Rather:  "6) tell your investors that you have invented a fantastic technology which safely allows 80,000 people to ride one bus with little or no difference in compfort and maintanance costs."


You bet it is.  That must be why Adam Back has ignored all hard technical questions about the "fee market" and the Lightning Network, and instead has been relentlessly spewing out baseless FUD and ad-hominems for the last two months.  "Let Blockstream decide the future of Bitcoin, because Mark is a CIA agent".  And it is working, it seems...


AFAIK not even Blockstream claims that Sidechains will solve the scaling problem.  Their pie in the sky now is the Lightning Network, which seems to be just as plausible as an 80'000 passenger bus...


Well, right now, if they control Blockstream, they control the future of Bitcoin...

Funny that no one seems curious about what the motives of Blockstream may be...



You own a bus company and you are losing money because you have 1 bus, only enough customers to fill 60% of its capacity on average, and 20 employees  with fat salaries in an expensive rented office. You are charging 1 $ of fare, but you are spending 200 $ for each passenger that your carry. However passengers are increasing and by next year you know that you will have to leave passengers waiting for hours at peak hours and times.  What would you do to fix your financial situation:

(1) get a loan from the bank, buy another 7 buses and hope that somehow it will be enough

(2) raise the fare to 200 $ per passenger

(3) wait until next year and then let the passengers negotiate the fare with the driver by a blind auction on the platform.

(4) fire all employees except 1 driver and 1 mechanic, close the main office and move to a back-office in the garage

(5) file for bankruptcy protection.


It is increasing at 0.2 MB per year, the max effective capacity is 0.8 MB, and traffic jams will start to occur when it is 0.6-0.7 MB

(Have a plane to catch, more on return)

The problem is that, as I tried to explain, there will not be a "Gavincoin" and an "Adamcoin".  Even if the chain splits, and the 1 MB branch survives for a while, there will be just a two-headed bitcoin. Even if you could trade and speculate on each coin separately, you cannot attack one in the market without harming the other. 


I cannot avoid being the comp sci prof here.  Technically, what Gavin and Mike say is right, and increasing the limit is necessary for two reasons that can't be ignored:

* With 1 MB limit, it is already too cheap and easy to carry out a spam attack.  A attacker with 10'000 USD budget could probably delay half of the legit traffic for days; an such an attack will get easier and cheaper as time goes on. With 8 MB limit, such attack would be at least 10 times more expensive for the attacker.

* Congestion would make bitcoin unusable for ordinary e-paymets.  The user base will stop growing; users wil be turned away not just because of fee increase, but mostly because of the long and unpredictable delays. 

In contrast, for these past 2 months Adam has been only spreading baseless FUD and ad-hominems.  For one thing, he always says "size" instead of "size limit", then claims that 8 MB blocks will have very serious effects (first it was on full node count, then on orphan rate, then mining centralization...) but does not offer evidence or numbers.  At the same time he ignores completely the two points above.  He promises that the LN will provide the growth that the 1 MB limit denies; but the LN will not work, and will not be bitcoin...

Because the transactions issued by clients are in principle executable in both branches.  Only a few transactions will be specific to only one branch: namely, those that involve coins that were mined in that branch, after the split.  

The incoming traffic is now ~120'000 transactions per day, on average.  For the last 12 months it has grown almost linearly at ~5000 tx/day per month.  If it continues growing at the same pace, by Jan/2016 it wil be ~150'000 tx/day.  The effective capacity of the bitcoin mining network, revealed in the recent stress tests, is ~200'000 tx/day or ~1400 tx/block.

Suppose that the miners split 90:10 between the 8 MB and the 1 MB branches.  The 8 MB branch will have ~130 blocks/day, which will have to carry ~1150 tx/block -- which is still less than the effective capacity ~1400 tx/block.  There may be "traffic jams" at peak hours or peak days, but the backlogs should clear after a few hours at most.

On the other hand, the 1 MB branch will have ~15 blocks per day only.  Each block would have to carry 10'000 tx, which is way over the capacity ~1400 tx/block.  So, every 10 minutes, on average, 1400 tx will be confirmed, and 8600 tx will be added to the queue.  The backlog will only begin to clear when the difficulty is readjusted and the block rate returns to ~10-15 min.  The adjustment happens every 2100 blocks, with is normally 2 weeks but in that chain will be 20 weeks.


Well, right now the 5 largest miners (who have ~70% of all hashpower) could cooperate to starve all the other miners, and take 100% of the block rewards intead of just 70%.  Or they could do much nastier things.

