reason why this type of search engine is so fast, has to do with the fact, that it is not actually documents that are searched, but the search index, as explained above. The search index needs to be manually created in the beginning and will then be updated by a regularly running task. The default value is 4 hours, but this can be changed to any other time frame.
Conclusion
The new search engine is a simple proof that Kentico wants to position its CMS to an enterprise level. If they do as well with future components as they have with the search engine, they will — no doubt — succeed.