Quote from: smooth on August 14, 2015, 08:47:07 AM

Quote from: smoothie on August 14, 2015, 08:40:52 AM

Quote from: smooth on August 14, 2015, 08:14:47 AM

Quote from: smoothie on August 14, 2015, 08:05:28 AM

Quote from: smooth on August 14, 2015, 08:02:09 AM

Quote from: smoothie on August 14, 2015, 07:57:11 AM

Quote from: smooth on August 14, 2015, 07:52:42 AM

Quote from: smoothie on August 14, 2015, 07:49:15 AM

Quote from: smooth on August 14, 2015, 07:42:12 AM

Quote from: smoothie on August 14, 2015, 07:34:13 AM

Did you think Satoshi put that limit in at 1MB to be there forever? Huh

No, I suspect he meant to add other mechanisms to control in-block spam in its place (or at least I might say, hoped to, but may not have known how to do it), but never got to it and no one else picked up that work either. Now we are kind of screwed.

How are we screwed if we have time to try to find solutions?

I meant screwed on the assumption that the 1 MB is about to fill up and cause problems.

But has anyone considered that a good solution may not even exist? Bitcoin is a new thing that has never been done before and is not a law of nature that Bitcoin must "work" in general (for example at larger scale).

In my experience in building software, often problems that get the kick-the-can treatment don't end up having good solutions later either. Maybe 1MB was satoshi kicking-the-can?

I'm not claiming this, just asking the question.

Well is good the same as perfect?

Certainly not

Quote

I believe there is a good solution but not a perfect one. Sounds like the solution may be a moving target (dynamic in nature).

I hope you are right but I don't really know if there is.

Most of these ideas that get thrown out are stuff that seems like a good idea but nobody can make a rigorous argument why they won't break.

Did you have any thoughts after reading Peter R's paper? It seemed to describe the major issues as they are in their current form.

Lots of math and tons of variables to reference. I found myself going back to the definition of VARs page constantly to follow his logic/math/etc.

It answered the question of whether a fee market will collapse under certain conditions, with some simplifying assumptions about the network. I don't believe it answered (or even asked) the questions surrounding decentralization at all. You can have a fee market (in fact it is the most obvious case of one) with a single node serving the entire world. Obviously that is not a desirable outcome and nothing in his paper addresses it (nor tries to, which is fine -- that what's what papers do, they don't need to answer every question).

We already have a fee market.

But it isn't one that is very competitive yet.

Basically do we keep 1 MB cap and create a competitive fee market or raise the limit with some mechanism and eliminate the need for a competitive fee market while not turning users away and being forced to use intermediaries as payment channels.

Yes that is the question of course. My question has always been if we get rid of the 1 MB what do we do about spam control. One answer is "don't worry be happy" arguing, in effect, that the original cap was unnecessary. I'm not sure I buy it.

For whatever reason, you seem to keep ignoring the very answers to your fears that I and guys like Peter have been detailing to you in excruciating detail over and over again.

As long as miners and nodes have the freedom to set what size TX's they will validate and relay, which they do, they can protect their mempools from crashing. As long as miners are free to set their own block sizes, which they are, they can defend themselves from spam, however they want to define it. As long as users have the ability to outbid a spammers fee, which they will in a no limit situation, they can choose to either do so or walk away. As long as core dev can't be involved in setting limits, which they won't, then they can't be involved in picking winners and losers,which is good.