Bitcoin Forum
May 03, 2024, 12:21:47 AM *
News: Latest Bitcoin Core release: 27.0 [Torrent]
 
   Home   Help Search Login Register More  
Pages: « 1 [2]  All
  Print  
Author Topic: bitcoin/application/user-program supporting own scripts available?  (Read 2838 times)
Peter Todd
Legendary
*
expert
Offline Offline

Activity: 1120
Merit: 1150


View Profile
January 06, 2013, 11:33:39 PM
 #21

Not just 'not very important' but would be actually against the goals. You pointed out some of the risks, so I know you get it, but it deserves to be emphasized: Tight time requirements would provide no value to the Bitcoin currency but would in fact introduce substantial centeralization as everyone would become dependant on centralized sources of time in order to reliably meet the rules. Manipulation or disruption of these time sources would be devastating.

Agreed. A time window of 4 hours would have been quite reasonable too, although of course it's not worth it to change things now.

On a lark I did write up a sketch of a fanciful way of doing a decentralized time service using PoW chain consensus, you might find it amusing: https://people.xiph.org/~greg/decentralized-time.txt

Ah, yeah I saw that earlier. Fascinating, if unfortunately impractical!

1714695707
Hero Member
*
Offline Offline

Posts: 1714695707

View Profile Personal Message (Offline)

Ignore
1714695707
Reply with quote  #2

1714695707
Report to moderator
1714695707
Hero Member
*
Offline Offline

Posts: 1714695707

View Profile Personal Message (Offline)

Ignore
1714695707
Reply with quote  #2

1714695707
Report to moderator
1714695707
Hero Member
*
Offline Offline

Posts: 1714695707

View Profile Personal Message (Offline)

Ignore
1714695707
Reply with quote  #2

1714695707
Report to moderator
Activity + Trust + Earned Merit == The Most Recognized Users on Bitcointalk
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
1714695707
Hero Member
*
Offline Offline

Posts: 1714695707

View Profile Personal Message (Offline)

Ignore
1714695707
Reply with quote  #2

1714695707
Report to moderator
1714695707
Hero Member
*
Offline Offline

Posts: 1714695707

View Profile Personal Message (Offline)

Ignore
1714695707
Reply with quote  #2

1714695707
Report to moderator
smtp (OP)
Member
**
Offline Offline

Activity: 70
Merit: 10


View Profile
January 07, 2013, 12:13:58 AM
 #22

But he designed something intellectually consistent: abstract block-time, completely self-contained and discoverable by looking only inside the blockchain and only backward.
Well, a model/design which have dozends, probably hundreds of arbitrary chosen parameters as closer you look turns out, I should praise/appriciate?
You are joking. :-/  I never believe Satoshi had aimed at / wanted such a chaos/complexity which now is realized.

Quote
Please post something that shows that you understand how time-invariance of blockchain verification contributed to the security of the Bitcoin protocol.
You miss the point. I claim a practical "same security level" can be achieved without allowing 2h - wrong time-stamps, believing in say 30 sec inprecisions.

... as everyone would become dependant on centralized sources of time.
Nonsense. It seems you have no knowledge how cheap/expensive and precise simple good quartz-clocks can be today.

BTW: IF bitcoin should ever get more in the wide-spread scale of a common real-world currency this average 10 minutes time-distance must be very significantly lowered and probably other practical timings narrowed. Else many typical purposes can not use it and bitcoin will remain an unimportant/small corner for most people - net propagation time will decrease with the years. One day, you will pay with your master-card and could this be in BTCs or are BTCs forgoten because their transactions would have had to be waited for many minutes to be confirmed?

smtp
2112
Legendary
*
Offline Offline

Activity: 2128
Merit: 1065



View Profile
January 07, 2013, 01:46:39 AM
 #23

But he designed something intellectually consistent: abstract block-time, completely self-contained and discoverable by looking only inside the blockchain and only backward.
Well, a model/design which have dozends, probably hundreds of arbitrary chosen parameters as closer you look turns out, I should praise/appriciate?
You are joking. :-/  I never believe Satoshi had aimed at / wanted such a chaos/complexity which now is realized.
Quote
Please post something that shows that you understand how time-invariance of blockchain verification contributed to the security of the Bitcoin protocol.
You miss the point. I claim a practical "same security level" can be achieved without allowing 2h - wrong time-stamps, believing in say 30 sec inprecisions.
OK, I think I now understand our miscommunication.

