CES 2018: Philips Hue Announces iOS App Redesign, 'Hue Sync' for macOS, and Outdoor Bulbs Coming Mid 2018

[err404]

Just a little redesign can solve the problem, there isn't a need for a bridge other than wanting customers to be able to use Hue as its own standalone environment.

I think the Hub is more of a legacy issue. When Hue was originally released, zigbee was the right technology for the task. However just recently a home WiFi mesh network has become a more robust direction. We will never be able to get away from needing a bridge for a Hue environment that includes the current generation of bulbs, but it is possible that a next gen light could also include WiFi.
Ultimately the Hub is not a big deal. They are cheap in the starter packs and do provide a consistent base for maintaining schedules and configuration.

 

[Quu]

Really deep into the Hue ecosystem here, all of our lights are now Hue lights. I'm really looking forward to new hardware products from Philips especially if they have an answer to the Leaf panels which look so good.

 

[henrikhelmers]

I don't understand the complaints about the Hue Hub. As you expand into Hue, you can easily rack up 30-40 bulbs, some even hit the single-bridge limit and run dual installations. If all of these were on the wifi network it could cripple it. If every bulb were able to act as a bridge (which seems to be another common request) it would make the bulb hardware much more complex.

Instead I think Philips should make their bridge much more powerful (to handle more clients) and work with Apple to improve the Hue/HomeKit performance (as far as I know things like single-bulb reachability is not exposed through the API, but based on slow guesswork).

 

[itsmilo]

I just installed my Starter Pack which came with a free Echo on top this week. I love it but getting Echo to understand the scenaries for each and every room is really difficult for sole reason so i am mostly going back to the app on my phone unless it is about turning a light on and off.

Who gives a crap about this tiny hub which comes basically free with the starter set. Just hide it behind the shelf

 

[poppe]

I love my Hue system but the app is appalling. Here's hoping that the new version is a huge improvement.

Too true! It's so funny because my homekit app shows a more accurate depiction of my light status than my Hue app. I always have to double check if I'm crazy because the Hue app will always say lights are on that aren't on or off when they are off. Then if you are scrolling down or up it might just change the intensity of the bulb when you are just trying to scroll down to a different room/area. It's really awful.

I wouldn't use the Hue app at all, but since HomeKit seems to thing that a tungsten color is either pure yellow or a really ugly almost blueish white, I get stuck having to use the Hue app to get towards the "read" color they have.

 

[Laird Knox]

How about just fixing the lights to do simple things light not turn every light in the house on after a power outage or remember the last color temperature setting if you turn them off and on with a standard light switch? You know, useful things that make sense.

I understand the reason for the lights coming on but I agree. Out of sixty or so bulbs I only have one that I control with both the hub and the wall switch.
[doublepost=1515605871][/doublepost]

That's not really how it would work though. The bulbs talk to each other in a mesh network, there would be only one connection needed in a perfect world. And that would be a Bluetooth Low Energy connection to an Apple TV or iPad for Homekit or other hub/bridge for other ecosystems. You could actually make a "bridge bulb" if you think about it, and install that bulb as the closest to the Apple TV or whatever.

That's still a hub, you are just packaging it differently.
[doublepost=1515606160][/doublepost]

I just installed my Starter Pack which came with a free Echo on top this week. I love it but getting Echo to understand the scenaries for each and every room is really difficult for sole reason so i am mostly going back to the app on my phone unless it is about turning a light on and off.

Who gives a crap about this tiny hub which comes basically free with the starter set. Just hide it behind the shelf

Configuration can be a bear but once you have things sorted out it is really nice. I have a mix of Hue, Lightify, Mi-light, Lutron, GE, Kidde, Ecobee and Wink. When I replaced my AC last summer I added the Ecobee and it was only last week that I got the Echo working with it.

Hang in there and once you get it ironed out it you will like it even better. 🙂
[doublepost=1515606377][/doublepost]

I think the Hub is more of a legacy issue. When Hue was originally released, zigbee was the right technology for the task. However just recently a home WiFi mesh network has become a more robust direction. We will never be able to get away from needing a bridge for a Hue environment that includes the current generation of bulbs, but it is possible that a next gen light could also include WiFi.
Ultimately the Hub is not a big deal. They are cheap in the starter packs and do provide a consistent base for maintaining schedules and configuration.

I hope not - I don't need an extra 60 or 70 devices on my WIFI network. I would much rather manage them through a single access point. (Well, more like four access points in my case but that's just because I have several different technologies working.)

