Of WordPress Search

One of my biggest issues with WordPress is about its search. The default search is not really useful, neither does it show the relevancy nor does it search the entire content for results. In the entire content I include the excerpt and the custom fields of posts at the least. It will become a serious issue as WordPress is getting more attention from the corporate world, and might probably become a factor for its comparisons with other candidates. Even from the usability perspective, it is one of the most popular ways of getting information and cannot be ignored. I have found that it is not only me, there have been rants and inputs and contributions from various corners and from various angles. Here are some contributions and ideas from around the web:

Advanced Search by Weblog Tools Collection

Weblog Tools Collection has a nice plugin (through a hack) for enhancing the search. It provides fuzzy searches and some nicer integration in the WordPress.

However, I have a couple of problems with this plugin:

• Changes WordPress Database Structure: It includes an upgrade script that changes the WordPress database structure. It actually changes the data types of the excerpt and title to FULLTEXT index. This is mainly because the search finesse of this plugin is based on the MySQL FULLTEXT search functions. The biggest disadvatage of this hack is that it will not work if WordPress changes the database structure that will disrupt this upgrade. And it still does not include the custom fields.
• No Boolean Searches:It does not allow boolean searches. Boolean searches are supported inherently by MySQL 4.0.1 onwards. These cannot be used by default since WordPress supports MySQL 3.23.23 version. However the plugin can use the wpdb variable to identify versions and use the inherent capability for version greater than 4.0.1 and David Altherr’s boolean search functions or Boolean Search Plugin by Beau Collins for the lower versions.

Search Reloaded Plugin by Semiologic

Denis de Bernady’s Search Reloaded Plugin is a more appropriate solution and is an actual plugin. The plugin actually does dynamic indexing. It reindexes whenever a post is edited, saved or published. This is fine, however, it will slow down the response for a heavier database. To be able to support this, indexing has to be carried out separately and probably as a scheduled task. However, it has its own complications and cannot be solved just through a plugin. And even this plugin does not look into other content types like the custom fields.

Latent Semantic Indexing

Elliott points towards Latent Semantic Indexing (LSI) for improving the search. It might not be a viable option today, but it might be future of the search industry. Natural Language Processing is already being used in certain search functionalities.

Using Google and Yahoo Search

Another option is using Google and Yahoo Search, something what WordPress.org search does. It is not very difficult to setup. However its problems lie elsewhere. It needs a public web presence, which might not be the case for corporate internal blogs. This has become more possible with the WordPress multiuser version out. Secondly, they will have their own restrictions, and thirdly the dependency on it. This is not going to help WordPress to make it a serious contender in the corporate space.

WordPress is one of the leading CMSs for building blogs. But blogs themselves have transformed into media publications and information management sites. With this happening, search will be a bigger factor to consider. WordPress has to support some solution inherently to be part of improved blogging solutions.

Technorati tags: search, search plugin, boolean search

Copyright Abhijit Nadgouda.