I don't disagree that the "same security level" can be achieved with better time accuracy. I agree on this point.

What you seem to be missing from my point is the mathematical beauty of the "abstract time measured by the proof of work". The rules of Bitcoin have a beautiful property of statistical resistance to all kinds of cheating unless done by the 50%-or-more majority. No minority group can subvert this mechanism by coordinated cheating with timestamps, neither by moving block time forward nor backward.

So the single technical mechanism guards both honesty against double spending and against false timestamping. You may think that this is too complex; but I think it is mathematically minimal. In my opinion what you propose makes the Bitcoin more mathematically complex with unclear financial benefit.

Everyone here knows by heart that Satoshi considered transaction "confirmed" when it is 6 blocks deep in the blockchain.

Can you see the beauty of the fact that it takes 6 blocks to recognize that the time-on-the-blockchain moved forward? 6 blocks is measured statistically as a median-of-last-11-blocks. Anything less than 6-out-of-11 blocks wouldn't move the time forward! If it wasn't a median-of-11 but say average-of-11 then the time would move by an equivalent of increasing transaction confirmations by a fractional number.

6 may be an arbitrary number. 11 only looks arbitrary. It is actually dual to 6 by the duality of blocktime vs. blockheight. So if somebody proposes to change 6 without the corresponding change to 11, you now know that that person didn't really understand Bitcoin.

Yes, Satoshi could write this explicitly in his paper. But he probably set such a trap to allow himself to quickly grade the proposed changes to Bitcoin. So be aware of this when you making your own proposals.

Please comment, critique, criticize or ridicule BIP 2112: https://bitcointalk.org/index.php?topic=54382.0
Long-term mining prognosis: https://bitcointalk.org/index.php?topic=91101.0
gmaxwell
Moderator
Legendary
*
expert
Offline Offline

Activity: 4158
Merit: 8382



View Profile WWW
January 07, 2013, 01:49:49 AM
 #24

BTW: IF bitcoin should ever get more in the wide-spread scale of a common real-world currency this average 10 minutes time-distance must be very significantly lowered and probably other practical timings narrowed. Else many typical purposes can not use it and bitcoin will remain an unimportant/small corner for most people - net propagation time will decrease with the years. One day, you will pay with your master-card and could this be in BTCs or are BTCs forgoten because their transactions would have had to be waited for many minutes to be confirmed?

Nonsense, many common payment mechanisms have horizons of _months_ before transactions become irreversible.  Moreover, you can't simply reduce the time as though it were a free parameter.  As the time between blocks falls below the time it takes nodes to globally communicate and validate new blocks the time until convergence tends to infinity. 10 minutes may have been more conservative than needed but it cannot be made arbitrarily low in a universe with a finite speed of light.

It seems to me that like many of the more aggressively opinionated newbies that show up you are both clueless about the bitcoin system and the requirements of payment systems, and are just going to waste the time of anyone who bothers reading your posts. **Plonk**
smtp (OP)
Member
**
Offline Offline

Activity: 70
Merit: 10


View Profile
January 07, 2013, 11:07:48 AM
 #25

Nonsense, many common payment mechanisms have horizons of _months_ before transactions become irreversible.  Moreover, you can't simply reduce the time as though it were a free parameter.  As the time between blocks falls below the time it takes nodes to globally communicate and validate new blocks the time until convergence tends to infinity. 10 minutes may have been more conservative than needed but it cannot be made arbitrarily low in a universe with a finite speed of light.
Whom want you to impress with such a beginner attitude stated in your last sentence? Almost all transaction in real life are much quicker than 10 minutes. Go to a train stop and image in 50 year use an  electronic card when go into the train to pay. Or even today if you are in the shop and want to pay with your master-card. These TYPICAL transactions should be last say 10-15 sec AT VERY MOST. Therefore the time limit which is to be aimed at should be say 1 sec -- of course we have some decades time to reach it. If BTC-money payment only can go down to say 1 min, far more than 99 % of transactions in real-world usage will need another currency which is then of course of much higher importance, and then I see no sense to support BTC anymore myself -- it is (will stay) only a specialized currency in a small corner in the real world.
Thus time needs to be like a free parameter (in the model) which over the years has to be decreased -- as I already indicated in relation to net-propagation speed and size/frequency of total usage of BTC. Sorry, but some people are obviously unable to have/imagine a great future-plan and already lay barriers in the beginning which prevent the possibility to have only a real chance to get one good world-wide paying system/currency.

