Author Topic: BitShares X Status Update  (Read 261034 times)

0 Members and 1 Guest are viewing this topic.

Offline alt

  • Hero Member
  • *****
  • Posts: 2821
    • View Profile
  • BitShares: baozi
hang again? reproduce:
1. short 1000usd , price 1000 usd/bts
2. buy 1000usd, price 0.0015 bts/usd


Ok, I have identified the problem that was hanging the blockchain market... more rounding issues.  The problem arises when I have to revalidate price calculations using rounded or truncated data.   There is no way around the fact that information has been lost and thus the validation step has to be 'fuzzy'.

My 'quick fix' was to increase the epsilon I use for comparing claim_by_cover outputs as the rounding error was 870 satoshi in this instance.  This means I have to choose between lower meaningful divisibility or switching over to 128 bit numbers for all balances.   I can then ignore the lower 64 bits for comparison purposes but still use them for price multiplication purposes.   This would add an average of 16 bytes per bid/ask and keep all of the rounding errors so small as to be inconsequential.   Perhaps 1000 satoshi is inconsequential for all practical purposes.   

The changes required to fix this problem will require all testers to recompile and break the blockchain.   

If you want to process the current blockchain you will have to update and recompile your client otherwise your client will fail to validate the hacked solution.   After Miami I will probably convert all amounts over to 128 bit numbers and just eat the bandwidth and storage costs to provide reliable and meaningful calculations.

Offline bytemaster

That has equal space and rounding issues.


Sent from my iPhone using Tapatalk
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline SlyWax

  • Jr. Member
  • **
  • Posts: 45
    • View Profile
Ok, I have identified the problem that was hanging the blockchain market... more rounding issues.  The problem arises when I have to revalidate price calculations using rounded or truncated data.   There is no way around the fact that information has been lost and thus the validation step has to be 'fuzzy'.

My 'quick fix' was to increase the epsilon I use for comparing claim_by_cover outputs as the rounding error was 870 satoshi in this instance.  This means I have to choose between lower meaningful divisibility or switching over to 128 bit numbers for all balances.   I can then ignore the lower 64 bits for comparison purposes but still use them for price multiplication purposes.   This would add an average of 16 bytes per bid/ask and keep all of the rounding errors so small as to be inconsequential.   Perhaps 1000 satoshi is inconsequential for all practical purposes.   

The changes required to fix this problem will require all testers to recompile and break the blockchain.   

If you want to process the current blockchain you will have to update and recompile your client otherwise your client will fail to validate the hacked solution.   After Miami I will probably convert all amounts over to 128 bit numbers and just eat the bandwidth and storage costs to provide reliable and meaningful calculations.

Why don't you use two numbers to represent a quotients ?
This way you won't have rounding errors, and no need to deal with them !
The calculus will just be made on the UI for display purpose.

Offline JetainM

  • Full Member
  • ***
  • Posts: 83
    • View Profile
I have downloaded and compiled the source codes of bitshares X, and it works on my computer now. I am glad to see the initial version of the BTS product! It is a nice work!

I want to ask how to further participate in the alpha testing. How to interact with the other BTS testers? And where should I send suggestions/bugs to?

So far I have the following suggestions:
1. to compile the source code, except for the specified required packages, readline package is also needed. On ubuntu 12.04, I need to install libreadline-dev to make the compilation complete.
2. it is not clearly written down on how to use the software after successful compilation. e.g. two binary files (bts_server, bts_wallet) are generated, but users do not understand what bts_server stands for unless reading the source code. A brief introduction on how to use the software is preferred.
3. in bts_wallet, I find it impossible to get my private key.
4. although I can generate as many new addresses as I want, I cannot view these addresses again from bts_wallet. Please advise on how to view these address if I want to use it again.

Until now I have not received any test BTS yet. my addresses: HD98kN9Y3Hhep7ysGa1drLdR7ft

The initial testing period is temporarily suspended until I fix a few bugs that have come up.  The next testing period will come with more robust instructions, binaries, etc. 

I would appreciate help from the community in documenting how to use the tools because there is a lot of work to be done and that will slow me down.

Thks a lot. It readline troubled me for several hours.Now i make it.
My bts addresses:MwqnmRTubvptyorR1F7kGPVqgWC
BIB==BM IS BUG

Offline aasl

  • Jr. Member
  • **
  • Posts: 45
    • View Profile
Ok, I have identified the problem that was hanging the blockchain market... more rounding issues.  The problem arises when I have to revalidate price calculations using rounded or truncated data.   There is no way around the fact that information has been lost and thus the validation step has to be 'fuzzy'.

