Our scraper back then was light years ahead of any others, both in regards to performance (think checkpointing of program state) + accuracy of parsing and JS execution (back then the target goal was a complete emulation of IE6+, sans ActiveX (exception being various host objects such as XMLHttpRequest))
This scraper was easier programming then anything else out there (think visual domain specific programming language). 7 years after having left that company I was called into an interview at a rather big HR company, with them essentially asking me to meet and greet a group of 20 people, all of them working with an old old version of said software, that I had worked on releasing - as a team of two - many years prior. Quite the bizarre encounter telling them how to work with a product that was IE 5.5 compatible, and drawing on old legacy knowledge for said purpose.