It seems to be that many people' error here in the community is: you need external time service (for closer timings / global faster transactions) -- only because in standard PC built-in time-hardware has too low precision.

smtp
smtp (OP)
Member
**
Offline Offline

Activity: 70
Merit: 10


View Profile
January 07, 2013, 11:59:19 AM
 #26

What you seem to be missing from my point is the mathematical beauty of the "abstract time measured by the proof of work". The rules of Bitcoin have a beautiful property of statistical resistance to all kinds of cheating unless done by the 50%-or-more majority. No minority group can subvert this mechanism by coordinated cheating with timestamps, neither by moving block time forward nor backward.
This abstract, mathematical model I appreciate greatly, indeed since more than 1.5 years I got knowledge of it - I may add I don't know all details of the model, but its projected aim looks convincing to me. But to use it in reality efficiently, you need to define (better) time(-distances) in it. Time rules reality -- and to give a more known economic proverb for our case: Time is money! Wink

smtp
DeathAndTaxes
Donator
Legendary
*
Offline Offline

Activity: 1218
Merit: 1079


Gerald Davis


View Profile
January 07, 2013, 01:39:38 PM
 #27

Quote
Nonsense, many common payment mechanisms have horizons of _months_ before transactions become irreversible.  Moreover, you can't simply reduce the time as though it were a free parameter.  As the time between blocks falls below the time it takes nodes to globally communicate and validate new blocks the time until convergence tends to infinity. 10 minutes may have been more conservative than needed but it cannot be made arbitrarily low in a universe with a finite speed of light.

Whom want you to impress with such a beginner attitude stated in your last sentence? Almost all transaction in real life are much quicker than 10 minutes. Go to a train stop and image in 50 year use an  electronic card when go into the train to pay. Or even today if you are in the shop and want to pay with your master-card. These TYPICAL transactions should be last say 10-15 sec AT VERY MOST. Therefore the time limit which is to be aimed at should be say 1 sec -- of course we have some decades time to reach it. If BTC-money payment only can go down to say 1 min, far more than 99 % of transactions in real-world usage will need another currency which is then of course of much higher importance, and then I see no sense to support BTC anymore myself -- it is (will stay) only a specialized currency in a small corner in the real world.

Bitcoin transactions are processed in a few seconds just like credit cards.  You will not the keyword above was irreversible.  It takes up to 180 days before CC transactions are irreversible.  On that scale 10-60 minutes doesn't seem excessively long.  It is certainly possible to accept 0-confirm transactions.  For physical world, and low value transactions that is likely all that is necessary.

Also remember that the entire network needs to synchronize.  That means propogation delays as well as verification delays need to be considered.  Due to the nature of Bitcoin (verify -> relay -> verify -> relay -> verify -> relay ... etc) it may require multiple "hops" before all nodes have access to the same block.   As gmaxwell pointed out 10 minutes is likely somewhat conservative one can certainly go lower but there is a cost and risk with going too low.   Anything below 1 minute is likely not viable without excessively high rate of blockchain forks and orphaned blocks.  

Lets compare some transaction times for settlement and funds availability

bitcoin 0confirm - ~ 1 sec
bitcoin 1confirm - ~10 minutes (varies 6 to 30 min within 1 SD)
bitcoin 6confirm -  ~1 hour

ACH (including "billpay") - 72 to 120 hours
Bank Wire (domestic) - 4 to 12 hours
Bank Wire (international) - 24 to 72 hours
Western Union - 15 minutes (unless delayed due to fraud prevention)
CC (inducing CC based systems) unconfirmed - ~1 to 15 sec
CC (inducing CC based systems) confirmed - 2880 to 4320 hours

Costs:
Bitcoin < $0.01 per tx
ACH ~$0.10 to $0.35 per tx
Bank Wire ~$10.00 to $35 per tx
Western Union ~7% to 10% of gross settlement
Credit Cards ~2% to 3% of gross settlement
Debit Cards ~1.5% of gross settlement

Bitcoins unconfirmed and confirmed settlement times are as good or better than other payment networks and at a fraction of the cost.