 

[Iconoclysm]

I think the Hub is more of a legacy issue. When Hue was originally released, zigbee was the right technology for the task. However just recently a home WiFi mesh network has become a more robust direction. We will never be able to get away from needing a bridge for a Hue environment that includes the current generation of bulbs, but it is possible that a next gen light could also include WiFi.
Ultimately the Hub is not a big deal. They are cheap in the starter packs and do provide a consistent base for maintaining schedules and configuration.

Zigbee is probably still the way to go but the hub itself isn't necessary if the bulbs can also bridge to bluetooth or whatever. Wifi mesh would use more power and I think there's a lot of confusion around how that works. The mesh part of that is a second wifi radio, so that increases cost and uses double the power of wifi versus just a fraction of the power. The hub is a big deal because it's clutter, and I know i have 5 hubs for different products which is just ridiculous. I try to keep my rack space clean in my house and these hubs are a mess.
[doublepost=1515606553][/doublepost]

I don't understand the complaints about the Hue Hub. As you expand into Hue, you can easily rack up 30-40 bulbs, some even hit the single-bridge limit and run dual installations. If all of these were on the wifi network it could cripple it. If every bulb were able to act as a bridge (which seems to be another common request) it would make the bulb hardware much more complex.

Instead I think Philips should make their bridge much more powerful (to handle more clients) and work with Apple to improve the Hue/HomeKit performance (as far as I know things like single-bulb reachability is not exposed through the API, but based on slow guesswork).

Only one bulb needs to be a bridge.
[doublepost=1515606748][/doublepost]

That's still a hub, you are just packaging it differently.

I don't know why you're stating this as a correction to my comment, because the entire point is to replace the hub with the same functionality. So, package it differently. The problem with the hub isn't that it performs a function of bridging zigbee to ethernet, the problem is that it's a wall wart, an ethernet port, and an ugly white block among many others in the same small space.

 

[mattyj2001]

As others mentioned, the big advantages of ZigBe over WiFi are cost, size, and power efficiency. I doesn't have anywhere near the bandwidth of WiFi, but you don' need that for sending home automation commands.

Add to that the way the bulbs build out a mesh network in your house, so your furthest away bulbs don't have to be within range of wifi, they just need to be within range of other builbs in the network. That's the only way my front porch light would ever work, wifi doesn't reach out there.

Also add to that the fact that you don't have to reprogram each of your bulbs if you ever change your network name, passcode, router, etc. The hub just makes things easier.

 

[MhzDoesMatter]

"Complaining about bridges is ridiculous" - same people lighting fires over "Dongle-gate"

 

[Laird Knox]

Just a little redesign can solve the problem, there isn't a need for a bridge other than wanting customers to be able to use Hue as its own standalone environment.

The bridge keeps all those devices off of your WIFI network. I don't need that many devices eating up IP addresses and resources on my network.

Sometimes I get the feeling that these companies do not have in house developers or that the turnover is so great that they can only keep the product updated instead of rethinking its design.

I expect that is part of it. Companies can do odd things with priorities. Until recently most didn't consider smart lights a priority even though they produce products.

However, I think the bigger issue is the behavior is more of a legacy design that made sense when smart lights first started to hit the scene. At that time your typical user had only a few devices and there wasn't Siri and Alexa to control things. If you walked into a room you would have to pull out your phone or go to the computer to turn on the light. This is why they were designed to turn on after a power failure. You could then flip the light switch off/on and the light would come on.

Now that it market is becoming more mainstream and people are using it as more than a curiosity that behavior becomes a liability. It will take a fundamental shift in the market's use habits before any meaningful changes are made. This will be great going forward but no help for the legacy devices.

Things have started to mature and we will see even better products in the future. For now we are stuck with some of the early design choices.
[doublepost=1515607645][/doublepost]

I don't know why you're stating this as a correction to my comment, because the entire point is to replace the hub with the same functionality. So, package it differently. The problem with the hub isn't that it performs a function of bridging zigbee to ethernet, the problem is that it's a wall wart, an ethernet port, and an ugly white block among many others in the same small space.

I may have grabbed the wrong quote. Lots of threads going here and I'm fighting the flu. 😉

I was trying to respond to the general "bridges are stoopid" replies. It sounded like you were advocating getting rid of bridges by adding a bridge to a bulb. 🙂

 

[BrentD]

I, too, get pissed off after a power blip when all my Hue lights come on, but....

