And how much solar power do they generate when covered by snow/ice? Your objection is short sighted...
My objection is about it taking more power to keep them clear than they could generate.
If they generate 48 watts per panel, but are drawing 150 watts to run the heating elements, they're losing 102 watts the whole time the heating elements are on.
Maybe they have figured out some way to require far less power per square foot to melt snow/ice on a flat surface than the roof heating systems I looked at for reference, but they'd have to be down below 10 watts per square foot to break even under ideal conditions. That is not much heat at all, and as others have pointed out in the sort of conditions where you'd need the heaters running the weather tends to not be anywhere close to ideal for solar so the chances you'd even get 48 watts are slim.
According to all the articles and press releases power generation is the primary purpose of these panels. They claim they'll have enough surplus to offset the energy usage of the entire town square. If they are consuming more power in an hour than they could generate in three, just to keep them able to generate power, that doesn't make a bit of sense.
Now if they were hyping this as an interactive LED sidewalk that's heated to stay clear on its own in winter, and it also happens to generate some solar power, that'd be an entirely different thing. That's not what they're doing though.
just how much snow and ice melting does it take to turn these into a net negative rather than positive generator of energy?
My thoughts exactly. This installation has 30 tiles over 150 square feet, so five square feet per tile, with each tile generating 48 watts total under ideal conditions. Let's be nice and round it to 10 watts per square foot.
Looking at a variety of heated driveway and heated roof systems it seems that most use somewhere between 30 and 60 watts per square foot to effectively combat snow and ice. That's 3-6 times the best-case power generation of these panels.
Your information is years out of date. I've been using an ssh application on an ereader and I've been getting around half a second refresh. There's also a debian distro for the pocket Kobo from maybe three years back that has an on screen clock that updates in seconds - so less than one second refresh there.
Router/switch activity lights blink at a rate I'm not entirely sure of but definitely exceeds 10Hz. 1-2Hz is not enough to be useful for the purpose IMO.
I have a Kindle Paperwhite, I know how quickly modern displays can refresh. I actually want to build a thermostat that uses an e-ink display because it makes perfect sense in that role. but for a network device's status indicator it's no good.
Even cheap electronic paper can be updated once per second with fairly low power requirements. For activity, the lights have basically been useless for decades: unless you're the only one on the network and are sending pings one per second, they're basically always on. It would be far more use to have a few more pixels and display a logarithmic scale bar of total throughput. For power on, something that alternated between - and | once per second would let you know that there was power flowing, without needing a static light.
I'm looking at the gigabit Cisco switch on the desk next to me and definitely have to disagree with you there. I can clearly see the difference in activity between for example the port my VoIP phone is on and the ports my server and router are on. I can see how heavy the broadcast traffic is based on how often all ports blink simultaneously. I don't know what their actual blink rate is but I can say for sure it's greater than 10Hz on a highly active port. Many times over the years I've used the lights to help locate the source of a network loop or broadcast storm. The fact that the lights can blink rapidly is the key to that working.
A LCD might be able to go fast enough, I'm not sure.
The utilization indicator definitely could work though, I won't deny that.
Make it e-paper, not LCD, then it will be readable under any light. If e-paper displays are cheap enough to put on store shelves as price tags, then they should be cheap enough to serve as a status display on a router.
E-paper would be a terrible display for this purpose. It can't change fast enough to work as an activity light, and since it maintains an image effectively forever until updated it's not trustworthy for lower rate status monitoring like power on. If the device crashed or even powered off entirely without resetting the display first it'd look normal at a glance.
Tuning adapters suck.
Tuning adapters suck for the same reason CableCard as a whole kinda sucked. Because the cable industry as a whole wanted them to suck. Ever notice how their own boxes never had the same problems, even during the time they were forced to use the same CableCard interfaces? Or how variable the support was between providers, with some providers happily shipping cards to consumers and offering self-service interfaces to activate them where others would insist on a truck roll and scheduled appointment (with standard cable company timing)?
Look at the same concept as implemented in Europe. Over-the-air, cable, and satellite television all use variants of the DVB standard. It even has an IPTV variant, though I'm not sure how widely it's deployed in that context. There's a standard interface for a service provider's encryption solution. Any consumer can use any compatible device with any television provider, and it works great.
For whatever reason (read: doesn't benefit the right companies) in this country we have a history of looking at problems Europe's already solved and saying "nope, we can make something much worse for consumers". See also GSM vs. CDMA and the fact that Verizon still insists to this day that they need to individually certify each device while the majority of cell carriers on the planet happily work all day with whatever phones happen to be compatible.
It's easy to be compatible if you want to be compatible. What these companies try to avoid saying outright is that they don't want to be compatible.
Are there a lot of cell towers in these areas where cell service for internet is a viable option?
I have 250/25 cable and theoretically 24/2 DSL (really 14/1.5) at my house. A friend of mine two miles away has no cable and theoretically 6/1 DSL that really delivers about 3/256k most days. The same T-Mobile tower covers both of our houses, off which my old Note 4 gets 65/30.
A computer without COBOL and Fortran is like a piece of chocolate cake without ketchup and mustard.