Well... A lot of people think that there IS a clear-cut difference. That is kind of understandable because that is the way the DSM (classification system) presents itself. It does acknowledge that a dimensional approach would be more accurate, but that is in the fine print, and most people don't read the fine print!

What you have just described is the 'stereotype' of DID. When clinicians are training... They learn the diagnostic criteria, they surely surely do, but what is even more crucial than the diagnostic criteria is the little accompanying manuals that provide 'stereotypic' cases. (This is why the DSM has the little blurb in the front to the effect of 'the diagnostic criteria cannot be applied to an individuals symptoms in a rule like (algorithmic) fashion. it is intended to be used by a clinician with the relevant training in CLINICAL JUDGEMENT') the notion is that the more an individual resembles the stereotype (in the diagnostically relevant respects) the more accurately it is considered to be an instance of that diagnosis. the less an individual resembles the stereotype (in the diagnostically relevant respects) the less accurately it is considered to be an instance of that diagnosis. clinicians are provided with numerous stereotypic cases in their training so they get the hang of the scope of the diagnosis and what features / symptoms to look out for.

Then they proceed to meet real patients and get practice with seeing how the psychiatrists / psychologists classify them. This is harder because people tend to vary in ways the stereotypic cases did not. More indeterminacy. Basically... Practice practice practice until your judgement comes into line with the judgement of other clinicians. This is what is known as 'inter-rater reliability' (the extent beyond chance to which clinicians agree with the diagnosis that they give to the same individual). Inter-rater reliability is not high for most of the disorders listed in the DSM.

So... basically... You describe the current stereotype of DID. different individuals match some or all of the features of the stereotype to a greater or lesser extent. Whether an individual is close enough to the stereotype to be dx'd with DID or whether they are more properly regarded as having a 'dissociative disorder not otherwise specified' is a matter of clinical judgement. There was some concern that some clinicians judgement was over-inclusive (they were too willing to dx with DID). In an effort to curb the 'dramatic increase in prevalence' they reinstated the amnesia requirement. So... The DSM says that there must be black outs / time loss / amnesia in order for a dx of DID. That being said... One can always use ones 'clinical judgement' to apply the dx regardless... Or one could always 'coax' the behaviours by strategic theraputic intervention (whether it is recognised as such or not).