Yes, two weeks is rather short.  Two months would have been better.  I suspect that the 75% will be reached very quickly, or never.  I suppose that Mike was afraid of some PR counteroffensive that would convince many 8 MB adopters to switch back to 1 MB, if the grace period was too big.


Once 75% of the miners have agreed to switch to 8 MB, agreement is irrelevant.  The remaining miners and the major services will choose to switch before the 8 MB limit is enabled, even if they didn want to, otherwise they will lose money.  They will be aware of the event and capable of upgrading in time.


No, I don't agree with that.  

Litecoin is a functioning coin, that non-experts can use as easily as bitcoin, with no confusion.  It has its own exchanges, a large user base, etc.. (It is even used by bitcoin arbitragers, I read somewhere, to quickly transfer value between exchanges, due to its shorter block period.)  

The "1 MB bitcoin", if it survives at all, will have a very long block period, a huge and growing backlog of unconfirmed transactions, practically no services and exchanges -- and very few people will know how to handle it.  

Indeed, if the 8 MB sizelimit prevails, its miners may find it worthy to put some of their hashpower to the task of jamming the 1 MB branch, to minimize confusion of clients who failed to upgrade, and to prevent its use for scams.

instantly the moment of the fork?
across all vendors across the entire world?
what if they are sleeping?
[/quote]

The XT fork will not be a "stealth fork" (like the BIP66 caper).

Again: if and when the blockchain tags indicate that 75% of the miners are ready, there will be a 2 week period during which everyone running Core (or other incompatible software) will be warned of the event and urged to switch to XT (or any compatible software).  If and when that happen, you may be sure that all sensible exchanges and services will upgrade, together with the sensible miners among the other 25%.  Perhaps even Luke.  Perhaps even Mircea.

So, OK, there may be a few "rebel" exchanges left to cater for some thoroughly confused clients who did not upgrade, and the few "rebel" clients who refuse to consider bigger blocks and are willing to put up with 1 block every 2 hours...  But I still don't see any great opportunity there.

(After the 2 weeks period, there will still exist only one branch, and hence only one altcoin, until someone mines a block larger than 1 MB.  That may not take too long.  The average block size in early 2016 should still be 0.7 MB, but variable block spacings and empty blocks may create a backlog of more than 1 MB in the queue at some point.)

Guys, the fork will NOT create two independently usable altcoins -- even if the blockchain splits into two branches that survive long enough to play with.  

Some clever bitcoin hackers may be able to create transactions that are valid on only one chosen branch.  However, most transactions issued by typical users will be executed in both chains;.  Thus they will not be able to "see" the two altcoins.  When one of those transactions, by accident, turns out to be valid for only one chain, the client will either not notice ever, or will get confused, and maybe end up with his wallet in an inconsistent state, maybe even lose coins.

Therefore, there will not be a functioning market for both versions; it would only cater for those clever hackers. Malicious hackers may be able to exploit the confusion of some ordinary clients to do double spends; but payment processors and exchanges will almost certainly deal with just one "bitcoin", the one with mining majority.  

I can't imagine the chaotic outcome that you predict.  Either BitcoinXT (and/or other 8 MB versions of the software) will secure approval of 75% of the miners, in 6-8 months at most, and then 100% of the miners will switch to it; or it will fail to do so, and any miners that switched to it will promptly switch back.  The "fake XT" software is a childish idea; miners are in the game for real money, not for lulz.

But (sorry if I repeat myself) the "fake XT" proposal underscores the silliness of blockchain voting.  It is typical of the hacker mindset: a complicated programming hack that is invented in an attempt to avoid the need for personal contacts and political negotiation. 

Well, sorry, but that does not work.  The size limit issue will be decided by *people*, not computers or algorithms.  The only effective way to obtain a consensus on some issue is to discuss it with the relevant *people*; listening to them, trying to understand their worries and goals, finding a proposal that most of them could approve, or at least accept, triyng to convince them to do so, and getting their pledge that they will do so.   

Gavin, even without being obviously a political genius, knows how to do that (because of his experience at Google, and with the 2010 and 2013 incidents).  That is what he did this time, e.g. getting the Chinese miners to agree to 8 MB.  That is apparently an approach that Greg and Adam, prototypical hackers that they are, do not seem capable of doing -- and, for that reason, hate it, and call it "populist tactics".

I bet that the Core × XT "war" will be resolved that way: the miners (the Chinese, and probably others too) will talk to each other, and eventually agree that they all should raise the limit (by downloading BitcoinXT or some other 8 MB version, or rolling their own); or that they all should stick to 1 MB.

