The original taxonomy was created by educational evaluators (the people who write tests for college courses). It was geared toward helping them share different type of test questions. I find it does not work as well for creating training.
This taxonomy makes a lot more sense than the original one. I like that this one doesn't assume you're incapable of operating at a higher level in the taxonomy without completing the lower level. I also like that it goes into much more detail about how to use it for learning and teaching.
It also focuses on a new purpose for the taxonomy--identifying whether the objectives, teaching activities, and assessments all align with one another--that's helpful when you're reviewing a course. However, it only goes so far as to identify whether the same cognitive processes are used--in your objectives, teaching, and assessments. For example, if your objective is to have students learn to write a paper about how political science principles play out in history, but your lesson asks students to design a bicycle, the method in this book would say they align because they use the same cognitive processes.
The taxonomy also adds a new dimension--a knowledge dimension (making the taxonomy a matrix). At any level of cognitive process in the taxonomy (remember, understand, apply, etc.), there four knowledge dimensions: factual, conceptual (concepts, principles, and frameworks or mental models), procedural (algorithms and heuristics), and metacognitive. I like that this help clarify that something can be factual but still be understanding, or can be conceptual but only at the remember level, etc. However, it does become a little awkward because most objectives at a particular cognitive level tend to apply to just one level. For example, even though you can draw on facts and concepts when you're applying something, most application-level objectives tend to focus on procedural knowledge. Similarly, most understanding-level objectives tend to focus on conceptual knowledge.
Also, the specific knowledge areas can be a bit confusing at first. "Conceptual" includes not only concepts, but also principles and mental models. "Procedural" includes both procedures and the heuristic portion of problem solving (mental models are already covered as part of "conceptual"). This means that there's no way to look at the placement of an objective in the taxonomy table and tell whether it addresses problem solving or not.
The taxonomy also completely revamped the sub-categories of the cognitive area. I found the restructuring here to be very thoughtfully done. The original sub-categories were not especially useful for creating training--as witnessed by the fact that most people didn't even know they existed. I'd never even heard of them until I made the effort to actually locate and read the original taxonomy (the 200-page book, not the basic six-category summaries you see floating around). In the new taxonomy, each sub-category is a slightly different cognitive process, and therefore requires a different instructional method (for example, the two sub-categories in "remember" are "recognize" and "recall"). If one were to expand the taxonomy table so that each cell represented a sub-category of cognitive process and knowledge type, it could become a effective task analysis tool to identify the types of skills required by different activities. You could then use these to craft objectives, teaching activities, and assessments that should (in theory) all align. Of course the caveat stated above--where papers on political science and designing bicycles might look the same--applies here too.