Technology for reading pdf files on the Internet with annotations?
By "online reading" I assume you mean no PDF reader plugin on the client. In this case, you will need to convert to HTML
http://pdftohtml.sourceforge.net/
If you don't mind losing the ability to copy text, then converting to PNG might give you more accurate rendering.
Regardless of the output format, you can control the search using the original PDF data. One technology for this is mnogosearch
Monogosearch uses pdftotext internally, you may find it useful if you want to write your own search routines. pdftotext is part of the Xpdf suite of utilities.
http://www.foolabs.com/xpdf/about.html
All of the above tools are available on Windows or Linux
a source to share
You may also be interested in the Vuzit DocuPub platform: http://vuzit.com/products/docupub_platform
The display technology itself is not open source, but they provide an API to access their service, so it might be worth exploring.
a source to share
PyPdf is really nice. It supports text reading as well as encryption, which I know itextsharp does not.
Of course, you will need to program in python as the IronPython class libraries are not entirely accurate so that you can reverse them from another language and use them. (But I think they will someday soon) PyPdf
a source to share
It's not open source, but check it out anyway. You can download a free trial version of your SDK to give it a try. Reading PDFs and their annotations is not easy and I wouldn't trust a production open source decoder application.
Here is an online demo. http://www.atalasoft.com/ajaxannotations/default.aspx
Another good PDF reader is FoxitReader.
a source to share