My 'quick fix' was to increase the epsilon I use for comparing claim_by_cover outputs as the rounding error was 870 satoshi in this instance.  This means I have to choose between lower meaningful divisibility or switching over to 128 bit numbers for all balances.   I can then ignore the lower 64 bits for comparison purposes but still use them for price multiplication purposes.   This would add an average of 16 bytes per bid/ask and keep all of the rounding errors so small as to be inconsequential.   Perhaps 1000 satoshi is inconsequential for all practical purposes.   

The changes required to fix this problem will require all testers to recompile and break the blockchain.   

If you want to process the current blockchain you will have to update and recompile your client otherwise your client will fail to validate the hacked solution.   After Miami I will probably convert all amounts over to 128 bit numbers and just eat the bandwidth and storage costs to provide reliable and meaningful calculations.

Nice work!

The latest source requires berkeleydb c++ library. In ubuntu, I installed libdb++-dev to make the compilation pass.

My new BTS address is: SGspNwacutDySLyC8h4x1gMXNHD

Offline voldemort628

  • Full Member
  • ***
  • Posts: 117
    • View Profile
  My sources tell me that Nxt is starting to encourage clones like we will.

Hi Dan, I dont doubt the reliability of your sources, however there have been trolls targeting Nxt and all those so called clones are nothing but trolls haha. I have stakes in both bts and nxt, so i follow the thread quite closely ;)

Offline stuartcharles

  • Sr. Member
  • ****
  • Posts: 281
    • View Profile
Ok, I have identified the problem that was hanging the blockchain market... more rounding issues.  The problem arises when I have to revalidate price calculations using rounded or truncated data.   There is no way around the fact that information has been lost and thus the validation step has to be 'fuzzy'.

My 'quick fix' was to increase the epsilon I use for comparing claim_by_cover outputs as the rounding error was 870 satoshi in this instance.  This means I have to choose between lower meaningful divisibility or switching over to 128 bit numbers for all balances.   I can then ignore the lower 64 bits for comparison purposes but still use them for price multiplication purposes.   This would add an average of 16 bytes per bid/ask and keep all of the rounding errors so small as to be inconsequential.   Perhaps 1000 satoshi is inconsequential for all practical purposes.   

The changes required to fix this problem will require all testers to recompile and break the blockchain.   

If you want to process the current blockchain you will have to update and recompile your client otherwise your client will fail to validate the hacked solution.   After Miami I will probably convert all amounts over to 128 bit numbers and just eat the bandwidth and storage costs to provide reliable and meaningful calculations.


Thanks for all the updates and good luck at the conference

Offline bytemaster

Ok, I have identified the problem that was hanging the blockchain market... more rounding issues.  The problem arises when I have to revalidate price calculations using rounded or truncated data.   There is no way around the fact that information has been lost and thus the validation step has to be 'fuzzy'.

My 'quick fix' was to increase the epsilon I use for comparing claim_by_cover outputs as the rounding error was 870 satoshi in this instance.  This means I have to choose between lower meaningful divisibility or switching over to 128 bit numbers for all balances.   I can then ignore the lower 64 bits for comparison purposes but still use them for price multiplication purposes.   This would add an average of 16 bytes per bid/ask and keep all of the rounding errors so small as to be inconsequential.   Perhaps 1000 satoshi is inconsequential for all practical purposes.   

The changes required to fix this problem will require all testers to recompile and break the blockchain.   

If you want to process the current blockchain you will have to update and recompile your client otherwise your client will fail to validate the hacked solution.   After Miami I will probably convert all amounts over to 128 bit numbers and just eat the bandwidth and storage costs to provide reliable and meaningful calculations.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline bytemaster

I have downloaded and compiled the source codes of bitshares X, and it works on my computer now. I am glad to see the initial version of the BTS product! It is a nice work!

I want to ask how to further participate in the alpha testing. How to interact with the other BTS testers? And where should I send suggestions/bugs to?

So far I have the following suggestions:
1. to compile the source code, except for the specified required packages, readline package is also needed. On ubuntu 12.04, I need to install libreadline-dev to make the compilation complete.
2. it is not clearly written down on how to use the software after successful compilation. e.g. two binary files (bts_server, bts_wallet) are generated, but users do not understand what bts_server stands for unless reading the source code. A brief introduction on how to use the software is preferred.
3. in bts_wallet, I find it impossible to get my private key.
4. although I can generate as many new addresses as I want, I cannot view these addresses again from bts_wallet. Please advise on how to view these address if I want to use it again.

Until now I have not received any test BTS yet. my addresses: HD98kN9Y3Hhep7ysGa1drLdR7ft

The initial testing period is temporarily suspended until I fix a few bugs that have come up.  The next testing period will come with more robust instructions, binaries, etc. 