Still it seems you aren't really interested in learning or understanding so ....  Yup bitcoin sucks and everything about it was chosen to annoy you.  You likely should click logout and ignore it for another decade or two.
deepceleron
Legendary
*
Offline Offline

Activity: 1512
Merit: 1028



View Profile WWW
January 12, 2013, 10:13:42 AM
 #28

Who's this "smtp"? It took one week from asking "what is this 'Satoshi code'" and not knowing about GPU mining, to bashing Bitcoin design and confusing core developers with noobs, while going right into the scripting language of Bitcoin for no reason related to transferring currency. Kind of like how Mohamed Atta didn't care about learning to land planes...

I'm going to guess Chinese from the style of ESL writing.

I thought I'd alert those who have responded to this thread that he has also made over 100 edits to the script Wiki (along with other pages), so the community has a chance to review his changes.

https://en.bitcoin.it/w/index.php?title=Script&diff=34921&oldid=34415

gmaxwell
Moderator
Legendary
*
expert
Offline Offline

Activity: 4158
Merit: 8382



View Profile WWW
January 13, 2013, 01:56:10 AM
 #29

I thought I'd alert those who have responded to this thread that he has also made over 100 edits to the script Wiki (along with other pages), so the community has a chance to review his changes.
Good catch. I've reverted— as in a quick review removed a bunch of useful information (e.g. removing the sizes of the returned types).

Chinese you say? Basted on the attitude, cluelessness, and wiki editing volume I would have guessed it was atlas.


smtp (OP)
Member
**
Offline Offline

Activity: 70
Merit: 10


View Profile
January 16, 2013, 07:36:02 PM
 #30

I thought I'd alert those who have responded to this thread that he has also made over 100 edits to the script Wiki (along with other pages), so the community has a chance to review his changes.
Good catch. I've reverted— as in a quick review removed a bunch of useful information (e.g. removing the sizes of the returned types).

Chinese you say? Basted on the attitude, cluelessness, and wiki editing volume I would have guessed it was atlas.

Surprisingly I read this by chance, when wanting to improve a bit the URL https://en.bitcoin.it/wiki/Script .
Your change,
(cur | prev) 2013-01-13T01:52:33‎ Gmaxwell (Talk | contribs)‎ . . (16,312 bytes) (-34,010)‎ . . (Blind reversions of edits by SMTP, see this thread: https://bitcointalk.org/index.php?topic=134843.msg1449369#msg1449369) (undo)
would be judged simply as vandalism on en.wikipedia.org, gmaxwell!

He removed much more, new, helpful information, with the only given reason I may/could have removed some hidden details (which I was not aware because I did not need them when writing an interpreter for this script-language!)? -- too bad for you. :-(  I personally suspect you have had private/emotional reasons to have acted so. I believe this bitcointalk-community is not worth my support anymore - too many guys who can't think analytically precise and rational reasoned is my feeling since being inside.
I make my decision to depart here only on gmaxwell's behaviour on https://en.bitcoin.it/w/index.php?title=Script&action=history, yes. A bit unfair probably,
but I judge my leisure time to be as too worthy to waste it in an edit-war with unable editor/author/lector, who is obviously dislike to keep objectively or improve articles like me!

BTW: It seems to me, that Gmaxwell is more clueless of what the op_codes precisely do than me who wrote a much preciser and more complete explanation (even with historical references) of the op_codes in the wiki-page after studying the bitcoin-source-script.cpp in detail - sorry, to call you publicly the bad guy with these lines. :-/

Fare well,
smtp
gmaxwell
Moderator
Legendary
*
expert
Offline Offline

Activity: 4158
Merit: 8382



View Profile WWW
January 17, 2013, 06:20:50 AM
 #31

He removed much more, new, helpful information, with the only given reason I may/could have removed some hidden details
You removed important information as I _specifically_ described here, and I didn't see anything important added. By commenting here shortly after I made sure that you wouldn't miss the change. You made _103_ consecutive overlapping commits to the the script page but change very little besides removing information with the result of the page being less complete and accurate.  Improving it is good, but not at the expense of accuracy.

One of the edits you've made on the wiki was so deeply incorrect— editing in direct opposition to the clear meaning of the pages— that I wondered if you were being intentionally wrong.   Your interest in Bitcoin is certainly welcome, but your ignorance of your ignorance and your disrespect towards others is unlikely to result in useful cooperation.

I do not oppose you continuing to participate here and on the Bitcoin wiki— but I think you will find more satisfaction if you participate with a softer hand.
Pages: « 1 [2]  All
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!