[Question] Why don't we have new local wallets

[C. Bergmann]

Hi there, developers,

today I heard people complain, that so many bitcoiner want someone else to create anything awesome and that they think, this must be so easy to do and developers are lazy, while they don't understand a bit of the work and struggle and complexity behind bitcoin-development.

So sorry, just for curiousity, as a bitcoin-blogger, and with a ton of respect for the work you do, but I wonder about the development of local wallets. I joined in sommer 2013 and use since than electrum, the bitcoin-core, armory, sometimes multibit. Since 1,5 years seemingly nothing has changed in local wallets.

I tried the windows phone, but there's no wallet for it (what is, considering that microsoft seems to be open for Bitcoin, stupid)

But why no local wallet? Why is there no designer, that swaps some visual basics around this wallet? Why does noone simplify the signing (and sending) of a message, implements a plattform for Plug-ins, a simple tool to write a message in the blockchain, to create an adress with some signs you wish? If it's about the money, isn't it possible to send ad with the wallet? Let everybody send ads while writing into the blockchain (and everybody reads the messages in his wallet - or the messages from some tag)? As a natural verification and message tool a bitcoin wallet could be a mighty software by itself ...

So, just for curiousity, and with the awarness that I doesn't know a peace of the complexity of how to create things like this - why seems the development of wallets so slow? All week opens a new exchange, payment provider, online-wallet, supernode, and all day a new altcoin, but no new local wallet since I joined

<-- sorry, I remind there was a thread about a multi-currency local wallet, so "no local wallet is wrong", I didn't read it, cause I don't think altcoins are the clue at the moment ... -->

What's so special with local wallets?

[bitcreditscc]

Hi there, developers,

today I heard people complain, that so many bitcoiner want someone else to create anything awesome and that they think, this must be so easy to do and developers are lazy, while they don't understand a bit of the work and struggle and complexity behind bitcoin-development.

So sorry, just for curiousity, as a bitcoin-blogger, and with a ton of respect for the work you do, but I wonder about the development of local wallets. I joined in sommer 2013 and use since than electrum, the bitcoin-core, armory, sometimes multibit. Since 1,5 years seemingly nothing has changed in local wallets.

I tried the windows phone, but there's no wallet for it (what is, considering that microsoft seems to be open for Bitcoin, stupid)

But why no local wallet? Why is there no designer, that swaps some visual basics around this wallet? Why does noone simplify the signing (and sending) of a message, implements a plattform for Plug-ins, a simple tool to write a message in the blockchain, to create an adress with some signs you wish? If it's about the money, isn't it possible to send ad with the wallet? Let everybody send ads while writing into the blockchain (and everybody reads the messages in his wallet - or the messages from some tag)? As a natural verification and message tool a bitcoin wallet could be a mighty software by itself ...

So, just for curiousity, and with the awarness that I doesn't know a peace of the complexity of how to create things like this - why seems the development of wallets so slow? All week opens a new exchange, payment provider, online-wallet, supernode, and all day a new altcoin, but no new local wallet since I joined

<-- sorry, I remind there was a thread about a multi-currency local wallet, so "no local wallet is wrong", I didn't read it, cause I don't think altcoins are the clue at the moment ... -->

What's so special with local wallets?

You are right of course, but for the most part the core development team focuses on the underlying stuff more than the QT. If you wish to have a modified version, ask for one from any decent developer, or create a bounty.

[Tarion]

I like the idea of creating a new wallet. I think current wallets lack a lot of options that one might miss. I could think of some features. But what are the features that are really missing?

Basically a UI based on the existing QT client should not be too complex. Maybe we can collect some missing features here so that one can pick them up. Maybe I will do.

[hhanh00]

Developing a new wallet is hard because it has to be flawless or people will lose money. At this point, my impression is that there isn't enough demand for a brand new desktop wallet. What are the killer features that you would like to see in order to switch to a new and yet unproven wallet?

