Bitcoin-Qt, the future Bitcoin client GUI [user input needed]

[JuanPabloCuervo]

Not really an alternative client, but an alternative GUI for the original bitcoin client based on Qt4, which will be merged into mainline eventually (0.4.x).

It now has all the functionality of the original client's wxWindows GUI, and I'm working to extend it beyond that.

Why Bitcoin its so slow downloading all the Blocks for the first time?

... it does not matter if i have in the same network/room another PC with all the blocks and a fast Base1000/T network connection..
it takes 1/2 day to download...
it should download a lot faster.
 

[1713946648]

1713946648

[1713946648]

1713946648

[1713946648]

1713946648

[1713946648]

1713946648

[Garrett Burgwardt]

Why Bitcoin its so slow downloading all the Blocks for the first time?

... it does not matter if i have in the same network/room another PC with all the blocks and a fast Base1000/T network connection..
it takes 1/2 day to download...
it should download a lot faster.
 

The idea with bitcoin is that you can't trust anyone. So on the first download of the block chain, it has to check to make sure it is being sent the correct information, and 'work things out' on its own.

[wumpus]

I'd like to see wallet encryption and multiple wallet support first. The rest is just icing on the cake.

Agreed -- the thing with wallet encryption is that it is still experimental, and not in any official release of the client yet (only in git), so I'll add a big experimental banner.

Anyway, thanks for doing all the work on this, it's really coming together nicely!

Thanks!

edit: splash screen would also be awesome. i know its mostly eyecandy, but the startup right now gives a totally horrible first impression.

And also easy to implement!

When can we expect Bitcoin-Qt to be merged into the main tree, and released as the official client?

I don't know. From what I understand from Gavin on the mailing list, he wants the upcoming releases to be dedicated to bug fixing not major new features. If that's the case, I'll do a parallel release for a while. It just costs some time to build and distribute proper binaries, which I rather spend on development, but hey if there's no other option

[JuanPabloCuervo]

as there are still some very important bug fixes there, but it'd be more clear to release a bitcoin-qt version for each core version.

i think the main problem with Bitcoin.exe, its that does not has a blkindex.bak file. like http://memory.dataram.com/products-and-services/software/ramdisk
or ECC mode like servers memory.

yesterday had a corrupted blkindex.dat (Power loss / forced shutdown / etc...)

eliminated the blkindex.dat, restarted Bitcoin,
and blocks started from 0 again, BUT this time Bitcoin.exe did not overwrite the blk0001.dat, just started from 0 to but added to it.

Problem with Bitcoin.exe is that does not have a failsafe feature, its too delicated.
pull 1 string everything falls appart.

havent had those kinds of problems with Linux, so far anyway.

[wumpus]

i think the main problem with Bitcoin.exe, its that does not has a blkindex.bak file. like http://memory.dataram.com/products-and-services/software/ramdisk
or ECC mode like servers memory.

I agree it could be more robust. No need to implement anything complex for it either, in case the block index gets corrupted it should simply re-download the corrupted blocks (like bittorrent does). But please don't make this a general bitcoin issues topic There's the github issue tracker for that.

BTW: splash screen has been implemented

[goodlord666]

This looks really cool so far!
Keep up the good work.

BTW:
isn't making the field for decimals narrower than the field for whole coins a little optimistic? :)

[fornit]

BTW: splash screen has been implemented

works here

btw: in the forums, i often find threads where people suggest using --rescan to solve whatever problem OP has like missing transactions etc. the usual answer is "wtf cannot find the --rescan button" or something like this. i understand thats usually done on startup, but is it feasible to rescan the blockchain with the program already running? or restart the programm and rescan? i think having a button for that would save a lot of less tech-savvy people some time and headaches. also an option similar to the datadir parameter would be very convienient. ok, thats probably one of the steps to multi-wallet support anyway, but i just wanted to mention because its also easy to do and the parameter is often used. or would be, if people knew it exists 

btw2: the program gets unresponsive after a while for me. client area doesnt update anymore and doesnt close on X, usually when i do something on the computer. if i just leave the computer alone completely it works. OS is win7 64bit. complied and run exactly as you instructed in github.

[wumpus]

isn't making the field for decimals narrower than the field for whole coins a little optimistic?

LOL, good point, I've made them equally wide now.

btw2: the program gets unresponsive after a while for me. client area doesnt update anymore and doesnt close on X, usually when i do something on the computer. if i just leave the computer alone completely it works. OS is win7 64bit. complied and run exactly as you instructed in github.

