When estimating effort for a task, it's common to estimate in terms of staff-time, such as a team saying "it will take 4 staff days to do this", meaning it might take 2 people 2 days or one person 4 days.
One of the problems with this form of estimation is that you make it when thinking that you are working on the problem in a focused way, which ignores many of the other things people have to do as part of their job. Attending meetings, carrying out recruitment interviews, some customer support - all can reduce the amount of time you can focus on your programming tasks in a given day.
With ideal time you specifically estimate in terms of focused quality time "if I'm doing nothing else, how long would this take me". We then map from ideal time to actual time using a load factor. So if you only got 6 ideal hours of work done in an eight hour day, your load factor would be 6/8 (0.75). This way people could estimate without worrying about external factors and we multiply by the load factor to figure out how long in elapsed time a task should take.
This, of course, raises the question of how we determine the load factor. Our advice was to measure it by using XpVelocity. You'd do this by looking at recent iterations, taking all the stories that were delivered, adding up the ideal time, and comparing that to the elapsed time to determine the load factor. For example, we looked at the stories done last week and added up the ideal time for them to be 23 hours. The elapsed time was 40 hours, so the load factor was 0.6 (23/40 to one significant figure).
Using idea time fell out of favor since StoryPoints were easier to calculate, less likely to be abused, and every bit as accurate.
You can find some more information on ideal time in the tasteful green book, although even by then we were preferring story points.