Review of ATL #11
Description
The presented tool is basically the same as the reference solution, however performance has been optimized by
- more sophisticated data structures (i.e. maps) for faster access
- usage of ATL helpers with caches
- refined initialization
The paper is extremely short, with the actual presentation of the optimizations spanning ~1/3rd of a page. It assumes, that the reader is already familiar with the reference solution, and with internals of the ATL runtime. A general introduction, how the reference solution works should be added. With this, the optimizations could be explained better, as mentioned helpers like getTree or getPartition would be in context of their usage.
Another interesting point would be, if those optimizations are general, i.e., if they are applicable to a wider range of transformation tasks.
Furthermore, in the paper unsolved "issues that are in the transformation code" are mentioned but not explained in detail.
Metrics
- Does the solution use the provided benchmark structure, i.e., is a batch execution supported?
→ Yes - Can all models be transformed within 10 minues?
→ Yes - What kind of result is produced? BDT, BDD, with fix order
→ not described - Is the transformation correct?
→ not described nor checked - Computed Metrics (like number of decision or assignment nodes)
→ Only execution time