By the way, in the BitcoinXT Github repository there is already a branch that has only the 8 MB limit increase, without the other controversial changes:

https://github.com/bitcoinxt/bitcoinxt/tree/only-bigblocks

(This is not a recommendation, of course; only an information.)

That would be funny indeed.  I do have a Github account...

Please someone do it, and remove the temptation.   

The same thing that would happen if a group of miners with majority power decided to play see-saw with the blockchain.  That is, they mine ten blocks B(N+1), B(N+2), ... B(N+10) normally on top of some block B(N), then go back and mine 20 blocks C(N+1), C(N+2), ... C(N+20) on top of the same B(N), then mine another 20 blocks B(N+11), B(N+12), ..., B(N+30) on top of B(N+10), then mine C(N+21), C(N+22), ... C(N+40) on top of C(N+20), then...

Every sane person would conclude that bitcoin is a joke, while true bitcoiners will find a way to view that as "good for bitcoin".  That is what would happen.

That is all they got in the 0.022 years since the 8 MB XT code was released.  They have only  316'000 minutes left before the end of the year.  What a flop.

It is not.  Any miner can make a copy of XT and remove the patches that he objects to.  Or make a copy of Core and add only those XT patches that implement vote-triggered 8MB limit.

If you are not a miner, you (or any programmer you trust) can make a copy of Core, simply raise the size limit to 8 MB,  and start running that version any time before the fork actually happens.

These developers are listed on Blockstream's "Team" page: Adam Back (President), Greg Maxwell, Pieter Wuille, Matt Corallo, Mark Friedenbach, Rusty Russell, Patrick "Intersango" Strateman, Jorge Timón, and Glen Willen. There may be others. Luke "Tonal" Jr works for Blockstream as contractor. Peter Todd works half-time for Viacoin, an altcoin that claims to be bitcoin done right -- e.g. with 25 times faster confirms.

Who is paying Wladimir's salary now?

Who are the developers with commit rights to BitcoinCore, and what are their affiliations?

To help you, these developers are listed on Blockstream's "Team" page: Adam Back (President), Greg Maxwell, Pieter Wuille, Matt Corallo, Mark Friedenbach, Rusty Russell, Patrick "Intersango" Strateman, Jorge Timón, and Glen Willen. There may be others. Luke "Tonal" Jr works for Blockstream as contractor. Peter Todd works half-time for Viacoin, an altcoin that claims to be bitcoin done right -- e.g. with 25 times faster confirms.


Note quite.  I don't recall writing "delusional fools", bluntly like that, although I do believe that many of them are.  I understand that, strictly technically, bitcoin is not a ponzi; but it is a pyramid scheme, also called "ponzi" in comon parlance -- and that is the only aspect that many bitcoiners seem to care about.  I believe that the price will go to (practically) zero eventually, and the sooner that happens,  the better; but I have no idea how long it will take, actually. 

And I am not "anti-Bitcoin".  I think its is a significant invention and a great technical experiment, and I would admire it if it had not been hijacked by speculators and turned into a giant pyramid scheme -- that has already transferred a billion dollars or two from some naive investors and unlucky speculators to the pockets of smarter and luckier guys, and burned at least half a billion dollars of electricity so that its users could save maybe a million dollars in bank fees.


As a computer scientist, it really bothers me that Adam Back and the Blockstream devs keep spreading all that nonsensical and groundless FUD about the dangers of 8MB, claiming that the "fee market" will work and that the LN will save bitcoin -- while ignoring all data, all that is known about saturated network behavior and service pricing, and all inconvenient questions about the viability of the LN.   

I have no particular admiration for Gavin and Mike; in particular, I cannot respect Gavin given his continuing support for the Shrem Karpelès & Friends Foundation.  But their "agendas" for pushing the block size limit increase are quite obvious: they want to keep bitcoin working as it has been working so far, and usable for the purpose it was created for -- and also reduce the risk of crippling spam attacks.  Whereas Blockstream's agenda, that makes them desperately wish for bitcoin to become unusable and vulnerable to spam attacks, can only be speculated upon.  (Maybe, as Napoleon would suggest, there is no malice, but only gross incompetence and irresponsibility.)


As you know, since I discovered Bitcoin I have been doubling my holdings every day.  (I was tempted to triple them during the recent dip, but I did not want to break a tradition.)

Not necessarily.  Once the 8MB version gets 75% miner approval, the other 25% must promptly switch to it, from self-interest.  Then hopefully no one will pay attention to the Blockstream FUD anymore.