Haven't heard this one before. This will be a hell to reproduce or debug if it happens randomly on your PC and not on a user action :/

[wumpus]

Recent transactions on overview page now

[nuke]

Why there are no Mac OSX instructions in the README file?
I wanna try it

[wumpus]

Why there are no Mac OSX instructions in the README file?
I wanna try it

Because no one contributed them

I guess it is pretty easy though, and people have done it before (bitcoin.pro contains MacOSX rules), just install the MacOSX Qt SDK, and build the project using qmake or in Qt Creator.

You need the same dependencies as for normal bitcoin except Wx (follow the steps in doc/build-osx.txt to install them).

I cannot give exact instructions as I have zero pieces of Apple hardware (well except for a dinky first-generation iPod )

[bittersweet]

- I've always wondered why Bitcoin has separate Credit and Debit columns. Why not one "amount" column? Is this really more clear?

I think the best solution would be to let people add and remove columns they want to see. Like, you right-click the column titles bar and have a list of possible columns to check/uncheck.

[wumpus]

I think the best solution would be to let people add and remove columns they want to see. Like, you right-click the column titles bar and have a list of possible columns to check/uncheck.

This does complicate the UI, so I'd add it only if really needed. I think the current columns suffice (Status, Date, Type, Amount).

What other information about a transaction could you possibly want in a column?  Note that you can always request the transaction details by double clicking on one.

[phorensic]

Finally got around to install Ubuntu on a VM to try this out.  Compiled first try.  Looks awesome so far. Once the blockchain downloads I'm going to give it the full runaround.  Looks good so far!

Update:  This client downloaded the blockchain approximately elevntybillion times faster than the official Windows bitcoin client.  Even though it seems to be just a wrapper on top of the official bitcoin client, that's awesome!

[elements]

Why there are no Mac OSX instructions in the README file?
I wanna try it

Me too!

Is someone working on a Mac version?

[twmz]

OK I need some advice; what would be best to implement next? What feature is most desired? I have a few things planned, but need to prioritize them;

My interest would be in these two things (in priority order):

• Integrate wallet encryption (is done in core, just needs GUI)
• Print/import/export private keys (is done in core with pull request, just needs GUI)

And for that matter, I would be fine with import/export only being available from the command line and not in the GUI at least initially.

[willphase]

come to this thread from another, just wanted to comment good job with your bitcoin-qt client.  The problem I have is that there is a lack of trust in non-official clients, you are not providing binaries (most people just want to download something, and not bother compiling anything).

I think the solution is to try and get some of the features you have into the mainline client - or at least try and solve the trust issues by having your client endorsed by bitcoin.org or digitally signed releases and vouched for pgp keys so people can start trusting your client.

Otherwise, coming into threads that are talking about additions to the mainline client and saying 'my client already does this' isn't too helpful... imho.

Will

[wumpus]

And for that matter, I would be fine with import/export only being available from the command line and not in the GUI at least initially.

There is a pull request on github that implements commandline/RPC functionality for importing and exporting keys:

https://github.com/bitcoin/bitcoin/pull/220

I'd need to merge this anyway if I want to expose it in the GUI.

I think the solution is to try and get some of the features you have into the mainline client - or at least try and solve the trust issues by having your client endorsed by bitcoin.org or digitally signed releases and vouched for pgp keys so people can start trusting your client.

If you don't trust me, that's up to you -- you're free to use any client that you want. Bitcoin is a distributed currency. There is no "official" bitcoin organization to endorse anything.

Everyone can audit all my code on github, that's how open source works. I would be incredibly stupid to backdoor it wouldn't I? Git remembers every change ever made to a code base. If I was really out to steal your coins there'd be much less time consuming ways that wouldn't put my reputation at risk.

I cannot help but take this as a bit of an insult. Good luck with your trust issues, but don't look at me for fixing them.

[willphase]

If you don't trust me, that's up to you -- you're free to use any client that you want. Bitcoin is a distributed currency. There is no "official" bitcoin organization to endorse anything.

Everyone can audit all my code on github, that's how open source works. I would be incredibly stupid to backdoor it wouldn't I? Git remembers every change ever made to a code base. If I was really out to steal your coins there'd be much less time consuming ways that wouldn't put my reputation at risk.

