It isn't essential for OpenCoin that ANYONE but OpenCoin (or approved lackeys) run servers.  It may be essential to your goals but OpenCoin has indicated that the network will work fine with just a handful of servers.   With decent hardware a handful (or one) server could handle tens of thousands of transactions a second without breaking a sweat.

If OpenCoin (and the source code dispaeared) tomorrow you really think the Ripple network could survive?  What someone is going to reverse engineer the server software?  At best that would be an incompatible hard fork.  The existing network operates solely at the will of OpenCoin.  Period.

Of course I don't.  It is closed source until it is released under open source license.  By your "logic" if they NEVER release the source code then it can never be closed source.  Hmm?  Yeah that makes sense.

Source code =/= open source.

Open Source is an open license.  It can't be undone.  If Satoshi showed up tomorrow and said "nope I want Bitcoin to be closed source in the future" he would have no legal authority to do so.  The MIT license grants others the right to modify and distribute the code.  Open Source is about free software (free speech not free beer).  Until it is licensed as open source you are just hoping they will do so in the future.  If they don't?  Well that sucks but you have no legal authority to modify or distribute the code.  Open Source can't be "reversed" and that makes it incredibly powerful.

Ripple Server is closed source until it isn't.  Maybe it will be open sourced in the future but .... maybe it won't.  That is the whole point.

It won't be released ... certainly not in the next couple months.  OpenCoin is already distancing itself from Joel comments about wanting (not promising) the source code to be released by the end of the year.  It won't die because it was always intended to be bait and switch.  Just like you are trying to tell me night is day (closed source binary isn't closed source) in a couple months you (and others) will be even more invested and have a dozen excuses/rationals on why the code can't be released yet.

Beta doesn't magically make it open source.  Bitcoin is considered an open-beta and the source code is available.  Imagine that.


Microsoft Windows 8.1 Preview (http://windows.microsoft.com/en-us/windows-8/preview)
Closed Source, Beta

Bitcoin
Open Source, Beta

Ripple
Closed Source, Beta

Simple.  Release a new version which drops connections from servers older than prior version.  Revoke access to the proprietary closed source binary to those you no longer want to run the server.

That is a dubious distinction.  It has been "released" the network is in use.  If it was an internal closed testnet only by companies employees that would be one thing but the network has been released to the public.  Real transactions are occuring everyday and they are occuring on proprietary transaction servers running closed source code.

Lets take your logic to the extreme.  Say 10 years have passed and the source code hasn't been made available "yet".  Is it still an "open source project"?  Yes or no.  If yes how about 5 years? 3 years? 1 year?

Open Source is open source.  If the binaries are not released under an open license and the source code made available then it isn't open source.  Period.  By the Open Source Initiative (http://opensource.org/), the Ripple server is not open source.  Hell by the definition of anyone other than a fanboy the Ripple Server is NOT open source.  Even Ripple (and supporters) don't claim the server is open source ... BECAUSE IT ISN'T.   The mere promise/claim of future open source doesn't make something open source.

That same way PayPal is secure; OpenCoin acts as a central authority.  They maintain "the books".  Mining is only necessary to achieve consensus among untrusted peers.
On PayPal if their servers say you have $100 then you have $100 and if they say you don't then you don't.  Ripple is no different. 

IT IS CLOSED SOURCE.  That is the DEFINITION of closed source.  It is a centrally controlled proprietary network.  The fact that the company has made some vague promises to release the code to the public at some unknown future date doesn't make it open source. If Microsoft promised they would make the source code of some future version of Windows on some future unnamed date would that make Windows "open source" today?  Of course not.  

Servers run closed source binaries only with limited and revocable permission of the central authority (OpenCoin).  The only transactions servers that exist are the ones that OpenCoin has allowed to operate (and can revoke at any point in the future).
What exactly would you call that other than closed source and centrally controlled proprietary network? The fact that it may (or may not) change in the future doesn't negate the reality today.

Yet you admit in the next post that the transaction server IS closed source and only available to approved lackeys of OpenCoin.  Replaced "owned" with "approved".  Nobody runs the closed source binaries with OpenCoins experess permission.  Permission which can be revoked at any point in the future for any reason. 

Closed source and absolute central control of the network.   Ripple is PayPal 2.0.

"Scientific Wild Ass Guess"

"By mid October".  So the first orders will ship in late September and go through mid October. 

Still the product being real doesn't mean the seller is legit.  You are concentrating on the wrong thing.
KNC is real, the product is real, the timeline is real, but you may still end up with nothing.

Provide a link to the download.  Saying there are 100 servers is meaningless.   Google's search engine runs on thousands of servers.  It makes about as much sense to say "if the conduct of Google, Inc is a problem at least the world will have their [Google, Inc] technology".  No they won't if access is restricted.

