[bitcoin-dev] Updating the Scaling Roadmap

Gregory Maxwell greg at xiph.org
Wed Jul 12 00:07:51 UTC 2017


On Tue, Jul 11, 2017 at 10:17 PM, Paul Sztorc <truthcoin at gmail.com> wrote:
> I don't understand this at all. This document attempts to do exactly
> what its predecessor did -- nothing more or less.

That might be your impression, then you've misunderstood what I
intended-- What I wrote was carefully constructed as a personal view
of how things might work out. It never claimed to be a a project
roadmap. Though as usual, I work hard to propose things that I believe
will be successful... and people are free to adopt what they want.

And to the extent that it got taken that way I think it was an error
and that it has harmed progress in our community; and created more
confusion about control vs collaboration.

With perfect hindsight I wouldn't have posted it; especially since
we've learned that the demand for increased capacity from many people
was potentially less than completely earnest. (The whole, can't double
capacity until we quadruple it thing...)

> As to your specific complaints, I have addressed both the security model
and the concept of mining centralization on this list in the recent
past.

I don't agree that you have; but for the purpose of this thread
doesn't really matter if I (specifically) do or don't agree.  It's an
objective truth that many people do not yet believe these concerns
have been addressed.

> I really don't understand what you are disclosing. That Adam asked you
> for feedback on the draft? And then, in the next sentence, that not

That Adam asked me to write publish a new "roadmap" for Bitcoin as
you've done here, with particular features and descriptions, which I
declined; and explained why I didn't believe that was the right
approach.  And Adam worked with you on the document, and provided
content for it (which I then recognized in the post).

Set aside what you know to be true for a moment and consider how this
might look to an outsider who found out about it.  It could look a
like Blockstream was trying to influence the direction of Bitcoin by
laundering proposals through an apparently unaffiliated third party.
Doubly so considering who participated in your drafting and who didn't
(more below).

I don't think in actuality you did anything remotely improper
(ill-advised, perhaps, since your goal to speak for developers but you
didn't speak to them, IMO--) but I think transparency is essential to
avoid any appearance of misconduct.

> But surely you can
> appreciate how bizarre your position on roadmaps is. What exactly, did
> you intended to create at [1]? Since it is described explicitly as "the
> roadmap in Capacity increases for the Bitcoin system", have you been
> disagreeing with it's characterization as a 'roadmap' this entire time?
> One wonders why you haven't said anything until now.

I did--

As Bryan pointed out... with the exception of the intro and end and a
couple sentences in the middle my entire response to your post, with
the position on "roadmaps" was written a long time ago, specifically
to complain about and argue against that particular approach.

> In my first email I list the benefits of having a roadmap. One benefit
> is that, without one, it is likely that a large majority of outsiders
> have almost no idea at all what is being worked on, what effect it will
> have, or when it might be ready, or to whom/what they should turn to for
> advice on such matters. Do you have a different way of addressing this
> communication problem?

I think you may be mistaking a roadmap with "communications"-- people
taking about research they are exploring is great! and we should have
more of it.  But like with RedHat and kernel features, we can't really
roadmap things that are unresourced and currently just unimplemented
exploration or test code-- without seeing things which are more or
less done the community can't rightfully decide if they'd want to
support something or not.  Perhaps they'll be good things perhaps
awful-- the devil is in the details, and until an effort is fairly
mature, you cannot see the details.

There have, for example, been signature aggregation proposals in the
past that required address reuse (could only aggregate if they're
reused).  I would strongly oppose such a proposal, and I hope everyone
else here would too.  So if I were a third party looking at your
message, rather than the person who initially proposed the agg sig
thing you're talking about, I wouldn't know if I could love it or hate
it yet; and probably couldn't be expected to have much of an opinion
on it until there is a BIP and/or example implementation.

To the extent that a roadmap differs from communications in general,
it is in that it also implies things that none of us can promise
except for our own efforts; Completion of implementations, success of
experiments, adoption-- basically assuming a level of top down control
that doesn't exist in a wide public collaboration.

One of the great challenges in our industry is that we don't have
rings of communication: We don't have much in the way of semi-experts
to communicate to semi-semi-experts to communicate to media pundits to
communicate to the unwashed masses-- at each level closing the
inferential gap and explaining things in terms the audience
understands. We don't have things like LWN.   We'll get there, but it
it took the Linux world decades to build to what it has now. I think
various forces in the industry work against us-- e.g. we lose a mot of
mid tier technical people to the allure of striking it rich printing
money in their own altcoins.

It might be attractive to try to end-run the slow development
appropriate web of reliable communications by deploying high authority
edicts, but it ultimately can't work because it doesn't map to how
things are accomplished, not in true collaborative open source, and
certainly not in an effort whos value comes substantially from
decentralization. Doing so, I fear, creates a false understanding of
authority.

(It's a common misunderstanding, for example, that people do what I
want (for example); but in reality, I just try to avoid wasting my
time advocating things that I don't think other people are already
going to do; :) otherwise, if _I_ want something done I've got to do
it myself or horse trade for it, just like anyone else.)

One of the great things about general communications is anyone can do
it.  Of course, unless they're talking about their own work, they
can't promise any of it will be completely-- but that is always true.
 If someone wants some guarantee about work, they have to be willing
to get it done themselves (and, of course, if it's a consensus
feature-- that much is necessary, but still not sufficient to get a
guarantee).

[from your other reply]
>> A fine intention, but I've checked with many of the top contributors
>> and it sounds like the only regular developer you spoke with was
>> Luke-Jr.  Next time you seek to represent someone you might want to
>> try talking to them!
>
> That is false. I could provide a list of names but I'm really not sure
> what would be gained as result. You yourself admit that it is an
> excellent list of research, almost all which you support directly.
>
> So I think your only real objection is that I didn't talk to you
> specifically.

Come now, this is needlessly insulting. I would have made the same
comment if you had talked to me because you didn't talk to most/all of
Matt Corallo, Wladimir, Pieter Wuille, Alex Morcos, etc.... e.g. the
people doing most of the work of actually building the system.  Before
making that comment I went and checked with people to find out if only
I was left out.  Talking to Adam (who isn't involved in the project)
and Luke-jr (who is but is well known for frustratingly extreme
minority positions and also contracts for Blockstream sometimes) isn't
a great approach for the stated goal of speaking for developers!


More information about the bitcoin-dev mailing list