The ordinary meaning of "dangerous" is "[f]raught with or causing danger; involving risk; perilous; hazardous; unsafe": The New Shorter Oxford English Dictionary, 4th ed (1993), vol 1 at 591, sense 2. It describes, when applied to driving, a manner or speed of driving which gives rise to a risk to others, including motorists, cyclists, pedestrians and the driver's own passengers. Having regard to the ordinary meaning of the word, its context in s 319 and the purpose of s 319, as explained in the Second Reading Speech, negligence is not a necessary element of dangerous driving causing death or serious injury. Negligence may and, in many if not most cases will, underlie dangerous driving. But a person may drive with care and skill and yet drive dangerously. It is not appropriate to treat dangerousness as covering an interval in the range of negligent driving which is of lesser degree than driving which is "grossly negligent" within the meaning of s 318(2)(b) of the Crimes Act. The offence created by s 319 nevertheless takes its place in a coherent hierarchy of offences relating to death or serious injury arising out of motor vehicle accidents. It is not necessary to that coherence that the terms of the section be embellished by reading into them a requirement for proof of some species of criminal negligence.
The Court of Appeal erred in R v De Montero (2009) 25 VR 694 in finding otherwise, an error repeated in the case discussed below.
Further Edit: I recently had drawn to my attention the strong, condemnatory words of the Alberta Court of Appeal in R v Field 2011 ABCA 48 [Macfadyen, Watson and McDonald, at 22]:
The signal that must be sent out for crimes of this sort cannot be a mixed message. The target audience of deterrence is, mainly if not exclusively, people just like the respondent. Sentencing cannot focus entirely on the offender’s youth, or his otherwise good characteristics. Driving motor vehicles is a privilege, not a right, as it is governed by a vast array of licensing requirements and limitations reflective of universal public awareness that even largely safe motoring is perilous to the users of the road. There is a wide public consensus as to the need for the highest degree of social responsibility when driving. The public has a right to xpect that when using public thoroughfares according to the law, their lives and security will not be threatened by unexpected reckless conduct by other thrill seeking drivers.
Accordingly, street racing has a high degree of moral blameworthiness both on the side of gravity of the offence and degree of responsibility of the offender. The driver demonstrates not merely gross negligence but a willingness and intention to subject his passengers, the other competitor racer, and the public at large to great hazards. Driving a ton of glass and metal through spaces where people can be expected to be present and at a speed where it is likely to be impossible to stop the vehicle in time to avoid calamity cannot be treated as a youthful indiscretion. Street racing is entirely avoidable. There is no need for it to ever happen. A driver thus has plenty of choice not to start this crime and plenty of chance to stop it. Nothing impels the behaviour. Accordingly, what has been called the controlling emotion of fear must be enlisted by the law to deter such a choice. Since awareness of the risk of catastrophe seems insufficient to discourage it, then the tool of stern punishment must remain. As to denunciation, it is necessary for the law’s face to be set against such conduct with no ambiguity.
The case is currently on appeal to the Supreme Court of Canada.
Edit: The Court of Appeal recently considered the cases (and approved them) in a conviction and sentencing appeal King v The Queen  VSCA 69.
Also have a look at the more recent case of Rodi v The Queen  VSCA 48. Some issues of causation got a thorough treatment. The appellant unsuccessfully challenged both conviction and sentence after a jury convicted him of dangerous driving causing death and serious injury.
Ashley JA [at 50, arper JA and Hargrave AJA in agreement]:
The gist of the submissions for the applicant was that an agreement could not properly be inferred. I do not agree. Acceptance that nothing was said explicitly about racing at either intersection does not mean, having regard to the conversations and the events which ensued, that it was not open to the jury to infer the making of an agreement of the nature alleged. That is so notwithstanding that it required an inference adverse to the applicant.
There is a specific offence under the Road Safety Act for drivers who participate in a race on a public road. It is found at s 68 and reads simply:
68. Speed trials
(1) A person who on a highway drives or is in charge of a motor vehicle which
is being used in a race or speed trial is guilty of an offence.
Unlike dangerous driving (discussed here a couple of months ago) an offence under s 68 doesn't seem to require any accompanying element of dangerousness. The penalty is a fine with no mandatory loss of licence.
Racing cars on a public road can be dangerous, and it's unsurprising that a surplus of confidence combined with a deficit in ability can lead to one of the
'competitors' crashing. What's the legal liability of the non-crashing driver in these circumstances?
In Guthridge v The Queen  VSCA 132 the applicant was the non-crashing driver who, a jury found, had engaged in a race with another car which had struck and killed a passing motorist.
