Taskets are not yet defined concretely, so nobody knows, if they are even equivalent (transformable) to OOUI descriptions, but the following reflects my current guess.
OOUIs differ from taskets in that they are independant of an UI approach. Taskets don't imply a certain kind of rendering (e.g. 2D), but they organize the UI and thus are limited to a specific UI approach.
E.g. text based applications are usually task-oriented, in that you first choose an action to perform and then select the object to modify. Most good graphical applications let you first select the object and then define what to do with it. You can't create a good graphical app by taking a text based app and rendering everything graphically. Thus, rendering taskets in 3D (and maybe changing the rendering in the content area) is likely to be insufficient to create a good 3D app.
IBM Research seems to have developed a very similar method for creating UIs, called OVID.
However, both methods were developed completely independently from each other, including the starting idea.
In fact, I only know of the existance of OVIDs through my research during working on the OOUI proposal. Until time of this writing, I intentionally don't know any details about OVIDs in order to avoid influences.
Update: Now, that the overlapping parts are finished, I looked over the description and it looks like OVIDs are a subset of the OOUI proposal. OVID is a methodology for designing UIs, but the introduction states “As structured as any methodology may be, however, design will always involve the judgement and inventiveness of designers.”. I believe, that this statement is wrong, and the OOUI proposal is based on that believe. However, the remaining, strong similarities between the proposals are interesting.