To elaborate on thermal bridging since it was brought up again... there are several ways to calculate it, but to keep it as simple as possible, here is how it affects your total wall's insulation. First thing, if you put in r20 rated insulation product, you are certainly not getting an r20 wall for cavity walls. Other products such as ICF and SIPs (some SIPs even have thermal bridges) are nearly thermal bridge free and have what is called continuous insulation. My home is an ICF house from the footings to the roof, and there are no thermal bridges anywhere until the roof line. This gives me an r28 continuous insulation. However with wood frame structures, you have a framing reduction to take into account. Most new homes, studs, heads and plates take up anywhere from 18-25% of the wall area. (lets keep the numbers whole and rounded for easy math) Say you select blown fiberglass in a 2x6 wall, gives you roughly r20. You have 1000sqft of wall surface. That r20 is only for the stud bays, and every 16" roughly you only have an r1 due to the stud (framing factor) So you have 800 sqft of wall at r20 and 200 sqft of wall at r1. . Your "r20" wall is performing at an r16 on average with the studs acting like heat sinks. Add just 1" of XPS foam to the exterior (r5). Your r20 turns into r25, and your r1 turns into r6. Now 200 sqft of your wall is an r6 vs an r1 and reduces the heat sink effect.The materials to add that 1" XPS would be roughly 500-600 for that 1000 sqft wall. The upgrade from blown fiberglass to open cell foam is typically more than that. For adding exterior foam, details are pretty simply up to about 1.5" thick. After that, the details change and become more costly. Luckily for you, in zone 4. 1.5 is about as thick as you need to go. 1" would be a good start, and your details really do not change. I have detailed 1" foam and windows 1 of 2 ways. Either install the sheathing and foam and then the window. Or with 1" foam, most window flanges are 1 1/8"+ wide, so you can install the window and then install the foam. This gives them about 1/8"+ lap for a caulk joint for trim to hide the foam. The trim stands proud of the window with this detail obviously. 1.5" adds another step. I like to detail a frame out around the windows and butt the foam into it. So frame the wall as normal, install sheathing, then rip a 2x4 in half to give them 1.5x1.75 furring to frame out around the window RO. Install the foam, butt into the furring around the window, and then set the window, nailing into the furring. Everything planes out, and quite simple. Obviously this can also be done with 1" foam too, its just not necessary. There are lots of details floating around out there. Check out green building advisor as well. You can also checkout my blog for more details and nerdy building information. Again, a very untechincal explaination, but you get the idea hopefully. Here is a link that might be useful: home building blog

"With the air tight IC recessed light cans you should still build air sealed enclosures around them because they leak a lot apparently. LEDs are a great choice for these locations especially without enclosures as they have a weaker "mini chimney" effect." in all my projects new and existing I've found many ways to seal IC cans. retrofit with inserts. then seal housing to sheetrock ceiling...from inside the house before putting trim kits in place. even 'air tight' trims don't stop the oversized cuts in sheetrock from leaking. I use a mastic tape made by Hardcast #1402 for all my holes in ceiling. more on this later. this seal described above is the best with no measurable air leakage. building boxes..maintaining the 4" air space on all sides & top of can light..also works, cheaper cost..harder install. problem is the 'legs' of the can lights that are attached to ceiling joists & properly sealing around them. this is where the air leakage occurs. I never use foam..because it shrinks over time, you have little control over where it goes & unless it is fire rated..shouldn't be next to heat producing areas. and yeah..I'm overly cautious about this. this method can be well done..and still leaky. foam installs I've seen online in videos put fg batts around recessed lights and foam them. I don't like this at all, and it is never done on my jobs. best of all is to use ICAT only, and use the hardcast tape to seal the cut between housing & sheetrock. Hardcast brand mastic tape #1402 is a 3" wide strong mastic tape. I perfer it over any other brand on the market. cut it into 1 1/2" wide strips and use it in full width depending on size of opening to be sealed. sticks to just about any clean dry surface...to the point of pulling paper off sheetrock if you try to remove it. so use it carefully. opeings at bath fans, supply boxes, stove vents, recessed lights..I use hardcast. use it once..and you'll understand why. not the cheapest on the market @ $22- $35 per 100' roll. but absolutely great stuff. while it is hard to describe the sealing of supply boxes with hardcast from inside the house..a picture shows it well. I'll attach one to this post. flgargoyle .. foil bubble foil is an overpriced product that performs no better than the cheaper radiant barriers. in my house with a tin roof..I have a single sided foil scrim fraft paper like what is used for ductwrap. I used button cap nails to attach it to the undersides of my rafters. with gable end vents..and 3 hurricanes later I've had to do a little repair on it..but it was worth the few bucks it cost me & the 2 mornings it took to install. good benefit in summer little to none in winter. my attic has fg batts on attic floor, one layer in joist bays the other unfaced layer laid across joists. no ducts, no equipment no recessed lights and a damn near perfect air barrier from living to attic space. ducts are in fur downs inside living space, vs air handler in side living space also. a bit over sized @ 2 tons with an whole house april air dehumidifier..also in conditined space. 4" media filter at return. 15SEER unit the addition I re-roofed last summer, batts in rafter bays 1" R-7 foil foam sheating lathes & gavalnized tin. as rafter/ceiling joists are 2x4 (yeah...go figure) the batts are R-13 with the R-7 gives me R-21 for this large bedroom. prior to this install the room was used for storage as it was too hot in summer & too cold in winter and too expensive & uncomfortable to use. with radiant barriers there always has to be a mininum of a 3/4" air space. thus the lathes on top of foil foam board under tin in addition & opposite install in big attic with rb under rafters with foil facing into attic space. gotta run...but wanted to share my take on things. oh here is the pic of the supply sealed from inside the house with hardcast tape. best of luck

