Yet another analyst :)

[lucif]

1. Very unlikely to happen, if they actually are able to produce working hardware they would be much better off continuing to do so.
2. Not going to happen, GPU miners will quit after they start become unprofitable not before it, hashrate would only go up.
3. Sure why not? But I'd not be too concerned with a selloff but with a permanent image problem alas a long grind down.
4. That would be stupid and even discussing it is stupid, overall just a stupid point.

I see points 1 and 3 real. Point 4 is a joke of course. But every joke has bit of thruth.

[notme]

Another portion of pessimism

1. Bitcoin has a bottleneck with unconfirmed transactions network buffer, firstly pointed by Microsoft research team. It can easily paralyze network on high load
2. Bitcoin has a bottleneck  with its database size and sync process. Vanilla client is almost unusable for common non-technicial Internet user.

If problem 2 may be solved by futher Electrum server development and deployment of such scheme, the problem 1 persists.

If it were so easy don't you think it would be attempted?  They invented a problem that doesn't seem to actually exist just so they could "fix" bitcoin.  Yes, I have read and understood the paper.

[lucif]

1. Very unlikely to happen, if they actually are able to produce working hardware they would be much better off continuing to do so.

Pls explain me why to sell chicken which brings golden eggs?

[ElectricMucus]

Another portion of pessimism

1. Bitcoin has a bottleneck with unconfirmed transactions network buffer, firstly pointed by Microsoft research team. It can easily paralyze network on high load
2. Bitcoin has a bottleneck  with its database size and sync process. Vanilla client is almost unusable for common non-technicial Internet user.

If problem 2 may be solved by futher Electrum server development and deployment of such scheme, the problem 1 persists.

Everybody knows about this issue and acknowledges it right now and until the block reward becomes irrelevant it isn't rally any issue. Bitcoin was never made to last that long only a few years. Satoshi said it.

I see points 1 and 3 real. Point 4 is a joke of course. But every joke has bit of thruth.

1. Real yes, but again unlikely. Unlikely enough not to worry about.
3. We agreed already it's entirely possible.

[ElectricMucus]

1. Very unlikely to happen, if they actually are able to produce working hardware they would be much better off continuing to do so.

Pls explain me why to sell chicken which brings golden eggs?

Because the eggs spoil very quickly but people will always buy chickens.

[lucif]

If it were so easy don't you think it would be attempted?  They invented a problem that doesn't seem to actually exist just so they could "fix" bitcoin.  Yes, I have read and understood the paper.

Just come here and press Uncofirmed transactions. Wait, drink a cofee, smoke a sig. And see Gateway timeout This wasnt in past. What next? If bitcoin grow to some average payment system size, number of broadcasted transactions would be overwhelming.

It may not overwhelm power pools servers, but will overwhelm regular clients. We go to centralized model with that stuff not fixed.

[notme]

1. Very unlikely to happen, if they actually are able to produce working hardware they would be much better off continuing to do so.

Pls explain me why to sell chicken which brings golden eggs?

Because during a gold rush there are very few miners/panners who strike a nice gold vein, but everyone who sells shovels and blue jeans gets rich.  Most miners barely scrape by and eventually give up.

[lucif]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

[ElectricMucus]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

I see 3 linear curves, if the overall thing tends to become exponential it has to be weighted against advances in hard-drive capacity and communication bandwidth both of which increase in a exponential fashion.

[lucif]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

I see 3 linear curves, if the overall thing tends to become exponential it has to be weighted against advances in hard-drive capacity and communication bandwidth both of which increase in a exponential fashion.

Yup. But what about usability of that for regular user?

[lucif]

Because during a gold rush there are very few miners/panners who strike a nice gold vein, but everyone who sells shovels and blue jeans gets rich.  Most miners barely scrape by and eventually give up.

Fork trap here if asics not a scam.

If bitcoin price goes to the heaven - only idiot will sell own ASIC technology which may bring him >50% of overall reward. Then we'll have a danger of 51% attack and most power located in few hands => prices go to hell.

If bitcoin price goes to the hell - they will sell technology. But prices go to hell.

The only way for bitcoin price to survive is to stay stable sideways for all priod of asic adoption.

[ElectricMucus]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

I see 3 linear curves, if the overall thing tends to become exponential it has to be weighted against advances in hard-drive capacity and communication bandwidth both of which increase in a exponential fashion.

Yup. But what about usability of that for regular user?

I'm not sure. I don't think you can say for certain either. I don't think it will become relevant for about lets say 20 years.

[lucif]

That is the point. Bitcoin is complex software for common user understanding. People will not buy potatoes for bitcoin in near future. Only hard technicials and criminals close their eyes for mentioned bottlenecks and understands the technology.

They agree to wait initial sync for two days (and growing). I dont count chain download via torrent - its centralized stuff and another type of brainfuck.
They agree to wait a sync when they didnt use vanilla client for a long time.
They agree that bitcoin client stucks their hdd and its almost impossible to do regular work on computer during this.
They agree to research what difference between confirmed/unconfirmed trans.
They agree that bitcoin consumes large resouces on their pc.