What do all of you propose they do about that and/or having the light come back on in its last state when you flip a switch? I mean, think about it. There are really only two options that could make that happen.

1) A battery in each bulb to remember the last state when mains power was removed so it can be reapplied when power returns, whether via switch or power restoration. This is a non starter because there's no way to design for this. You might have a power outage or turn the light off for 45 minutes or for 3 days. They can't implement a battery that accounts for all of those variables.

2) Some of you say the hub should remember the last state of a light and restore it when power is reapplied. The only way that would work is to design the bulbs so they don't turn on UNTIL they reconnect with the hub once power comes back on or the switch is flipped. They might be able to get the time for that reconnection to the hub down to 5, MAYBE 3 seconds. How much more pissed off would you be when you flipped a light on if it took 3-5+ seconds for anything to happen? Yes, it would solve the issue of a power outage, but it would compound the issue of flipping lights on and off with the switch.

 

[thisisnotmyname]

I understand the reason for the lights coming on but I agree. Out of sixty or so bulbs I only have one that I control with both the hub and the wall switch.

There's just no excuse for it not being something that can be controlled via non-volatile RAM on the device. The whole "what if there's an emergency" thing is such a cop-out, emergencies often involve a lack of power which is why first responders carry their own light sources. Default could be that they automatically come on via cycling of power but those of us who have smart switches should be able to permanently disable that "feature".

In any situation the device should be smart enough to remember the last light-temperature though.

Anyway, I know I'm preaching to the choir. It's incredibly frustrating that so many of us have complained about this for literally years and Philips refuses to address it. LIFX seems to be coming on strong, if they build a strong enough reputation I could be convinced to dump all of my Hue and change. Anyone looking at Hue right now I'd encourage you to wait for a bit and see how this plays out if your needs aren't urgent.

PS - for those of you unfamiliar with the issue, imagine a power outage happens at 2:00 AM, even a momentary one (not that uncommon). When the power comes back on Hue sees this as being the same thing as turning the lights off then back on again and turns on EVERY HUE LIGHT IN YOUR HOUSE AT MAXIMUM BRIGHTNESS. Yeah, that's fun.
[doublepost=1515608835][/doublepost]

I, too, get pissed off after a power blip when all my Hue lights come on, but....

What do all of you propose they do about that and/or having the light come back on in its last state when you flip a switch? I mean, think about it. There are really only two options that could make that happen.

1) A battery in each bulb to remember the last state when mains power was removed so it can be reapplied when power returns, whether via switch or power restoration. This is a non starter because there's no way to design for this. You might have a power outage or turn the light off for 45 minutes or for 3 days. They can't implement a battery that accounts for all of those variables.

2) Some of you say the hub should remember the last state of a light and restore it when power is reapplied. The only way that would work is to design the bulbs so they don't turn on UNTIL they reconnect with the hub once power comes back on or the switch is flipped. They might be able to get the time for that reconnection to the hub down to 5, MAYBE 3 seconds. How much more pissed off would you be when you flipped a light on if it took 3-5+ seconds for anything to happen? Yes, it would solve the issue of a power outage, but it would compound the issue of flipping lights on and off with the switch.

nonvolatile memory. You don't lose all of the data stored on your phone when your battery dies either.

 

[BuddyRich]

Just a little redesign can solve the problem, there isn't a need for a bridge other than wanting customers to be able to use Hue as its own standalone environment.
[doublepost=1515600365][/doublepost]

Sometimes I get the feeling that these companies do not have in house developers or that the turnover is so great that they can only keep the product updated instead of rethinking its design.

Ummm... while that is a downside a big upside is having the light come on when you flip the switch (or power is restored).

its a design (if not safety) choice to have the light come on at full brightness in a warm white when it is turned on. its probably the only reason why i got WAF for it with a wife that justs wants it to work via the switch and could care less about an app.

 

[MrX8503]

I still have to laugh about people complaining about the need for a bridge. Sorry, I'd rather have one connection to my network from Hue, rather than 11 (one for each of my bulbs).

I used to be in the bridge less camp, but then I saw how many issues people were having with bridge less setups. I'm willing to bet the bridge is needed for a rock solid connection.

 

[nuckinfutz]

I fell in love with Hue bulbs a few years ago but they have stagnated and the competition has more than caught up. When LIFX became Homekit compatible I tried them out because they offer higher lumen bulbs AND no hub. Unfortunately, a bit too late for me as I have a house full of Hue but I won't be buying anymore unless they come out with a 90-100w equivalent flood. I do only gift LIFX now.