Do I really need to clarify with "OpenCoin and approved lackeys"?

No they won't the transaction server is closed source and only runs on OpenCoin owned hardware.   

Generally quantum computers do a poor job of breaking symmetric encryption and hashing functions.  Their real threat is against asymmetric cryptography like ECDSA used by Bitcoin for signing and verifying transactions.   https://en.wikipedia.org/wiki/Shor's_algorithm  Shor's algorithm allows finding a private key given a public key in polynominal time which is many magnitudes faster than classical computing solutions.

D-Wave however isn't a general purpose quantum computer, it can not implement Shor's algorithm and is absolutely useless for breaking cryptography.  ECDSA can be broken with a large enough quantum computer but nobody has even broken 16 bit keys using a Quantum computer yet much less the 256 bit ECC keys used by Bitcoin.

It is not that you CAN'T it is that it won't be very efficient.  Remember there is no such thing as free electricity even if you have a free/waste heat source; there is always a capital cost and all equipment has a finite lifespan.  So if your $10,000 turbine produces so little electricity over its economical lifespan that it has an average cost of $1.00 per kWh it doesn't really make sense.   The hotter the input and the colder the output the more efficient the energy conversion is (thermal energy to electrical energy).


The waste heat from miners is simply too "cool" to be useful for power generation.   It can be useful for heat (either heating the air in a house, or heating water) but even then you have some material costs and you likely need to "water cool" your rigs.  Trying to capture waste heat from the air is an excercise in futility.

On edit: Bitweasel beat me.

To date all rigs have either been standalone (the rig has ethernet and you just connect it to a network) or use USB to connect TO a computer (host).  None have been installed inside a computer.

It is not even close.  Depending on the ASIC model, a GPU uses about 50x to 500x as much power.
https://bitcointalk.org/index.php?topic=281279.0

A 2.5W USB Block Eruptor outhashes many 200W+ graphics cards.


No.  Well maybe if they already own GPUs and they have free power but even then it is getting close to the point it isn't worth it.

Yeah 2 MW/h is a nonsense unit.  Maybe that is why we didn't want to ship them.

2 MW = measure of energy
2 MWh = measure of power
2 MW/h =


This.  It is just an assumption that the 1 PH/s will be built out using existing chips.  An assumption that I think is unlikely otherwise we would be seeing the hashrate climbing the last couple months.  My SWAG is that AM is near its power capacity and that is why it has been selling of blades rather than expanding hashing power.  It will remain ~50 to 60 TH/s max until 28nm? 55nm? blades start to replace the existing 110nm ones.  For each new blade deployed they will sell off the existing used 110nm blade until the entire farm is "switched over".  Depending on the relative efficiency that would expand the farm's output by 4x to 15x without increasing the load. 

Yeah it perfectly clear from the OP but it would appear the power supply is standard ATX PSU.  right?  

I haven't seen a manually switched ATX power supply in about ten years.  Power supply companies hate them because you know some x% will forget to switch it and that just means x% more RMAs.

IF BFL is actually shipping manually switched (god I would hate to see how inefficient those low end pieces of garbage are) power supplies to customers in Europe without either setting them to 110V or putting a warning sticker on the plug connector well that is just sad.

What is "fit in one unit"?  To use more than 16 boards requires 2+ m-boards.  All 400 GH/s orders shipped with additional m-boards. 

The additional boards were shipping with the original order.  The individual boards were rated @ 25GH/s they seem to do on average 22GH/s (some more and some less).  

Starter kits (25 GH/s total vs 44 GH/s shipped):
Advertised: 1 host + 1 board
Shipped: 1 host + 2 boards

Full Kit (400 GH/s total vs >= 400 GH/s shipped)
Advertised: 1 host + 16 boards
Shipped: 2 hosts + 18-19 boards (enough to ensure total output is >= 400 GH/s)


Still the bet was poorly worded and hopefully people in the future will start making better worded bets.  Bets should be written as contracts so there is no possibility of ambiguity.

What is a "unit"?  It would appear the company delivered >=400 GH/s using 2 hosts (M-board) and 18-19 hashing boards (H-board) instead of 1 host (M-board) & 16 hashing boards (H-board) as designed.  The hardware was on time (for 5+ orders per the bet), at least 400 GH/s was delivered per 400 GH/s order, and the efficiency was <= 1J/GH (looks like ~0.7 J/GH). 

The only thing in dispute is doe delivering hardware using 2 hosts + 18 boards not meet the criteria of a "unit"?  The OP complaint isn't that not enough hashing power was delivered or it wasn't efficient enough.  His complaint is that on the 400 GH/s sales it shipped as 2 hosts + 18-19 boards instead of 1 host + 16 boards.  Is 2 hosts + 18-19 boards "one unit"?  I would say probably yes but bets should be worded better so there is absolutely no ambiguity.