While a pay per feature it seems a reasonable model for say an individual customer needing features X and Y, the world is more complex then that. It would be a disaster for standardization and code/solution exchange (which is already an issue). And it would be bad for multiple parties working on the same platform for the same customer as they all would need to be in synch database feature wise. It complicates matters in the developing process beyond what we have now.

For the same reasons I am strongly opposed to differentiating a product line (standard/enterprise/…) based on anything that you can programmatically do. Instead they could differentiate in matters of capacity or optional management/ security features, think for example of backup compression.

Right now, the standard editions of SQL Server are already severely handicapped compared to the enterprise editions. Things like table partitioning, compression, column store indexes or star join optimizations are not present in the standard editions. The huge gap in pricing between the standard and enterprise editions won’t for most customers justify the jump to that more expensive edition.

And as most people truly start to learn during their work, a whole generation of developers grows up never using these features or are even totally clueless about their existence. All Microsoft does with their licensing and product segmentation is handicap their product and in the market give competitors ample time to catch up.

It is strategically dumb…if only they were as paranoid as Google is for the emergence of new competitors! And in the end, only the paranoid will survive they say!