Neave and Redlich JA, Coghlan AJA [at 5]:
5 Although the men did not know each other, the Crown alleged that they had agreed to participate in a race along Sydney Road, which involved driving at a high speed in an industrial and residential zone, weaving in and out of traffic and failing to keep a proper look-out, and resulted in the collision causing the victim’s death. Although it was [the other driver], who collided with the victim’s car, the Crown case was that by becoming involved in a race with [the other driver], the applicant had acted in concert with [him] or alternatively had aided and abetted [him] and that they were therefore both criminally liable for the death of the victim. The applicant’s driving was thus said to be a substantial and operative cause of the victim’s death.
The appeal was advanced on the ground of inadequate (or erroneous) jury directions. The trial was conducted a few days prior to the Court of Appeal's decision in R v De Montero  VSCA 255 and so (unsurprisingly) the directions failed to accord with the test of dangerous driving the Court of Appeal outlined in that case (discussed here). The appeal was allowed and the conviction for dangerous driving causing death quashed.
The Court of Appeal went on to consider the principles of complicity as they relate to a non-crashing driver engaged in racing. The applicant had submitted that his mere presence at the collision location was insufficient to found a charge of causing death against him.
After considering traditional authorities that establish mere presence is insufficient to constitute aiding and abetting, the Court continued [at 107],
A participant in a two person race is not simply ‘present’ in the race with the other participant but engages in a process in which each person attempts to outdo the other. Whatever the precise goal of the race between two participants, it necessarily involves a notion of mutuality. Each participant attempts to drive faster than the other. In turn, this may encourage the other participant to attempt to outperform their competitor. It was a matter for the jury whether the applicant’s driving encouraged [the other driver] to drive in the manner that he did at the time of the collision.
108 The view we have expressed has been accepted in England. In R v Lee  EWCA Crim 240 witnesses had seen the appellant and the deceased driving their motor cycles at excessive speeds and doing ‘wheelies’ before the accident. The deceased was killed when the men collided and the deceased drove into a brick wall. The prosecution case was that the appellant could be convicted of dangerous driving causing death either on the basis that the death was caused by the dangerous driving of both men or on the basis that the appellant’s dangerous driving encouraged the deceased to drive dangerously, even though the death may have been directly caused by the deceased hitting the back of the appellant’s motor-cycle. The English Court of Appeal upheld the appellant’s conviction for causing death by dangerous driving, having accepted that the question for the jury was ‘whether each was, by his driving, encouraging the other to drive in a similar manner’.
109 In R v Haynes  EWCA Crim 1218 a number of witnesses observed the deceased and the appellant, driving at high speeds, before one of the men was killed when he lost control of his car. The men did not know each other. The Crown case was that the men were racing each other and that the appellant had aided and abetted the deceased man’s bad driving. The Court of Appeal refused to set aside the conviction of the uninjured man for causing the death of the other man by dangerous driving. One of the submissions of counsel for the appellant was that the men might have started off racing but by the time the man who was killed lost control of his car, the appellant had withdrawn from the race. Saunders J, who delivered the judgment of the Court, said that the incident of racing had lasted ‘at most a minute. The idea that over part of it the appellant was racing and then withdrew from it towards the end, is in our judgement unrealistic’.
110 Liability on the basis of aiding and abetting was also assumed to apply in R v Milburn  RTR 431, where the person who was killed was a third person who was hit by one of the two men who was racing.
111 We have already said that there was sufficient evidence for the jury to find beyond reasonable doubt that the men were racing, when [the other driver] collided with the victim. We therefore consider that it was open to the jury to have convicted the applicant of the offence of dangerous driving causing death, on the basis that the applicant encouraged [the other driver] to drive dangerously.
The Court stated a preference for the allegation put by the Crown to be put on the basis of aiding and abetting, rather than acting in concert, unless there is 'cogent' evidence of an actual agreement [at 112]:
112 In our view, where an accused participates in a race on a public road in which another participant in the race directly causes the death of the victim, it is preferable that the Crown present the accused on the basis that he or she aided or abetted the principal offender. Accessorial liability of the non-colliding driver as an aider and abettor does not carry with it the artificiality of reliance on an implied agreement or understanding. It has the further advantage that it would simplify jury directions. The jury would only be required to consider whether they were satisfied beyond reasonable doubt that the actions of the accused encouraged the colliding driver to drive dangerously. As we have already stated, in a case where there is cogent evidence of an agreement or understanding, it will be appropriate for the Crown to rely upon concert.
Abettors in summary offences are liable to be punished as principal offenders under s 324 of the Crimes Act 1958. (It's hard to believe more of these procedural provisions haven't been moved into the Criminal Procedure Act 2009). I haven't heard of a driver being prosecuted for another driver's crash in the summary jurisdiction, but this case may encourage it.