As per above, it will be very difficult to find such miners.  If and when the chain actually splits, every miner who cares about his money will be on the 8MB branch. 


Since those coins will exist on a chain that will be issuing only 1 block every 2 days, "slightly more than its market value" may be zero plus 10%.

Not worth pointing out the problems with the other steps...

Right now 1 company has control over bitcoin's future, and has been spreading baseless FUD and ad-hominem because it does not intend to relinquish it.  If that is not bad enough already...

Garzik's proposal, BitcoinCore and BitcoinXT will be the equivalent for a while; the last two wil be equivalent until 2016.  After the first 1.2 MB block gets mined, it may be that Garzik and XT still agree.  Even BitcoinCore could be patched to retroactively accept that block and the branch built on top of it, orphaning any small-blockian branch that may have been mined after the fork.

That's because there is no way to tell, looking at a block, what was the block size LIMIT adopted by the progam that generated it.  One can only see the block's SIZE; if it 0.75 MB, it will be valid for both versions, and for any other custom version out there that did not reject it for its reasons.

indeed, if you don't trust Mike and Gavin's XT code, but (for some unfathomable reason) you do trust Adam and Greg's Core, you can make a copy of the latter and apply to it the patch of XT that implements 8 MB blocks after bla bla bla.  Then you can use that version (just as you could use XT) whether XT gains 75% acceptance or not.  You will be in trouble (as if you used XT) only if XT gains 75% majority, some jerk solves a 1.5 MB block, and  then there is massive switch back to Core (or to some other small-block version).  But if that happens you are better investing in lottery tickets than in bitcoin anyway.

Indeed, if you are not a miner, you can take a copy of Core and change the line "MAX_BLOCK_SIZE = 1000000" to "MAX_BLOCK_SIZE = 32000000", start using that modified version anytime before 2016, and never worry about the issue again until 2022.  Your transactions will be fine, and no one will be able to tell that you are using an heretical Core.  You will run a small risk of some mining jerk solving an oversize block before the due time, that will be accepted by you but rejected by all other players; but that block will be quickly orphaned, so that event will not be different from the ordinary orphaning that has been happening 2-3 times every day since forever.  (And that is why you should wait for 6 confirmations to be reasonably safe.)

Indeed that is the point.  The way the fork is programmed, every transactions will be in principle valid under both sets of 'consensus rules' and executed in both branches of the blockchain.  Sometimes, by accident or crafty art, a transaction may be executed in only one branch; or the same input may be sent to two different adrresses, one in each branch.  However, ordinary clients will have no control over such events (and may be confused by them).

So, it does not make sense to call either branch an altcoin, or to invent a symbol for it.  Splits of the chain (ophaning) happens all the time even when everyone is running the same version of the software.  

It is also not logical to say that one branch is an altcoin because it uses different 'consensus rules' than those that were in force before.  Once a block smaller than 1 MB has been solved, it is impossible to tell which software was used to solve it.  (There is a version field in the header, but the software can set ti to anything it wants; it is ignored by the protocol.) Thus one can claim that the limit has always been 8 MB since genesis, or 17.3 GB; and both branches would then be valid according to these 'consensus rules' too.

Or also: suppose that someone posts a version of the software -- let's call it BitcoinZZ -- that does NOT cut the block reward in half when it should.  If that version attracts enough miners, but not 100%, the chain will split into two branches, one with 25 coins of block rewards, as in the common trunk, and the other with 12.5 coins.  Which one will be the true Bitcoin: the one with a change in the reward, or the one whithout the change?

This question is connected tp a famous logical/philosophical paradox of Bleen and Grue.  Everybody knows that emeralds are grue, and sapphires are bleen.  But a curious thing will happen on March 17, 2017, when emeralds will suddenly become bleen and sapphires will turn grue.  Some weirdos say that things are 'green' if they are grue until that date and bleen afterwards; but why would one give a name for a color that becomes a different color at some date? That is stupid...

Because the bitcoin traffic, if it keeps growing as it has over the last 12 months, will hit the effective capacity of the mining network by mid 2016.  Then traffic jams will be frequent, and the average time to 1-confirmation will jump from ~10 minutes to hours depending on the time and day.  Bitcoin adoption will stop growing and bitcoin's image will suffer.


The stress tests were not directly related to the blocksize issue.  The company that claimed responsibility for them (Coinwallet.eu) has some wallet software that, they claim, will compute the correct fees to get over traffic jams like the ones that they did.

No, they have worked hard to make bitcoin unusable for ordinary payments, and to drive all bitcoiners to some lousy centralized banking system, that no bitcoiner would want to use if he had the choice.