To me Hue's two weaknesses, which don't seem to be fixed here are ditching the need for a hub and brighter bulbs for home's with recessed lighting.

Hue's weakness is not having an in-wall decora light switch that provides constant power to the circuit.

Hue uses Zigbee which is a mesh network. That is not a hub that comes with it it is a bridge that connects wifi networks to Zigbee networks. Hub and Bridges are vastly different networking devices.

Lifx doesn't have a snowballs chance of supplanting hue. Hue Entertainment will be huge and because it's based on standards (Zigbee) it's going to interoperate with many more products.

If your home needs a 100 watt floodlight you're doing your lighting wrong.
[doublepost=1515609487][/doublepost]

FWIW I'd like for Hue's bridge strategy to be the best technically, because I just recently bought into the ecosystem

It is the best strategy which is why Lutron deploys the same bridge strategy. Bridges allow for multi-protocol support and the saving of lighting data (scenes, schema etc) without requiring a cloud backend. They also make remote access easy as well.

Without a bridge you have

1. Cloud requirements (See TP-Link bulbs for example)
2. Bluetooth or Wifi ..of which neither have the range of Zigbee or Zwave in the home and require more power.
3. Slower support for existing or new protocols. Bridge based devices get faster access to homekit, smarthings etc.

 

[44267547]

I love my Hue system but the app is appalling. Here's hoping that the new version is a huge improvement.

Definitely need improvement on the application part of it. The product is great, but if the app is troublesome, it doesn't make the process any easier of enjoying the product.

 

[Laird Knox]

There's just no excuse for it not being something that can be controlled via non-volatile RAM on the device. The whole "what if there's an emergency" thing is such a cop-out, emergencies often involve a lack of power which is why first responders carry their own light sources. Default could be that they automatically come on via cycling of power but those of us who have smart switches should be able to permanently disable that "feature".

In any situation the device should be smart enough to remember the last light-temperature though.

Anyway, I know I'm preaching to the choir. It's incredibly frustrating that so many of us have complained about this for literally years and Philips refuses to address it. LIFX seems to be coming on strong, if they build a strong enough reputation I could be convinced to dump all of my Hue and change. Anyone looking at Hue right now I'd encourage you to wait for a bit and see how this plays out if your needs aren't urgent.

PS - for those of you unfamiliar with the issue, imagine a power outage happens at 2:00 AM, even a momentary one (not that uncommon). When the power comes back on Hue sees this as being the same thing as turning the lights off then back on again and turns on EVERY HUE LIGHT IN YOUR HOUSE AT MAXIMUM BRIGHTNESS. Yeah, that's fun.
[doublepost=1515608835][/doublepost]

It isn't just Hue. That's pretty much the standard for smart bulbs. I only have five or six Hue bulbs - the other 50+ also come on full after a power interruption.

As I said in another post I don't see that behavior changing until more people adopt the technology on a larger scale. One or two bulbs? Not a problem. A full house of bulbs? Are you kidding me?! The industry is starting to mature from first adopter status to mainstream product but it isn't quite there yet.

I deal with similar problems at the office all the time. I explain that the "quick" solution is a bad idea because of the long term implications. My suggestions are ignored and two years later everybody is in a panic because the quick hack is causing all sorts of problems. It gets even more fun when you are dealing with casino products that handle monetary transactions.

For now I've resigned myself to the fact that the lights do not behave like I would prefer. The benefits outweigh that particular problem.

Some day they'll get it all worked out but I'll still have all my legacy bulbs to deal with. 😛

 

[err404]

Zigbee is probably still the way to go but the hub itself isn't necessary if the bulbs can also bridge to bluetooth or whatever. Wifi mesh would use more power and I think there's a lot of confusion around how that works. The mesh part of that is a second wifi radio, so that increases cost and uses double the power of wifi versus just a fraction of the power. The hub is a big deal because it's clutter, and I know i have 5 hubs for different products which is just ridiculous. I try to keep my rack space clean in my house and these hubs are a mess.

You need a long range, low power, mesh protocol for the lights. Zigbee fits the requirements well, but you need a second protocol to connect your phone. WiFi is the only real option since it is common to all devices in your home. Without a Hub, BT would break too many other control options. From speakers to web interfaces.
I get the frustration with Hubs. I just pulled out a Sonos Hub and cell repeater from my network last month.
Frankly the solution may be to keep the dedicated Hub, but rev the Hub to be wireless to your WiFi. This way you don’t consume the port and can put it wherever you want.

