Hold on David, I take exception to some things:
Too often it seems that database professionals appear to be defending the SQL status quo rather than standing up and acknowledging that "SQL is broken. We need something better".
SQL isn't broken, but that isn't to say that it solves EVERY need. What I see most is tweaks to existing features.
I can only recall one ocassion when someone other than myself raised the question of the need to break with the SQL past and move on to something new.
Like what? The statement "we need to change" is useless unless there is some alternative. I think most of us would welcome a discussion is less abstract terms, something new like......
My own wishlist would include:
A truly relational data model and language
As long as you don't lose the set based operations otherwise all you get is RBAR and quite frankly THAT is a step backward.
General purpose integrity constraints with optimiser support
Please elaborate, this sounds interesting..
Strong typing and type inheritence
Maybe..
Self-optimising storage engine with more flexible physical storage options
Ok, first part self-optimising, most try to do this now, sounds like a side discussion of the post about "shouldn't it just work". More flexible storage options, ok, sounds good, what have you got in mind?
I like to work in specifics, I think that the deadening silence from the DBA crowd is from the fact that nobody has really offered up anything really as an alternative.
You brought up cloud technologies, the use of those has appeal when what you want to do fits into what they provide, many of them are relational like Azure and the new Amazon DB service. I looked at Amazon Simple Db and found that it didn't fit ANY of my requirements. You can often find a technology that fills a very definable niche, some of the non-sql variants do this, but very few of them fill the role that SQL Servers do, so when someone comes up with something that does that, THEN I think there will be a discussion..
CEWII