I can tell you from my own experience that there are systems that routinely not tested as whole. Like, how do you do full test of a system that suppose to work in a zero-G environment? You can't.
Yes...you can. Zero G environments are simulated all of the time in the atmosphere.
Further more, many systems are taken into space on DOD missions and tested there. Yes...systems can be and are tested in zero G.
For some, a "full test" is the real deal and there is only one shot. Look at the Mars Opportunity rover. Everything has to work perfectly on the first attempt, and it is achieved without having done the so call "full test" .
Apples and Oranges, Engineer.
Is the PLAN intending to attack a carrier on the face of Mars?
No...of course not.
How many times did the NASA crash into Mars in proceeding to the Opportunity Rovers? How many smaller, shorter range, and shorter term NASA landers went to Mars? How many times did NASA simulate environments where they could full-up test those rovers? The answer in all cases is that they did it multiple time in leading up to somthing like the Opportunity Rover on Mars.
Testing is intrinsic. You cannot prove anything's reliability, and then improve on it, without testing. Sometimes circumstances demand that the real world environment and in operations constitutes the test..but that is rare, and most times not.
That is also not the case here. There is nothing keeping the PRC from establishing a test range out in the Pacific, as the US and other nations have done, and then shooting their missiles out there to test them.
Testing anti-shipping missiles is routinely done. First for the missiles themselves and showing that they work. Second to improve on them, and finally to ensure that the people using them are adequtely trained in their use...up to and including life fire tests.
I have worked on all of the following systems in the real world in the design/engineering environment.
The A-7 Corsair II Naval Attack aircraft
The US Army Multiple-Launch-Rocket System (MLRS)
The THAADS (Theater High Altitude Air defense System)
The NSSN program that became the Virginia Class attack submarines.
Before any of those systems went operational, a lot of prototyping and testing was done. A lot. In some cases vehicles and systems were lost as a result of the tests. But the tests allowed us to know exaclty what was working...and many tests were perfomred to understand how those systems performed in varying environments...both atmospheric, electronic, and for numerous other criteria.
The DF-21D is a system, with a big reliance on C4ISR that would absolutely require significant testing, from the component level on up, to ensure it operates properly and then to improve on the invariable problems such a complex system will have.
There is nothing "narrow," about this. There is no progganda involved in this discussion. This is straight design/engineering pracitse in any development, particularly for military systems.
Anyhow...I am repeating myself.
The Chinese have made huge strides and progress in their military systems. Qualitatively and technolofgically they are orders of magnitude further advanced and better than they were just 15 years ago. They have gotten there by not taking short cuts and I respect what they have done and are doing.
If they plan for the DF-21D to be a system that they must rely on, I do not expect they will shortcut it. They will test it as much as they possibly can.
Anyhow..as I already said...I have said enough.