 

[Laird Knox]

If your home needs a 100 watt floodlight you're doing your lighting wrong.

Oh?

I have 100 watt equivalent task lights in my kitchen (A19, not flood) that are controlled by a Lutron dimmer. Most of the time they are set to around 30% but if I need the light they bring it. Not sure what is wrong with that.

I have motion activated floodlights on my driveway (Ring Floodlight Cam). I don't know what the equivalent wattage on them is but I want better than a 60 watt light in that capacity.
[doublepost=1515612883][/doublepost]

Frankly the solution may be to keep the dedicated Hub, but rev the Hub to be wireless to your WiFi. This way you don’t consume the port and can put it wherever you want.

Some manufacturers do that.

I can (kind of) see the issue of saving costs by only using Ethernet. Then again, I have an $8000 laser cutter that only works over WIFI and has no Ethernet. I can't imagine the cost savings was that much in a device that expensive. Manufacturers can get odd ideas. 😉

 

[H3LL5P4WN]

That's not really how it would work though. The bulbs talk to each other in a mesh network, there would be only one connection needed in a perfect world. And that would be a Bluetooth Low Energy connection to an Apple TV or iPad for Homekit or other hub/bridge for other ecosystems. You could actually make a "bridge bulb" if you think about it, and install that bulb as the closest to the Apple TV or whatever.

To be clear, I'm not complaining about the bridge, but all of these bridges are a pain in the ass if you really go all in on automation. Eventually, they need to be eliminated.
[doublepost=1515600017][/doublepost]

Most of the bulbs don't talk directly to the bridge, they talk to each other and the bridge in a big mesh which gives them much better range around the house.
[doublepost=1515600086][/doublepost]

And if it were wifi, embed that bridge in a bulb. But really, they just need a "homekit" version that relies on Bluetooth as the bridge.

Zigbee isn't cheap though, there are licensing costs that I doubt exist with wifi.

Not a shabby idea. At the current time, literally the only bridge I have is for Hue, so I can't complain.

 

[tzm41]

I like Hue, and I like the idea of game/music synced lighting.

 

[henrikhelmers]

[doublepost=1515606553][/doublepost]
Only one bulb needs to be a bridge.

This bulb would then need to be sold in all the configurations that customers want. And have a reachable button, or other means of authentication.

In addition you risk that someone turns off this special bulb, or even worse— if someone unfamiliar with Hue replaces it with a standard bulb.

I bet Philips gave this some serious thought when they decided to go for the separate bridge.

 

[Iconoclysm]

This bulb would then need to be sold in all the configurations that customers want. And have a reachable button, or other means of authentication.

In addition you risk that someone turns off this special bulb, or even worse— if someone unfamiliar with Hue replaces it with a standard bulb.

I bet Philips gave this some serious thought when they decided to go for the separate bridge.

I bet they're still giving this thought after the bridge. At the time the bridge came along, there was only one bulb and one configuration anyway. Also, it doesn't take much to think this through further, the device could even just be a bulb extension or in the very, very least, be a wireless plug like what Logitech uses for the pop buttons.
[doublepost=1515614843][/doublepost]

You need a long range, low power, mesh protocol for the lights. Zigbee fits the requirements well, but you need a second protocol to connect your phone. WiFi is the only real option since it is common to all devices in your home. Without a Hub, BT would break too many other control options. From speakers to web interfaces.
I get the frustration with Hubs. I just pulled out a Sonos Hub and cell repeater from my network last month.
Frankly the solution may be to keep the dedicated Hub, but rev the Hub to be wireless to your WiFi. This way you don’t consume the port and can put it wherever you want.

Yes, the second protocol would be BLE to the Apple TV. Then Apple TV to iCloud...then your phone or whatever other application. Hue is trying to have its own ecosystem as a stand alone, which is fine. But that Hue Hub should be an option not a requirement. The wireless option has been done, it's still not ideal but it's definitely better than more wires - you lose the ethernet cable and you even lose the AC adapter cable if it's just a stand alone wall plug device like Logitech's.
[doublepost=1515615024][/doublepost]

Ummm... while that is a downside a big upside is having the light come on when you flip the switch (or power is restored).

its a design (if not safety) choice to have the light come on at full brightness in a warm white when it is turned on. its probably the only reason why i got WAF for it with a wife that justs wants it to work via the switch and could care less about an app.

