People are always looking for location aware services for their mobile devices, including local info. But frankly the UIs on small mobile devices often are poor. When you are on a cell phone, voice to a smart person is the interface you often want.

So here’s a possible location aware service. Let people register as a “local expert” for various coordinates. That’s probably folks who live in a neighbourhood or know it very well. They would then, using a presence system on their own phone or computer, declare when they are available to take calls about that location.

Somebody sitting with a cell phone in a location could call a special 900-like number. Their phone could just transmit their location, or they would quickly say it to a human for entry. Then, their call would be routed to a local expert who is marked as available for calls. (In some cases it may simultaneously ring several experts of possible but unsure availability and give the call to whoever answers first.)

Then they could, for a fee (perhaps $1/minute?) ask the expert questions.

• “Where’s the best Thai food?”
• “How do I get transit to such and such location?”
• “What’s a good Taxi company to call? Can you call me one?”
• “Is there a shop around here that sells widgets?”
• “Is this museum worth it?”
• “What parts of the area are dangerous?”
• “How much is real estate here?”

The expert would be expected to know how to answer questions about most of the restaurants, bars and shops. And they could also — so long as they disclosed any kickbacks very clearly — provide coupon codes to people that would rebate the cost of the call.

At the end of any call, the caller would stay on the line and be asked to rate the quality of the expert. They could also rate later. Experts would gain reputations for their skill, and the ones with the highest ratings would be given more calls, or be able to charge more.

Charging could be per minute, fixed-rate, or as noted, rebated with validation from a recommended merchant (though I would want to design a system so that advice is never biased by this.)

This could also be done by texting, which would be easier for experts to do, and probably be cheaper, but of course is slower for the mobile user. Many mobile users are getting pretty good at their texting. The experts would presumably be at computers with IM clients, but they could be at mobile phones as well.

To make this cheaper, one could arrange for trading minutes. Which is to say, if you put minutes into the system advising others, you can in turn use minutes getting advice when you need it. Some people might prefer to do this in a friendly way rather than charge or pay.

Experts could very well be just around the corner, physically, if they are being an expert on their local neighbourhood. It’s not out of the question they could then agree to help in person. In this case you would need to have some way to certify they’re not up to something nefarious. The fact that the call is logged and you know the home address of the expert in the database should be enough. The client might be up to something nefarious, but this seems a pretty low risk.

Just on the heels of my prior post on the bad math often found around alternative energy, I see a Google Blog post on Google’s solar installation. It claims Google with save money with their 1.6 megawatt solar installation.

I would be very interested to see Google’s numbers — what are they paying for this PV system, and what do they pay the power company for their grid power? Did they get rebates on the PV install? Rebates can help a single customer save money but they do it at taxpayer expense which makes it a wash, other than as a means to try to increase the market for solar and bring down the price.

Now, I’m not in any way saying that it’s bad for Google to go solar. Large grid-tie solar arrays are quite green, with minimal emissions (only those from their manufacture, shipping and install) and so it’s good to have them, even if they are more expensive than non-green grid energy.

But I want to know, is my math bad, or is Google’s? If companies can really save money with a PV array they should be springing up like weeds.

Today I also read announcements of companies hoping to bring to market new solar panel technologies with thin films that are vastly cheaper than existing tech. When that happens, the panels really should sprout everwhere, and to very positive effect.

Update: The press releases say the system is 1.6MW, and provides 2.6 million khw/year for a saving of $393K per year (about 15 cents/kwh which is about right in California.) The press release also says the system will pay for itself in 7.5 years, which at 7% interest rate means its total cost was $2.2M. (Truth is Google is able to make far better than 7% with its money, I suspect.)

This means an astounding $1.38 per watt for installed solar. I’ve never heard of anything remotely like this. Even with a bad-math 0% interest rate, 7.5 year payoff is $1.84/watt so it’s not just bad math here. Even with the California rebates of $2.60/watt and 30% federal tax credit, it’s still amzingly cheap — and almost all the savings are coming from the taxpayer.

The release also suggests that 393K per year will result in 15 million saved over a 30 year lifespan. I can’t figure the math in this number. The bad-math 30*393 is under 12 million. The real saving over 30 years at 7% interest has a present value of 4.8 million. The future value, in 30 years time, of $393/year is well over 30M at 7%. You need an interest rate of 1.5% to have a FV near $15M. I suspose the risk-free-rate-above-inflation might correspond to this but it’s not typical in expressing these numbers.

So what are the real numbers?