Hmmm. Don't agree with this.
1. Spider Oak has built its business on zero-knowledge (Full Disc: not an employee or a fanboy, but a user. Like it, except for non-zero-knowledge on mobile/web)
2. There _is_ research going on about ways to compute on data without knowing the contents of the data. It's entirely likely that someone will solve search on zero-knowledge encrypted data, even though you and I don't yet know how it might work. (one way that comes to mind: zero-knowledge encrypt the query, then bounce the encrypted query against the encrypted ciphertext. This would probably suck 'cause it'd require ECB mode or something similar, and that's pretty weak, and such a technique would leak information like a sieve, but OTOH, not _impossible_ right out of the gate).
+1 for SpiderOak, but please know that their Android client is not Zero-Knowledge. It means that mobile use is...not quite as clean as one would like.
I think you're right to be concerned about the rainforests due to what is already an increasing demand for palm oil.
However, I put the blame on business looking for monoculture farming, and a generally unsustainable US consumer culture. It's not a secret that Americans have stretched resources to and past the breaking point; that we have demanded everything be constantly available, and cheaper every year. It should be obvious to anyone with basic arithmetic skill that that cannot continue indefinitely.
I realize that regulation is now a dirty word, but that is, in fact what is needed. I realize that the international scope of the problem will make that difficult, but the scale of the problem, the size of the disaster looming ought to make it a priority.
I'm sure someone will weigh in, pointing out that shareholder value demands frosting in a can, at the expense of our global carbon sink. Please. Go ahead and make that point.
Commodore made the Colt (I think) PC (as in "IBM Compatible PC") during the same era as the Amiga 2000, so it's valid to question what's being talked about here, IMO.
"You look like you're trying to drop a steamer! Would you like me to help?"
Good point. Making the RPi talk RS232 to the ancient WattsUp took only a level shift and decidedly retro baud rate (the summary said "outdated"). (FYI, there's now a "net connected" WattsUp, but they want you to go through their proprietary portal: NFW! Hunk of junk!)
I probably should've mentioned that I initially used Google Analytics to build the graphs. It's really not designed for that.
The duct tape-like approach I took with hardware I had sitting around to bend it to my will (vs. buying something off the shelf - not to mention final fit and finish) make it feel like a hack to me.
But OK, point made. I DQ myself.
YOU'RE NOT MY REAL FATHER!!!
Well, both, really.
I guess I should say accuracy, or repeatability or noise floor of the humidity sensor - since I originally thought I was seeing some capacitive artifact. It actually does a pretty good job.
Put it all together for near-real-time track of how much it costs to keep my basement at a given humidity.
The Raspberry Pi caches readings in a local database in case it can't connect to the web, then stores in a database on my web server. The database ingestion also keeps a 2-hour running average to smooth things out a bit.
When I set it up, I thought it wasn't working right - I saw sawtooth-like patterns in the humidity data. Turned out, it was working perfectly: the resolution of the humidity sensor was good enough that I could watch the humidity in the room rise until the dehumidifier kicked on!
The Terrorists Win!!!!!
...all those Goddamned batteries...
IMEI. Definitively identify a device, without even requiring an account sign-up. There's a bunch more.
It's important to understand that the context of use for a mobile web page is different from that of a successful mobile app.
There are a bunch of dumb apps developed for online news sites - as though I'd ever want to go to just one news site, vs. have the mobile web spread out before me.
OTOH, good mobile apps do things that the mobile web doesn't or can't. Perform read/write operations on local data. Use local processing power - as much as is available. Access local sensors not available to mobile web. Aggregate data from multiple sources - perhaps blending web and local data.
Even apps that do nothing more than provide deep search - if the vertical market for the app is well defined (Movie geeks: IMDB, small investors: finance apps...), it's possible for a mobile app to excel over a web site by providing native gadgets and a platform native UI that doesn't have to leave room for the (admittedly minimal) mobile browser UI.
Finally, mobile apps can scrape a bunch more information from the user's device than can mobile web. Definitely a help in monetizing a popular vertical, if you roll that way.
Disagree with your comments on CG UI.
I used Lightwave 3D for quite a while. It's built with one of these cross-platform Frankenstein UIs, and frankly, that part sucks.
I want an app to behave on a Mac like a Mac app, on a Windows PC like a Windows app, and on a Linux box with whatever feeble attempt at consistency it can muster.
Ditto goes for iOS and Android. Can't stand seeing the iOS "BACK" button on an Android screen. Dumb, dumb, dumb.