I know that's why but it should be optional. In my case, I replaced the light switches with the Hue button things, so it's not necessary. I'm sure everything that everyone's asking for is going to happen one day but Hue development is so slow that they could potentially lose their market to an upstart.

 

[araadt]

How about just fixing the lights to do simple things light not turn every light in the house on after a power outage or remember the last color temperature setting if you turn them off and on with a standard light switch? You know, useful things that make sense.

I live in an apartment. The way this is set up means that if my landlord sends someone in to do maintenance or a repair they can turn on the lights using the switch installed in my wall in 1955 (because flipping the switch on and off resembles a 2am power cut to the bulb—or maybe it's the other way around) instead of me having to add Joe/Jane Handyman to my HomeKit network. For all intents and purposes Joe Handyman doesn't even know I have fancy bulbs, etc.

 

[Iconoclysm]

The bridge keeps all those devices off of your WIFI network. I don't need that many devices eating up IP addresses and resources on my network.


I expect that is part of it. Companies can do odd things with priorities. Until recently most didn't consider smart lights a priority even though they produce products.

However, I think the bigger issue is the behavior is more of a legacy design that made sense when smart lights first started to hit the scene. At that time your typical user had only a few devices and there wasn't Siri and Alexa to control things. If you walked into a room you would have to pull out your phone or go to the computer to turn on the light. This is why they were designed to turn on after a power failure. You could then flip the light switch off/on and the light would come on.

Now that it market is becoming more mainstream and people are using it as more than a curiosity that behavior becomes a liability. It will take a fundamental shift in the market's use habits before any meaningful changes are made. This will be great going forward but no help for the legacy devices.

Things have started to mature and we will see even better products in the future. For now we are stuck with some of the early design choices.
[doublepost=1515607645][/doublepost]
I may have grabbed the wrong quote. Lots of threads going here and I'm fighting the flu. 😉

I was trying to respond to the general "bridges are stoopid" replies. It sounded like you were advocating getting rid of bridges by adding a bridge to a bulb. 🙂

These devices should continue to use Zigbee or, if possible, BLE 5.0. WiFi has too much overhead for these bulbs, and it's not just the IP addresses but also the number of simultaneous connections that can slow everything down. I definitely believe this needs to be bridged, it's just a matter of how it's bridged.

I agree with all of your assumptions it's just been half a decade without revisiting or improving the underlying architecture and people are getting restless.
[doublepost=1515615354][/doublepost]

I live in an apartment. The way this is set up means that if my landlord sends someone in to do maintenance or a repair they can turn on the lights using the switch installed in my wall in 1955 (because flipping the switch on and off resembles a 2am power cut to the bulb—or maybe it's the other way around) instead of me having to add Joe/Jane Handyman to my HomeKit network. For all intents and purposes Joe Handyman doesn't even know I have fancy bulbs, etc.

That's why there are so many new Homekit compatible switches and sensors. At this point, these should be options - advanced options, but options. Even if I have to script a JSON file and upload it to the hub or something, just make it possible.
[doublepost=1515615666][/doublepost]

I, too, get pissed off after a power blip when all my Hue lights come on, but....

What do all of you propose they do about that and/or having the light come back on in its last state when you flip a switch? I mean, think about it. There are really only two options that could make that happen.

1) A battery in each bulb to remember the last state when mains power was removed so it can be reapplied when power returns, whether via switch or power restoration. This is a non starter because there's no way to design for this. You might have a power outage or turn the light off for 45 minutes or for 3 days. They can't implement a battery that accounts for all of those variables.

2) Some of you say the hub should remember the last state of a light and restore it when power is reapplied. The only way that would work is to design the bulbs so they don't turn on UNTIL they reconnect with the hub once power comes back on or the switch is flipped. They might be able to get the time for that reconnection to the hub down to 5, MAYBE 3 seconds. How much more pissed off would you be when you flipped a light on if it took 3-5+ seconds for anything to happen? Yes, it would solve the issue of a power outage, but it would compound the issue of flipping lights on and off with the switch.

3) The hub should remember the last save state, as in option 2. The lights can all come back on warm like they do until they sync up and return to the original state with some intelligence to tell if it was truly an outage or just the switch being flipped (did ALL of the lights go out? Are they all in the same "room"? Did the hub also lose power? etc). And finally, make this an optional behavior choice.