I’ll echo Paul’s comments on the NTSB’s new CAROL system for searching accident reports. It comes across as designed by someone trying to maximize the difficulty in accessing accident information. As part of my work, I read 100 NTSB aviation accident reports a month. The previous search system was fast and intuitive. The CAROL system is anything but intuitive and relies on has a multi-page instruction manual filled with arcane jargon that simply isn’t helpful. I’ve spent several hours working to figure out CAROL and can use it for most of what I need but am constantly frustrated by its shortcomings and the fact that it hides information a user needs to do a search (good luck figuring out how to find accidents between two dates - you can do it, but figuring out how is time-consuming).
I’m also concerned that a number of accident reports for the time period near the end of that covered by the old system and the first year or so of CAROL have been lost or cannot be accessed. I’ve no data, but in 15 years of searching and reading accident reports monthly, I’ve gotten a feel for how many accidents will have occurred for a type of airplane during a year. When the system returns only a fraction of that number, I get concerned.
What has made it worse is that when I’ve tried to reach out to the NTSB through the contact information in CAROL. I’ve never gotten a response to my questions.
The NTSB had an excellent system for searching aviation accident reports - it screwed it up with CAROL. While it might be fun to assert that it was done on purpose, my experience with large bureaucratic organizations in the public or private sector is that it’s just the result of incompetence along with high-level supervisors insisting on having things done their way rather than in ways recommended by subordinates who actually know what they’re doing.