Why XT Failed


Active Member
Aug 19, 2015
I suppose it is now safe to say that XT failed. The miners have categorically rejected it. The only way the economic majority can force XT is by the nuclear option, but we do not need to go there because there is unanimous agreement that the blocksize limit should be increased across all sections of bitcoin and because I think BU has the right and very elegant solution to the blocksize question so hopefully we will manage to persuade the whole network to upgrade.

However, if we are to succeed, we need to understand why XT failed. As you all may remember, the process began with an initial suggestion by @Gavin Andresen to raise the limit to some 20mb with some 40% yearly growth. I am not sure if there was any public consultation on what the first suggestion should be, but after the suggestion was made it seems that after some consultation, not sure if any in public but presumably at least some in private, the yearly growth was dropped and only 20mb was publicly suggested and so started a heated and an ongoing public debate. The miners then, presumably after some private consultation, agreed, in blood so to speak, to 8mb and publicly announced their agreement.

It seems that after some public feedback whereby users asked for 8mb with a growth rate, Gavin decided to counter the suggestion of the miners with what the miners say is 8gb "we asked for 8mb he gave us 8gb, hell". Now we all know it was not going to be 8gb until some 20 years, it was a limit not a size, technological growth, etc. but we do not know what sort of consultation happened in private between Gavin and the miners. In any event, some sort of miscommunication, we all humans, we make mistakes, so all is still sort of fine, but what happened next was decisive.

The miners were somewhat angry, but still very much open to a solution. They suggested 4mb, doubling every 4 years, up to 32mb. Mike Hearn however turned around and stated that XT was no longer open to any negotiation. That the parameters were not going to be changed, even if all miners agree, they can go and create their own client (I paraphrase and if I misstate is whole unintentional). This is where the slow road to XT's failure began. First in private, then in public, the miners started making it very clear that they were not going to adopt BIP101 or XT and in line with Hearn's request they came up with BIP100 which the economic majority rejected and so on we come to where we are.

Gavin, as we all know and to his credit, remained open to any suggestion, including Adam Back's 2-4-8 to whom he said get everyone to agree to it and we go from there.

As I hope can be seen, the failure of XT was, in my opinion, due to a lack of user feedback and user engagement. Instead of a reasoned, public debate in an appropriate forum, what public engagement there was from the devs was carried out on reddit, a platform more suited to short discussions than in-depth engagement with the aim of reaching an agreement. Further debates were carried out on the mailing list, so insulated from the users, where the issues were limited to technical aspects rather than a holistic view of all, and especially economic, considerations. And much debate was carried in private or on irc.

It seems to me that, over the years, the developers have insulated themselves further and further away from the users, so having their discussions in fleeting irc, or limiting their discussion to code alone in now heavily moderated mailing lists. The effect has been the creation of what can be termed a god complex. The developers, rather than being mere maintainers of code, have been turned into celebrities and have been given such god like attributes as "the smartest" or even geniuses, that they have infinite knowledge of crypto or whatever, that only a handful can do the job and other god like qualities. The expected order of the day is deference and even submission to the walking gods.

The effect of this is the creation of a wall between users and code maintainers and the effect of this wall is what Jeff Garzik called the greatest disconnect between users and developers he has ever seen in his 20 years of working in open source. The effect of this great disconnect is the ugly debate we have so far seen, where the blocksize question has been turned into some pr campaign with back and forth between the camps where once one side seems to have decisively won it is then seemingly decisively wounded and thus the other side wins which in turn is seemingly decisively wounded and so on, in the process creating a far greater division than necessary, but worse, fostering obfuscation through soundbite talking points.

Bitcoin Unlimited has, in my opinion, learned from these lessons and offers not only a technical solution, but also a social solution. It tries to bridge the gap between the two divides of small and big block on one and user and expert on the other. It offers a constitutional democracy as opposed to dictatorship or tyranny or "whatever I say it is" consensus, with the aim of keeping the developers in check and accountable, so bringing them down from their ivory towers and into our accessible spaces. It further offers an interdisciplinary approach where code is one aspect of bitcoin with economical, mathematical, political, legal etc considerations being just as important.

I give credit to Gavin for now engaging with us, although some still complain that he is inaccessible but one must think he is very busy, and I hope other developers see the wisdom in our ways and choose to engage in public forums where feedback can be given and thus this debacle can be avoided from happening again.

If there is one thing to learn from the failure of *desktop* linux is that the ios the linux developers created contained and still contain so many little annoyances that need not in any way be there and are very easily fixable. However, lack of ordinary users feedback, lack of engagement, and the lack of an interdisciplinary approach as needed for any "product" to succeed destined it to be of use to a very tiny minority.

I hope we learn from both failures and incorporate our lessons in our approach forward and in so doing heal both of the two divides in our community and carry this debate forward in an open, public and accessible manner.


Active Member
Aug 30, 2015
I don't agree with the premise. Some miners are conservative, want to delay change as long as possible, and they also need to do the same as other miners. The game is underway.
  • Like
Reactions: freetrader


Active Member
Oct 10, 2015
XT has not failed,yet, we need to get XT supporters to look at BU, if we can promote BU to them and they make the switch as joined front we can start taking on core. What hurts XT is one person - Hearn being supreme leader.

Core has centralized we have not, once we do then Core will have to look over their shoulders as we make a run on number of nodes that are out there.

BU 1st has to tie down the code, then promote to BU to guys like Oliver Jansen and let them work on XT users after that we take on Core. We have to centralize to have the code decentralize Bitcoin.

We need well spoken knowledgeable debates no trolling on users including ones who are here.
  • Like
Reactions: soullyG


New Member
Aug 28, 2015
Agreed, XT may not be dead yet. But I still feel that the greatest disaster of that campaign was releasing XT with undocumented functions - or at least poorly documented ones. Underestimating the vehemence of the Core devs' opposition was another problem - but that only became apparent after the fact. At least with BU we can understand that, and deal with it from the start.

An important lesson for BU development is to limit its features to only those absolutely required. The temptation to pop in various other 'fixes' ( and bitcoin could do with a few) must be controlled because they can lead to PR disasters.


Staff member
Dec 16, 2015
I think one of the key problems with XT is that it seems stagnant at the moment. If there are initiatives underway to promote it they seem to be rather low-key, and there is no clear leadership. It seems Hearn has other commitments, and Gavin has not really announced much, so as far as I can see it's up in the air. Jonathan Toomim is actively doing good work establishing facts, but unfortunately he does not consider BIP101 to be a viable contender for miner consensus.

BU has the strong advantage that it's configurable to what miners want to set it, and it's got a lot of activity going on. It needs to sustain this, esp. in the areas of testing, review and promotion of its basic idea.
  • Like
Reactions: Aquent and Peter R

Zangelbert Bingledack

Well-Known Member
Aug 29, 2015
I wouldn't write off XT just yet. Recall that part of the goal with BU is to get other implementations to BU-ize themselves by making any controversial parameters user-selectable: Core Unlimited, XT Unlimited, btcd Unlimited, etc. I think BU's goal is not to be the next Core, but to be just another variant among the popular implementations - their popularity springing from them not attempting to force policy on the user.

Once XT chooses to merely *default* to BIP101 scaling and let the user customize it BU-style, I would say it becomes a key player again as long as Mike/Gavin or Jeff want to keep working on it. They're doing some promising things with thin blocks, for example. Under the BU paradigm, BIP101 is merely a suggestion, not a point of contention with the node client devs as it isn't their place to decide.
  • Like
Reactions: Mengerian