In this case, the circumstances made it very difficult to avoid a horrible tragedy.
My opinions on the findings:
1) While there may have been people that did stupid things and lost their lives, few of the deaths were in mobile homes or vehicles. That just gives a sense of its power. In addition, it took only a few minutes for the tornado to go from formation to a catastrophic monster, so early warning could not have helped much (unlike, say, Greensburg). Even the best warning response, given the circumstances, would not have helped. So this really doesn't apply here. The people and the NWS cannot be blamed here.
2) Sounds fairly reasonable. However, reaction time would be an issue here, since it touched down at 5:34 and became a monster by 5:38. So the individual warning texts seemed reasonable based on 5:30ish details. The later warning at 5:48 was well-worded (although the 5:42 update could have used stronger wording), but too late in both cases, and an earlier warning was impossible.
3) It would have been of little help here, but in many other cases (think the 4/27 report coming late fall), it might come into play.
4) Agreed the warnings need to be clearer and better communications is necessary, although again it would have been of little help here.
5) Inter-office communication definitely needs to be improved. That has proven to be a huge issue in the past.
6) Not that it would matter here, but rules need to be followed and clear.
7) Seems that NWS Springfield did a good job considering and that the rules were followed. The SPC forecasts were reasonable (a High Risk was probably not necessary since the threat was not widespread enough, and a PDS watch would have been of little benefit in this case and would have barely verified).
8) While strong wording was issued, a Tornado Emergency with extreme wording could have definitely been used at 5:42 and 5:48 on surface observations and afterward. It would have been too late for Joplin though, and there was no way such would have been justifiable on any warnings before then when the tornado was not yet on the ground. The 5:39 wording could not have had it since it was based on weak damage without a clear scan.
9 and 10) Probably the most important points. Radar scans update every few minutes, and that made the critical difference. If a scan was available from, say, 5:36, it would have been extremely helpful - but even then they would only have seconds to react. The scans available were at 5:34 - weak tornado touchdown with radar not suggesting anything more than an EF1 or so, and 5:43 - catastrophic EF5 tornado on the ground and radar-proven. Huge difference.
In conclusion, the report does make a lot of valid points - just even the best practices possible would have had resulted in an extremely short reaction time possible, like less than 1 or 2 minutes. With the data available, it was virtually impossible to issue the warranted warnings and statements necessary in time. It may have helped for southeast Joplin, but certainly not in the hospital area.