Tags: ,

Although your website has an engaging design, a content managing platform which you have chosen so you can create and modify the pages your own way, users complain about how difficult it is to find the information they are interested in.

At the beginning, the project’s choice about saving all the contents in database seemed winning !

But with the passing of time all the important information started to become available in different formats:

  • Report e studies in PDF
  • Analysis and white paper in DOC
  • Presentation and case studies in PPT

All this data isn’t present inside the database and so the search engine we made isn’t sufficient to cover the new needs.
Don’t despair! To obviate this inconvenience IBM and Yahoo have found a solution to this problem which seems simple to install and able to index up to 500,000 documents in more than 200 formats, easy to integrate on your website and especially because it is FREE!

This amazing software is called: IBM OmniFind Yahoo! Edition and it is advertised as an enterprise search solution. One of the project’s key points is the simplicity of its installation and its configuration which guarantees you to be working in only a few minutes.

The installer is very significant, this means about 100 MB, since it includes a web server, a crawler, LUCENE framework and the administration interface. Once downloaded, the installation is very easy and all you have to do is choose which port needs to remain connected with the server web (easy if you already have Apache or IIS installed on your PC) and data authentication.

Schermata di login

After logging in, we go into the applicative’s heart where we can decide which domain or filesystem path we want to start indexing documents in. In the following screenshot you can see the result of an indexation trial which is able to scan a 69 document file.

Scansione di un sito web

Now that we have a base on which we can make our searches, let’s see what OmniFind makes us available:

  • Search engine look & feel customization
  • Definition of synonyms
  • Creation of suggested links
  • Searching statistics
  • Classification rules

Thanks to the web interface in AJAX style (DOJO libraries are used) we can customize the search engine’s design and the one which holds the results of the same search, in order to match it up with our website.

The advanced characteristics of search keys management allows you to define synonyms so that similar terms will recall the same keyword and the chances of finding important documents will increase. In a similar way, the suggested links will allow you to understand which is the most pertinent content to a specific search key which will be presented to the user in an area of prime importance just like Google does with the sponsored links.

Search statistics and the classification rules are used to monitor which searches have been made the most (even those that do not provide results) in order to eventually configure some suggested links ad hoc; instead, the rules allow fine tuning for ranking results, preferring more or less, the most recent documents, those which are closer to the websites’s source and those that have more links.

So far OmniFind proved itself the best for those who wish to create a search page for their own website following Google’s style but what if we desire to have a simple search box and if we want to present data in other modalities? Nothing easier! IBM developers have also thought about this necessity and have made available a complete API which allows the search in the engine’s index and to implement XSLT result transformations (on REST modality).

Let’s summarize this product’s significant features:

1. Easy to install.
2. Indexation up to 500,000 documents of more than 200 different formats.
3. Search page and results customization.
4. API to build custom applications.
5. FREE.

…and now you are a bit more conscious!