Take a good look at BitcoinSpinner. It has exactly what you are asking for, including easy backup/restore using QR codes. The app has the private key, server does the heavy lifting of managing the block chain.

It may be a great mobile app, but this is certainly NOT what I'm talking about because it needs your server to operate.

I think what is needed is something that provides the best of both worlds…easy to use and by default uses a service for the back end interface with the p2p network…but also having the option that people can easily setup their own back end and use it without needing the service.

on a side note, the site's missing G+ button. and also stats how many tweets, FB liked and +1'ed so we can see some relative numbers right on the site )

Why do you say mobile clients shouldn't contain anything about the block chain? That's not apparent to me at all. I'd suggest just letting clients develop and see what works out. It boils down to a tradeoff between server availability/fee charging and performance, but I think in the long run the two approaches we see today will converge.

@DeepBit:

It is not reasonable for a normal user to store private keys. They will get hacked or fail at backup diligence, and lose their money either way. I strongly recommend that we do not encourage what will inevitably end in disaster. Sure, fraud in banks will appear, but it will be less of an issue than with fiat money, since nobody can bend the truth as much by inventing numbers on paper.



@Topic:

Haha, I love how "Computer Skills" goes up to "Command line hacker". It's exactly what I see everywhere, I love the irony of that table. (Placing Visual C below is strange though. Okay, VS really sucks for C, but the debugger alone makes the order look like irrational console/emacs/vi/whatever fandom.) On a more serious note, I don't think it's right to assume perfect correlation between financial and technical understanding. I know people whom I would place "off the chart" in one of "knowledge of currencies" or "computer knowledge", but rather low in the other.

Also, I believe good design should never have to care about assumptions on the users' low-level tech skills, such as C hacking or webdesign. Ending up at such a question is a strong indicator that something went wrong. (And I don't like the idea of non-verifiable code when there's finance involved, but I know I'm fighting a lost cause.)

If I may, let me propose what groups I believe are around, why, and how big their numbers are if advertisement is done right:

• Casual user -- 10 M or more thinkable. (Yes, I think 1M is conservative)
  
  • Wants a small amount of Bitcoins for an arbitrary reason, and does not want to invest a lot of effort.
  • No assumptions on platform or technical skills are reasonable. Needs web interface to Bitcoin bank or similar, and to be guided in terms of security.
  • "A friend told me about the new bit coin..."
• Geek -- 100 000 or so people readily available as market
  
  • Likes Bitcoin just because it's Bitcoin.
  • Mostly wants to store Bitcoins himself. May get bored if there aren't enough ways to spend or earn some.
  • This group is already strongly represented, although many more would join in if given more incentive.
  • "Whoa, 1337 crypto, Satoshi-sama!"
• Small financial speculator -- 10 000 people maybe?
  
  • Is interested in Bitcoin because of the massive potential upside of new, disruptive technology.
  • Safe to assume knowledge on financial instruments.
  • Not safe to assume any technical knowledge.
  • Expects to be able to move five- to six-digit USD sums in or out of Bitcoin without trouble.
  • "What's the spread on call options at 3 USD/BTC?"
• Investors and project leaders -- little known about who and how many might be there.
  
  • Is an expert on at least one related topic.
  • Has an idea that needs support, either in terms of infrastructure or funding.
  • Is capable of coping with some issues, but will drop the project if problems pile up.
  • "How is the legal situation of Bitcoin in Norway?"

This kind of classification removes the coupling between financial interest and IT knowledge and instead sorts by intentions and existing groups/stereotypes. I find it easier and more efficient to classify people this way.



Now, what I would like everybody to keep in mind: the important groups are casual users and investors! Geeks and speculators is what we currently have, and these are easy to draw in anyways. But the real power lies in building a bridge between the ideas of the investors and the needs of a casual newcomer! This means it's very important to make direct Bitcoin usage available to even the casual users. We need a comfortable path between projects and their potential customers.

That said, thanks for doing this! I hope you will have a lot of users!

Edit: added some kinky example sentences for each stereotype.

I think there are some assumptions in here which can be invalidated by technical progress:

• Nothing stops an app from catching up with the chain at nighttime, when it's plugged in to the wall for charging. When you open it during the day it only has to catch up on (probably) around 100 blocks, which is fast.
• In future Bitcoin nodes can be asked to filter blocks before sending them, ie, only sending CMerkleTx structures given a set of interesting script templates + possibly bloom filters. This isn't exactly a difficult protocol addition, but it does involve some coding and then waiting for people to upgrade. I described this protocol months ago on this forum.
• Currently processing the block chain on a mobile is slow because BitCoinJ (the only codebase I know which does this) does some things quite inefficiently. It can certainly be sped up a lot without protocol additions.

That said, there's no denying that with the state of various codebases today, what Jan says is true.

Wow the website is improving every time i refresh it, nice job!

It's not reasonable for a normal user to store cash. They can lost it, invest it, be robbed, and so on. Let them use credit cards and paypal instead, right - someone should be in control of user's money :)

Mybitcoin, btcex, bitomat, mtgox - sure, they aren't like "normal users", so they can't get hacked or fail at backup diligence. Oh, wait...
Currently possibly tens of thousands of bitcoins were stolen/lost by online wallets and exchanges.
But that's not the point. I just think that big enough portion of users should have all the control. I'm not against all online wallets, but I would prefer them not become the "normal" way to store funds.
Also it makes bitcoin easier target for governments and other enemies.

@Bit-pay: You're welcome; I hope my thoughts help optimize the site! I added some examples and fixes in the list.

One more comment, which I just noticed on vocabulary: be careful with words such as "bad" or maybe even "beginner". Okay, the latter might encourage people who are not tech-savvy, but it must be used with caution. This may seem a minor detail, but the psychological impact of such subtleties should not be underestimated.

And yet another thing I noticed: the web design is fixed-width and sticks to the left side of the screen. This will leave massive blank space to the right on big screens. Scaling sites, such as this forum, have an advantage in that respect. A quick and common fix to this is to center the current 1024px design. Many pages use this method and just add neat graphics to the sides. https://www.casascius.com/ would be an example of such a page.



@DeepBit:
Comparing the required security knowledge between cash and Bitcoin is like comparing counting marbles to Fourier analysis. Do a test on normal people: give one group an item to keep safe, and transmit data to the other. Then look for ways to break their security. Do you seriously suggest that the difficulty will even compare? Who breaks into people's houses in the first place, and even then, good luck finding the valuables. But almost everyone either fails to backup or leaves ridiculously huge security holes open on their computers!

The average person will lose money if you put this to the test.