Some of them agree with that because of geek fanatism. Some because there is no legal way to offer their services.

But for regular user decentralized bitcoin client is just a chineese letter. Only centralized services are better for understanding - but they do regular scam. If they dont do scam - they will be closed for money laundering (hello, MtGox!). The most attractive news about Bitcoin are related to scam and illegal use (see google trends). They bring a stereotype that Bitcoin is something related to scam and illegal stuff.

That is what Bitcoin for regular user: something complex, hard, illegal and scam-related.

Sad but true. I think market is waiting for big disappointment, as many ppl think that is GOLD20, replacement for cash, etc, etc. But Bitcoin is initially ILLEGAL coz of its anonymous nature. It will be always a heaven for criminals and scammers. And this is the true niche for this technology. And the road of understanding this lays via big disappointment.

IMHO.

[lucif]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

I see 3 linear curves, if the overall thing tends to become exponential it has to be weighted against advances in hard-drive capacity and communication bandwidth both of which increase in a exponential fashion.

The second part of my person is expirenced Unix sysadmin.

I saw many crying clients, who were cry that there is enough space in hdd to keep video content, but server is hanged coz of heavy io.

HDD has two parameters: space and bus throughput.

Home users don't take into account second one. Never.

You know what is database double-tripple bigger then RAM size under read/write load? It is a death for IO scheduler if there is no good array of 15k rpm disks.

And even with array...100G database is really hell for a home PC even with 100TB single HDD.

So I ask one more time: where this curve leads to?

[lucif]

Just imagine Bitcoin network serving 100m customers.

New transaction arrives every second. Software have to verify its inputs by queriyng huge database. And every client do this job for every trans. Every transaction will cause millions of nodes to lookup bunch of keys in 100G database....

Also every client must keep this transaction buffer.

Well, let's wait that times.

[adamstgBit]

Just imagine Bitcoin network serving 100m customers.

New transaction arrives every second. Software have to verify its inputs by queriyng huge database. And every client do this job for every trans. Every transaction will cause millions of nodes to lookup bunch of keys in 100G database....

Also every client must keep this transaction buffer.

Well, let's wait that times.

"millions of nodes" 

buy buy buy, millions of nodes! Millions of nodes!!!

[lucif]

I remember the times when blockexplorer replied instantly on any query. Now it gives denial of service of about 30% of my queries. And that is what fate of every regular client in future if things will not change. And this future comes with exponental speed.

Disk space isn't problem - Satoshi was right. But disk IO is a problem. Big problem.

What I want to say with all of that? Bitcoin is beta and officially not recommended for serious investments by Gavin if I don't mistake. Too sad if I do. Too sad that press release on halving day from Bitcoin Foundation contains only squeal about Bitcoin price rise.

The big disappointment will reach fanatic believers if they loose connection with reality. See my sig.

[ElectricMucus]

Anyone say where this exponental curve leads to?

http://blockchain.info/ru/charts/blocks-size?timespan=all&showDataPoints=false&daysAverageString=1&show_header=true&scale=0&address=

I see 3 linear curves, if the overall thing tends to become exponential it has to be weighted against advances in hard-drive capacity and communication bandwidth both of which increase in a exponential fashion.

The second part of my person is expirenced Unix sysadmin.

I saw many crying clients, who were cry that there is enough space in hdd to keep video content, but server is hanged coz of heavy io.

HDD has two parameters: space and bus throughput.

Home users don't take into account second one. Never.

You know what is database double-tripple bigger then RAM size under read/write load? It is a death for IO scheduler if there is no good array of 15k rpm disks.

And even with array...100G database is really hell for a home PC even with 100TB single HDD.

So I ask one more time: where this curve leads to?

Triple redundant parity raid, for now later even higher redundancy. But that is true for any database with exponential growing size. The immediate solution is "ultraprune", simply that clients do not run the full database but rely on others to confirm transactions.
This issue is not unique to Bitcoin but is a known challenge for all IT. I don't know the answer but I think that eventually we are going to move away from hard-disks at all. Flash memory will in the short run archive enough bandwidth to do away with data degradation in time. In the long run there will be processor arrays which have their own non-volatile memory attached to each node.
I would expect bitcoin to be obsolete in it's current form by that time, which again will take about 20 years I think.

But that doesn't mean the blockchain couldn't be stored in multiple nodes communicating with error correcting codes. That's as far as I can imagine, everything after that is uncertain.

[notme]

Just imagine Bitcoin network serving 100m customers.

New transaction arrives every second. Software have to verify its inputs by queriyng huge database. And every client do this job for every trans. Every transaction will cause millions of nodes to lookup bunch of keys in 100G database....

Also every client must keep this transaction buffer.

Well, let's wait that times.

It's called an unspent transaction cache.  You don't have to query every transaction, just keep a running list of how much each nonempty address holds.

Yes, there are scalability issues to be addressed before we get to 100 GB, but they are being addressed and the software is improving constantly.

[myself]

HDD has two parameters: space and bus throughput.

People still use HDD? 

nah most ppl use this  http://www.ramsan.com/products/rackmount-flash-storage/ramsan-810