Per 2010 IRC Code, Polyurethane spray foam insulation may be used in an UNVENTED roof per below: SECTION R806 ROOF VENTILATION R806.1 Ventilation required. Enclosed attics and enclosed rafter spaces formed where ceilings are applied directly to the underside of roof rafters shall have cross ventilation for each separate space by ventilating openings protected against the entrance of rain or snow. Ventilation openings shall have a least dimension of 1/16 inch (1.6 mm) minimum and 1/4 inch (6.4 mm) maximum. Ventilation openings having a dimension greater than 1/4 inch (6.4 mm) shall be provided with corrosion-resistant wire cloth screening, hardware cloth, or similar material with openings having a minimum dimension of 1/16 inch (1.6 mm) minimum and 1/4 inch (6.4 mm) maximum. Openings in roof framing members shall conform to the requirements of Section R802.7. R806.2 Minimum area. The total net free ventilating area shall not be less than 1/150 of the area of the space ventilated except that reduction of the total area to 1/300 is permitted provided that at least 50 percent and not more than 80 percent of the required ventilating area is provided by ventilators located in the upper portion of the space to be ventilated at least 3 feet (914 mm) above the eave or cornice vents with the balance of the required ventilation provided by eave or cornice vents. As an alternative, the net free cross-ventilation area may be reduced to 1/300 when a Class I or II vapor barrier is installed on the warm-in-winter side of the ceiling. R806.3 Vent and insulation clearance. Where eave or cornice vents are installed, insulation shall not block the free flow of air. A minimum of a 1-inch (25 mm) space shall be provided between the insulation and the roof sheathing and at the location of the vent. R806.4 Unvented attic assemblies. Unvented attic assemblies (spaces between the ceiling joists of the top story and the roof rafters) shall be permitted if all the following conditions are met: The unvented attic space is completely contained within the building thermal envelope. No interior vapor retarders (Class I or II) are installed on the ceiling side (attic floor) of the unvented attic assembly. Where wood shingles or shakes are used, a minimum 1/4 inch (6 mm) vented air space separates the shingles or shakes and the roofing underlayment above the structural sheathing. In climate zones 5 and 6 any air-impermeable (eg.- spray foam) insulation shall be a Class III vapor retarder, or shall have a vapor retarder coating or covering in direct contact with the underside of the insulation. Either Items 5.1, 5.2 or 5.3 shall be met, depending on the air permeability of the insulation directly under the structural roof sheathing. 5.1. Air-impermeable (eg. spray foam) insulation only. Insulation shall be applied in direct contact with the underside of the structural roof sheathing. 5.2. Air-permeable (eg. fiberglass or cellulose) insulation only. In addition to the air-permeable insulation installed directly below the structural sheathing, rigid board sheet insulation shall be installed directly above the structural roof sheathing as specified in Table R806.4 for condensation control. 5.3. Air-impermeable and air-permeable insulation. The air-impermeable (eg. spray foam) insulation shall be applied in direct contact with the underside of the structural roof sheathing as specified in Table R806.4 for condensation control. The air-permeable (eg. fiberglass or cellulose) insulation shall be installed directly under the air-impermeable (eg. spray foam) insulation. TABLE 806.4 INSULATION FOR CONDENSATION CONTROL CLIMATE ZONE MINIMUM RIGID BOARD OR AIR-IMPERMEABLE INSULATION R-VALUEa 4 R-15 5 R-20 6 R-25

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).