I cannot help but take this as a bit of an insult. Good luck with your trust issues, but don't look at me for fixing them.

I have no issues trusting you, I've seen your posts here and on the bitcoin-dev mailing list and wouldn't have any issues either running a binary supplied by you or just simply building your code from source - in fact I suggested that you put binaries up to make your client more accessible to the masses.  I think my point perhaps was poorly explained in my post.  My point is that the average user is going to 1. Google for bitcoin, 2. click on the top link, and 3. download the windows client.  If we want to work on the average user's user experience it's important that we concentrate efforts on improving the mainline client.  The alternative to this would be to place easy to access and easy to install binaries on the bitcoin.org website so people can have a choice between your client and the mainline one.  The only criticism I had was more a philosophical one that if someone comes in and says 'hey I'm going to fix this bug in mainline' or 'hey I'm going to add this new feature in mainline' then you pop into the thread and say 'oh I already have that in my client' - but very very few people are using your client (see above) then it makes volunteers far less likely to contribute to mainline (or at least that was how I felt when I said I was going to implement a usability feature in mainline in another thread)

I would prefer to see more activity on mainline, more people contributing e.g. to the test suites, stability improvements, UI improvements, audit of the code, assurance that mainline is being improved and released in a timely manor, than for the scant development resource to be thinly spread over several clients.

I totally understand why you started your client, probably a combination of the inertia of pulling code into mainline, and the belief that having client diversity is good for Bitcoin as a whole (both reasons I agree with, by the way )- but right now I go to the Newbie thread and try and help people and I see the same issues being reported over and over again and they really need to be fixed...

Apologies if my post came across as an insult, it wasn't meant as such.

Will

[wumpus]

I have no issues trusting you, I've seen your posts here and on the bitcoin-dev mailing list and wouldn't have any issues either running a binary supplied by you or just simply building your code from source - in fact I suggested that you put binaries up to make your client more accessible to the masses.  

OK I misunderstood that part, I read it as binaries being more trustable so I thought huh

I think my point perhaps was poorly explained in my post.  My point is that the average user is going to 1. Google for bitcoin, 2. click on the top link, and 3. download the windows client.

Yes -- I do intend to do a binary release, probably in parallel with the next mainline release (so that there is a wx and qt version). I'm already working on a cross-compilation environment to built it in, but it is quite involved. I do have some help with Linux packages, but for Windows I'm on my own (and myself I hardly use it).

Apart from that it will be easier to get my version into the official repositories of Linux distributions such as Ubuntu as it does not depend on some arcane Wx version.

It's all being worked on but there's a lot to do and so little time...

If we want to work on the average user's user experience it's important that we concentrate efforts on improving the mainline client.  The alternative to this would be to place easy to access and easy to install binaries on the bitcoin.org website so people can have a choice between your client and the mainline one.

My UI is scheduled for inclusion into the mainline client eventually. I'm just not sure when. This is why this thread is called "future bitcoin client GUI". Only if you want a sneak peak, you need to build it yourself for now.

But for that reason it's indeed a waste of time to implement usability features in the mainline client. So if I discouraged you in doing that, that's exactly the point So that you can direct your development efforts to some place they make sense...

The only criticism I had was more a philosophical one that if someone comes in and says 'hey I'm going to fix this bug in mainline' or 'hey I'm going to add this new feature in mainline' then you pop into the thread and say 'oh I already have that in my client' - but very very few people are using your client (see above) then it makes volunteers far less likely to contribute to mainline (or at least that was how I felt when I said I was going to implement a usability feature in mainline in another thread)

I don't have exact statistics, but by the amount of feedback I get, it seems a lot of people are using my client already (which isn't too strange, as I have some features that have been requested for ages).

I would prefer to see more activity on mainline, more people contributing e.g. to the test suites, stability improvements, UI improvements, audit of the code, assurance that mainline is being improved and released in a timely manor, than for the scant development resource to be thinly spread over several clients.

My client shares all the network code with the mainline client, so test suites and stability improvements apply to both. I regularly merge in the mainline repository... And have also contributed back some patches.

I totally understand why you started your client, probably a combination of the inertia of pulling code into mainline, and the belief that having client diversity is good for Bitcoin as a whole

Not entirely. I started my client to make a more attractive, usable and feature-rich UI for Bitcoin. It has existed in limbo much longer than originally planned, but that's indeed due to the inertia thing...