Hi,
I also very interested in how DITA people implements outputting HTML and how they implement search features in their publication.
In search area, it is implemented in front-end JavaScript. In DITA-OT processing the HTML5 plug-in generates a JSON file that contain DITA topic contents such like this:
{ "title" : "About the contents of this chapter",
"body" : "This chapter provides a description of all parts, including options. ... continued.",
"url" : "topic_9903.php#topic_dzx_lky_bgq" }
The front-end JavaScript inputs this JSON and when user submit a search query, it scans this JSON and displays the search results. This function uses JavaScript String.indexOf() function. (In other words, it is so simple. It does not have language specific services.)
Another user uses cloud-based commercial search engine. They generates same level JSON and put them into AWS Elasticsearch index. When a user submit a search query, it is posted into AWS Elasticsearch engine and Elasticsearch returns the search results and relevant topic URLs. The front-end JavaScript display the result as the search result page.
Elasticsearch is very powerful and customizable. It supports many language specific features containing Japanese (very complex) or other CJK languages.
Also another user uses OXygen WebHelp product. Oxygen WebHelp is known as very popular implementation for XML editor users and works very fine.
I know that IXIASOFT adopts Oxygen WebHelp for their manual page. (Yes, it was originally written in DITA)
If someone introduce the search implementation, it is very appreciated.
Regards,
--
/*-----------------------------------------------------------------------------------
Toshihiko Makita
Development Group. Antenna House, Inc. Ina Branch
Web site:
http://www.antenna.co.jp/
http://www.antennahouse.com/
------------------------------------------------------------------------------------*/