[doof]

I tried the windows phone, but there's no wallet for it (what is, considering that microsoft seems to be open for Bitcoin, stupid)

I started one.  Shipped a wp8 wrapper for coinjar and I'm trying work on one with another dev on here.  Its just time, and no real user base, and no way to monetize it.

I can write other commercial code and get paid.

[C. Bergmann]

Developing a new wallet is hard because it has to be flawless or people will lose money. At this point, my impression is that there isn't enough demand for a brand new desktop wallet. What are the killer features that you would like to see in order to switch to a new and yet unproven wallet?

I'm by far not anyone defining killer features. However, I can help to find ideas ...

- an open plattform for plugins would be very great, so every developer who has an idea can make it real without coding a whole local wallet
- a better UI - better design
- easy coin control
- easy writing messages in the blockchain
- show messages written in the blockchain
- some api that let a wallet connect with thunderbird (for sending bitcoins via email)
- possibility to type in euro instead of bitcoin
- integration of bitID

[C. Bergmann]

I tried the windows phone, but there's no wallet for it (what is, considering that microsoft seems to be open for Bitcoin, stupid)

I started one.  Shipped a wp8 wrapper for coinjar and I'm trying work on one with another dev on here.  Its just time, and no real user base, and no way to monetize it.

I can write other commercial code and get paid.

Great to hear! I love the windows phone.

I thought about money being the problem. Isn't it possible to ask microsoft for support? Or is it possible to let a wallet show advertising?

Also I run a blog, in germany, with about 2.000 readers every day. Maybe it can help (a little peace) if I start a crowdfunding-campaign for better wallets?

[C. Bergmann]

I like the idea of creating a new wallet. I think current wallets lack a lot of options that one might miss. I could think of some features. But what are the features that are really missing?

Basically a UI based on the existing QT client should not be too complex. Maybe we can collect some missing features here so that one can pick them up. Maybe I will do.

If you do I will ask on my blog for funds for you. It won't be much, but maybe we can start something like a campaign ...

[dabura667]

- an open plattform for plugins would be very great, so every developer who has an idea can make it real without coding a whole local wallet
- a better UI - better design
- easy coin control
- easy writing messages in the blockchain
- show messages written in the blockchain
- some api that let a wallet connect with thunderbird (for sending bitcoins via email)
- possibility to type in euro instead of bitcoin
- integration of bitID

- Electrum
- I like Electrum's UI
- Electrum let's you select inputs in 2.0
- Electrum let's you write up to 40 characters with "OP_RETURN "xxxxx" " format in 2.0
- Electrum's transaction status dialogue will show the OP_RETURN contents in 2.0
- I see no need for this... why would anyone need this.
- Electrum allows typing in Amount in many currencies in 2.0
- I don't know much about bitID.

You should download the 2.0 beta source from Electrum's github and install it.

Edit: here's a peek at how OP_RETURN is treated in 2.0
http://imgur.com/a/73N4e

(First image is this transaction: https://blockchain.info/tx/d29c9c0e8e4d2a9790922af73f0b8d51f0bd4bb19940d9cf910ead8fbe85bc9b?show_adv=true )

[jim618]

The OP asks a very good question.

Gary Rowe and I (the two authors of MultiBit) have been working away on a new version that we call MultiBit HD for quite some time now.

It includes:

+ hierarchical deterministic wallets created from 12 wallet words
+ a simpler UI which we hope will 'demystify' Bitcoin for people.
+ Integrated contacts with gravatar lookup, tags
+ Trezor support (in 0.0.5beta)
+ we bundle a Java runtime so there is only one thing to install.
+ more exchanges
+ more display options e.g. BTC, mBTC, uBTC, satoshi etc.  Even 'bits' if you want.

