For a certain class of problems this works.... but where it fails:

1) trusting strangers with your data / programs means it will not be usable by industry
2) I doubt it will be more cost effective than centralized alternatives.

Actually the core class of problems this will not help with are those in which the transmission and subdivision overhead of data transfer outweigh the speed benefits of distributed computing. This is the primary problem with distributed computing itself.

great project.

what time do you expect to finish it?

It will be completed in stages.  Depending on how much time I get to spend on it, these might arrive quicker or slower. Right now its a spare time project because I'm working on a full-time project for a client.

First of all it will just be a viewer for the various markets and assets, presented in a format familiar to traders.
The next version will have browser signed transactions and other technologies to allow a web-wallet and on site trading.
The next version will have bitUSD->bitcoin on/off ramps either by using an exchange or by directly purchasing bitUSD.

All of this is subject to the conditions of running a delegate for pay (and the price of BTSX), which appear to be in flux.

Cheers, Paul.

so DDOSing is not an issue IMHO

maybe campaign trash talk bashing .. but not DDOS

Here is one such exploit:

Enumerate all nodes on the network by following connections from the client. Delegates have to be online to produce blocks, so you'll be able to gather their IP addresses along with everyone elses.

Then, simply run through that list, DDOS bursting at each IP address and watch every single delegate for stats dropping. You will be able to find a map from IP to delegate with quite a high probability with enough DDOS probes.

edit: might cost a lot if you don't own a botnet, but the risk is there.

If true, then 5 delegates in the top 100, is a nice little dilution earner.

I think they'd need to be created again to get more than 3%.

Hug? Since when are delegates supposed to pay a fee besides registration fee, which is pretty tiny:

Quote

  "delegate_reg_fee": "14.09802 BTSX",

Sorry, I thought that's what this thread was discussing? I must have misunderstood

designing for testability negates the needs for such a stringent code revew system. IMO its better to start by offering rewards to valid unit tests and then work on an integration test suite.. block box and white box testing prevents most problems in the field.. after that some system should be inplemented to offer rewards via reviews but shouldnt be the main focus right now

I couldn't agree more. Actually - providing rewards for unit tests by external parties which reproduce a failure would be hugely valuable.

Ok, the previous error was due to having an old version on binutils on the path before the new one.

Now I've got what appears to be a compile error:

Scanning dependencies of target real128_test
[ 60%] Building CXX object libraries/fc/CMakeFiles/real128_test.dir/tests/real128_test.cpp.o
Linking CXX executable real128_test
/usr/lib/../lib64/crt1.o: In function `_start':
(.text+0x20): undefined reference to `main'
collect2: error: ld returned 1 exit status
make[2]: *** [libraries/fc/real128_test] Error 1
make[1]: *** [libraries/fc/CMakeFiles/real128_test.dir/all] Error 2
make: *** [all] Error 2

Seems that the unit tests have the wrong include path for the definition of the boost unit test marco BOOST_AUTO_TEST_CASE, it should be:

#include <boost/test/included/unit_test.hpp>
not

#include <boost/test/unit_test.hpp>
In files:

bitsharesx/libraries/fc/tests/task_cancel.cpp
bitsharesx/libraries/fc/tests/real128_test.cpp
bitsharesx/tests/deterministic_signature_test.cpp
bitsharesx/tests/dev_tests.cpp
bitsharesx/tests/nathan_tests.cpp
bitsharesx/tests/wallet_tests.cpp
bitsharesx/tests/bitcoin/wallet_import_tests.cpp

However, after these issues were fixed up, I am now in business


Cheers, Paul.