If you've never been in a room that was badly proportioned due to too-high ceilings, you may not really believe there is such a thing. But there is, and it feels awful. Just very wrong -- but hard to put your finger on why. As a general rule, 9' is safe and comfortable for just about any room size. Higher ceilings can be are nice in large rooms; and 8' ceilings are wonderful for cozier spaces. One trick we've learned in this house is that a small area with lower ceilings at the entry to a larger space really 'opens up' the room with 8' ceilings. Our downstairs hallway has 7'6" ceilings, which we thought was a problem until we lived there a while. But now the 8' ceilings in the kitchen and den at either end feel much nicer, and the 'transition spaces' really function as transitions.

The structural support of typical sloping roof rafters must resist horizontal forces as well as vertical forces in order to prevent the walls from bowing outward at the top causing the roof to collapse. In the current condition the ceiling joists act as horizontal ties to resist the horizontal force. There are only 4 ways to remove the ceiling joists and use the existing roof structure. 1. Remove the ceiling and leave the ceiling joists in place possibly removing 1/2 or 2/3 of them and reinforcing the top plate as needed. I've made the beam brace and reinforced top plate approach work with horizontal members as far apart as 10 ft. but it would be foolish to do that without the assistance of a good structural engineer even if a permit could be obtained without it. 2. Sister the existing rafters with larger members (2x10, 2x12, etc.) and raise the old ceiling joists to about the 1/3 or 1/2 point of the rafters (this approach is spelled out in a footnote to the IRC building code rafter tables). "Collar ties" in the upper 1/3 of the rafters provide little resistance to rafter forces; they are essentially substitutes for metal hangers to secure the rafters to the ridge to resist wind uplift. The connections are critical and often under-designed so they should be designed by a structural engineer. 3. Install a beam under the ridge with posts at the ends that carry the load to the foundation. Then there are no longer any horizontal forces to resist and the ceiling joists can be removed. The beam and its supports are required to be designed by a structural engineer. 4. Remove the ceiling joists and install diagonal braces from the top plates at each rafter heel to the middle of the rafter on the other side of the roof. Apply a new ceiling to the bottom of the new braces. This is a field built scissor truss that cannot and should not be designed by anyone other than a structural engineer. If the members and connections are not adequate the roof could collapse. So, the first step is to accurately draw the existing conditions in plan, elevation and section; then ask some builders what the costs might be; then hire a structural engineer to design it; then bid the project.

Thanks for the great replies. I should have mentioned the climate. I'm in Western Oregon (Portland) so overall things are pretty mild here most of the time. We typically get a week or three of 90-100 degree weather in the summer and maybe touch the high teens every few winters with little snow. While it rains a lot we basically never get hot and humid conditions like you see in some parts. We're apparently "zone 4c". Since we have the finished attic (where we sleep) I assumed that ventilating will be a benefit in the summer when the sun tends to bake the roof and make it pretty hot upstairs, but that's open to discussion. @maingrower - no, the existing decking seems to be in good shape, but 2 out of 3 roofers want to replace it anyway. I'm not clear on why. SIP panels look interesting, as does the spray foam... but how does the roofing go down on top of that? Another layer of decking? Any suggestions on relevant Building Science sites? What I've found on the first pass seems mostly relevant to new construction. Thanks again!

I realize this is silly and not even close to relevant. But... From you... "I just measured the temperature of a heatsink in an open trim can at 140° and if there were nowhere for that heat to go, the bulb would fail sooner than later." and also from you, "Failure is failure, and a condition of continually too hot is supposed to be even harder on them regardless of what kind of an exotic name they give their test." and finally one more, "When you have 120° degree summer heat pounding down on a cathedral ceiling roof, and the heat sink is 4" underneath that with insulation on the sides, it keeps the heat from escaping the bulb and it WILL heat up to the failure point." Now from the report that you provided, "The lamps were operated continuously at an ambient temperature of 45°C" and "While none of the LED lamps exhibited catastrophic failure—and all of the benchmarks did—the early degradation of performance is concerning." If you examine the actual study further you find that only two of the 15 samples were below purported luminosity after the test. Remember NONE FAILED. If someone maybe said earlier in the thread, "The concern with operating LEDs above their optimal temperature is usually luminosity degredation not failure. Wait, someone said that earlier you say? Could it have been me? Surely not, because you shot back with, "Put the lamp in a ceiling-mounted fully enclosed fixture and set the timer for when failure will occur." To put this in perspective, Assuming you operated the light for four hours per day in ambient temperatures in excess of 113 degrees, and 8 hours on the weekend at those temperatures, and the temperatures remained that high for 33% of the year, AND you chose one of the 13% of bulbs that don't last long, you will only get 9.6 years before your lights are noticeably dimmer (remember not failed, just dimmer). Of course, if you picked one of the others that were predicted to fail you can extend that another 9 years or so. While if you got lucky in your bulb selection, you are good for the long haul. I never argued that LEDs were good, I stated that they were better than any other alternate bulb. And they are... Now, please take your time, come up with some more evidence that completely contradicts everything in your previous points while you deftly switch your stance (since you have managed to go from led bulbs are bad because they fail, to led bulbs are bad because they degrade, to led bulbs are the best but can lights are bad in this application).