This is in beta, currently at 0.0.4beta: https://beta.multibit.org
The Trezor work will appear in the next release (0.0.5beta)


Other teams have been working away steadily all year - have a look at their github repo's pulse for activity:
+ Electrum 2.0 https://github.com/spesmilo/electrum/pulse/monthly
+ Armory https://github.com/etotheipi/BitcoinArmory/pulse/monthly

+ our pulse (MultiBit HD) is here: https://github.com/bitcoin-solutions/multibit-hd/pulse/monthly

There is also hive wallet which has both a local wallet and a webwallet.
Also on bitcoin.org under 'Desktop' is mSIGNA which I must admit I haven't used but would be interested in hearing other people's opinion of !

UfaSoft are actively supporting their multi-coin wallet.
I am sure there are other projects too - sorry if I have missed yours !

[DeathAndTaxes]

The reality is that writing a wallet (just wallet functionality depending on a compliant node) is relatively easy however writing a full node is very complex and the risk is very high because it must exactly match the network facing behavior of the reference client.  Currently the code for running a node is intermixed with the code for running a wallet.   One can thank Satoshi's original design for that.  This means decoupling a wallet from the reference client is difficult.   The core developers have worked hard to separate out the node logic from the client logic.

Optimally there would be node library and that node library is used to create the reference node and other implementations.  Then the client/wallet would communicate with a node (either locally or remotely).  This would vastly reduce the complexity of writing a custom wallet as the scope would be limited solely to wallet functionality instead of building the entire stack as a monolithic project.  I have no doubt eventually it will happen just no idea how long it will take.

[C. Bergmann]

- an open plattform for plugins would be very great, so every developer who has an idea can make it real without coding a whole local wallet
- a better UI - better design
- easy coin control
- easy writing messages in the blockchain
- show messages written in the blockchain
- some api that let a wallet connect with thunderbird (for sending bitcoins via email)
- possibility to type in euro instead of bitcoin
- integration of bitID

I love electrum, this is my favorite wallet, it has many well-thought ideas and implementations. I am very satisfied with nearly everything. Except that it's not completely clear on what addesses my bitcoins are. (For the controll of adresses I prefer multibit). Maybe this changed with electrum 2.0, but I struggle with installing programms from github.

The UI is ok for me, but I think it's not that good for everyone. e. G. you have to type in the transaction fee, what can be, if you don't know anything about it, some problem.

I'm excited about the new features you told about electrum 2.0.

- Electrum

Is electrum an open plattform for plugins?

- I like Electrum's UI

Me too ... how about some "skins" like winamp had? It looks a bit like windows in the 90ties.

- Electrum let's you select inputs in 2.0
- Electrum let's you write up to 40 characters with "OP_RETURN "xxxxx" " format in 2.0
- Electrum's transaction status dialogue will show the OP_RETURN contents in 2.0

Awesome! Do you need to type it in the console, or is there a button like ("Write this and that into the blockchain!")? I think when both partys agree this tool can serve the same role as a good old letter - proof of message. Or, if done right, it could be used as an advertising service ...

- I see no need for this... why would anyone need this.

Sending bitcoins via email is a great tool, but as long as it needs some centralized server to send out a mail with an invitation to some online-wallet it is more paypal than bitcoin. If there was some way to do this with thunderbird (let's say you attach an encrypted private key that the receiver can easily import in his own wallet, or, better, with his thunderbird plugin that automatically encrypts the file. Pay your bills per Email) This was just an idea, and I think, for people that created something like the wonderfull rescue-phrase in electrum it could be possible.

- Electrum allows typing in Amount in many currencies in 2.0
- I don't know much about bitID.

BitID is a tool to verify yourself with your private key (similar to signing messages). I think it could revolutionize the loggin in the bitcoin-world.

[C. Bergmann]

The OP asks a very good question.

Gary Rowe and I (the two authors of MultiBit) have been working away on a new version that we call MultiBit HD for quite some time now.

It includes:

+ hierarchical deterministic wallets created from 12 wallet words
+ a simpler UI which we hope will 'demystify' Bitcoin for people.
+ Integrated contacts with gravatar lookup, tags
+ Trezor support (in 0.0.5beta)
+ we bundle a Java runtime so there is only one thing to install.
+ more exchanges
+ more display options e.g. BTC, mBTC, uBTC, satoshi etc.  Even 'bits' if you want.

