It is difficult to address this aspect in present time as CachéRDD is yet to be tested by other developers. But, yes, I can shortlist what advantages CachéRDD may offer for an (x)Harbour application over its native or other 3rd party RDDs.
(x)Harbour RDDs suffer greatly for lack of scalability. Only recently Advantage RDD has given some strength to this fact but still more is desired. The key factor to today's business success is based on the fact that a very high rate of data-sharing among other agencies be possible. (x)Harbour RDDs cannot provide this feature at all because of proprietary nature of its table structures which can be accessed by those applications which uses the similar like tools which are dying day-by-day.
(x)Harbour RDDs can manage data on a very limited number of platforms. This also raises a question mark if one who is using this technology falls short of means for expanding his business operations for mere lack of hosting data to a different platform.
For ages, since Ashton Tate introduced dBase in early 80's, the index management of tables created with any dialect of xBase, be it dBase, FoxPro, Clipper, Xbase++, or (x)Harbour, has remained a point of lengthy discussions but without any real-time solutions. Indexes in (x)Harbour tends to be corrupted here-and-there and hence a lot of developers time is spent in supports just to maintain them. A lost index means lost data. Advantage Database RDD has greatly provided strength to overcome this deficiency but still not completely effective. Native RDDs, DBFCDX, DBFNTX are prone to this defect considerably, specially in networking environment.
None of the (x)Harbour's native RDDs have this capability at all. Advantage RDD has this feature but with few exceptions and limitations. Every Jack-and-Harry in database application development world knows about the importance of transaction processing.
This is a sever limitation. (x)Harbour RDDs do not support more than 2 GB physical size and around 400 million records per table whichever is greater. Recently the size of table has been increased to 4 GB. Most of the time developer is struggling to break-up the tables in smaller chunks and then writing the code to retrieve data as if those were one. I hope that Caché does not have such limitation.
(x)Harbour native database is not designed to adopt to newer technologies. CachéRDD overcomes such barriers. Developer can stick to his favourite code syntax for base development yet adopt newer arrivals to connect to same database.