This is in beta, currently at 0.0.4beta: https://beta.multibit.org
The Trezor work will appear in the next release (0.0.5beta)


Other teams have been working away steadily all year - have a look at their github repo's pulse for activity:
+ Electrum 2.0 https://github.com/spesmilo/electrum/pulse/monthly
+ Armory https://github.com/etotheipi/BitcoinArmory/pulse/monthly

+ our pulse (MultiBit HD) is here: https://github.com/bitcoin-solutions/multibit-hd/pulse/monthly

There is also hive wallet which has both a local wallet and a webwallet.
Also on bitcoin.org under 'Desktop' is mSIGNA which I must admit I haven't used but would be interested in hearing other people's opinion of !

UfaSoft are actively supporting their multi-coin wallet.
I am sure there are other projects too - sorry if I have missed yours !

Nice to hear that there is some action! I will download multibit HD and test it.

[C. Bergmann]

The reality is that writing a wallet (just wallet functionality depending on a compliant node) is relatively easy however writing a full node is very complex and the risk is very high because it must exactly match the network facing behavior of the reference client.  Currently the code for running a node is intermixed with the code for running a wallet.   One can thank Satoshi's original design for that.  This means decoupling a wallet from the reference client is difficult.   The core developers have worked hard to separate out the node logic from the client logic.

Optimally there would be node library and that node library is used to create the reference node and other implementations.  Then the client/wallet would communicate with a node (either locally or remotely).  This would vastly reduce the complexity of writing a custom wallet as the scope would be limited solely to wallet functionality instead of building the entire stack as a monolithic project.  I have no doubt eventually it will happen just no idea how long it will take.

I am really blind on the code-side, but do you mean something like toshi (https://toshi.io/)?

[jim618]

I think the previous post about toshi illustrates the point nicely.

The first things you need to decide with a wallet is:
1) How am I going to manage the keys ?
2) How am I going to manage the blockchain ?

Keys and seeds are small. The main problem with them is keeping them safe and secure.
The full blockchain is big now - anyone running Core or Armory knows this.

So you either introduce back end server(s), like Electrum does, or use bloom filtering/ SPV mode, like MultiBit does, to keep things managable.

If you are running a back end then your local client is small and manageable but you have introduced another dependency. If you want to port to lots of devices you want to code in something portable, which is why people use python or Java or write a browser based wallet.

[dabura667]

The UI is ok for me, but I think it's not that good for everyone. e. G. you have to type in the transaction fee, what can be, if you don't know anything about it, some problem.

afaik 1.9.8 will automatically enter the fee for you as long as you set the fee to anything above 0 in the preferences.

Is electrum an open plattform for plugins?

Yes it is. In fact, this article just recently came out.
http://talkera.org/crypto/tutorial-creating-an-electrum-bitcoin-wallet-plugin/
However, Electrum has been accepting plugins for a long time. Just not many developers know how to make it. Any developers who know enough to make a plugin are making their own wallets.

Awesome! Do you need to type it in the console, or is there a button like ("Write this and that into the blockchain!")?

Check the imgur album I posted. you just type in place of the address [OP_RETURN "text you want to write"]  without the brackets, of course.

[ICFiedler]

I enjoyed reading this thread and am happy to follow up on your discussions.

Without having much to add, I like to highlight what I find most important: that for mass adoption -- the key to bitcoin's success -- people need to be able to use bitcoin without understanding what it is. Wallets should thus be like apple's OS: intuitive for everybody but still open for lots of settings for the expert. I think UI cannot be valued highly enough in this regard.

[wonderloops]

It's like hhanh00 said; the hard part is making sure it's flawless. For my wallet, I'm currently working on a build-in visual transaction builder.
Here's a old screenshot of the project, I upgraded the GUI eversince (looks much better now).