I would appreciate help from the community in documenting how to use the tools because there is a lot of work to be done and that will slow me down.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline bytemaster

Im new to bitshares but am very enthusiastic about its potential as well as the potential of subsequent DACs. I'd like to invest however Im uncertain about the relative development stages of bitshares and its future competitors mastercoin and nxt. I particularly like the transparency of this company in comparison to the development of the other teams. I think that the idea of limiting 16 assets to a block makes bitshares more scalable than mastercoin which might be trying to do to much with the bit coin blockchain itself. It is also a good selling point for investors in pts and ags because they get a stake in each block chain. So im pretty much sold on bitshares over mastercoin, although i do like the idea of overlaying protocols in the same way as the internet. But im not entirely sure how nxt and bitshares differ. Nxt seems to provide the same services as invictus. Since i don't fully understand the technical aspects of transparent forging i am convinced that its either a great innovation that helps with scalability or its an equally great flaw the threatens the security of the block chain. Mr. Larimer, I know that you have said that you do not want to provide estimates on the timeframe of development but could you at all speculate on your companies relative stage of development when compared to your competitors. Additionally, I would like some clarity on the networks ability to store value, the fundamental problem with bitcoin. If i have read correctly, in order to lend bitassets into existence the client must have twice as much collateral in bitshares. If that is the case, ones money only gains half the stability of whatever their preferred assets might be while still maintaining half the volatility of bitshares. I imagine that bitshares could be more stable than bitcoin but its price is still a rather unpredictable variable. If i have indeed misunderstood and the collateral is only held for short positions then how can you guarantee that each long is paired to a short as i believe i have seen you state?  Lastly this might be a stupid idea but in order to facilitate the adoption of bitshares is it possible to distinguish each new node so as to allocate bitshares amongst new users. I understand that you have switched to proof of stake and do not want to dilute ags and pts holders stake in the DAC  but for the purposes of adoption, particularly because of the difficulty  in acquiring bitcoin

Sent from my RM-917_nam_usa_100 using Tapatalk

Welcome to our forum.  I post daily updates in this thread regarding the status of our development.   I am glad you already see the scalability issues of everyone else attempting to do everything in a single chain.  My sources tell me that Nxt is starting to encourage clones like we will.

Here is my analysis of Nxt's proof-of-stake system.... it is very good, perhaps the best of all 'mining' coins.   Despite being proof-of-stake, Nxt still uses mining to decide who produces blocks and pays transaction fees to the miners.   As far as I can tell, only Ripple and BitShares pay dividends to 'shareholders' where as everyone else either dilutes share holders or pays 100% of revenue to miners.   

I consider Ripple style consensus to be superior to Nxt's approach for any blockchain that implements markets.  The reason for this is that mining centralizes control of block contents one block at a time and in the case of Nxt... that miner has a high probability of knowing when they will get to select transactions and thus they can gain a market advantage.   There is also block propagation delay.

Nxt assets are more like colored coins than decentralized market-pegged digital assets.   

If I were a betting man I would consider myself more of the Steve Jobs of the Crypto industry with a strong sense of the importance of ease of use and Brian Page will probably become know as one of the leading marketers in the Bitcoin space as his work bears fruit over the course of the next several months.

Your understanding of BitAssets and how their peg works is incomplete, but there are other threads in this forum that will answer your questions.  Effectively, if my understanding of economics is correct, BitUSD will fluctuate between +/-5% of the USD price depending upon relative flow into and out of the network.  Much more stable than you believe. 

As for our current development status, I would wager that we will have a minimal viable product out on the market prior to Mastercoin, Ethereum.  Our products are different in function than Nxt and others and thus these other coins do not provide the same service and utility.  In other words, they are selling cars and we are selling trucks.... while they both provide transportation they serve different markets.  I think our market is bigger ;)

One major difference between us and the competition is that we have a half dozen different DAC models and more on our roadmap.  Investing in AGS or PTS gives you so many ways to pivot where as investing in the competition gives you much less diversity.     

Compared to the competition I believe we are currently the 2nd best capitalized (Mastercoin has more money due to getting their BTC prior to the 10x gain)... but with AGS I expect that we will excede Mastercoin's funding level before the Beyond Bitcoin Conference we are putting on in July.

No one else is doing dividends but Invictus and dividends will make all the difference in the years ahead.
For the latest updates checkout my blog: http://bytemaster.bitshares.org
Anything said on these forums does not constitute an intent to create a legal obligation or contract between myself and anyone else.   These are merely my opinions and I reserve the right to change them at any time.

Offline aasl

  • Jr. Member
  • **
  • Posts: 45
    • View Profile
I have downloaded and compiled the source codes of bitshares X, and it works on my computer now. I am glad to see the initial version of the BTS product! It is a nice work!

I want to ask how to further participate in the alpha testing. How to interact with the other BTS testers? And where should I send suggestions/bugs to?

So far I have the following suggestions:
1. to compile the source code, except for the specified required packages, readline package is also needed. On ubuntu 12.04, I need to install libreadline-dev to make the compilation complete.
2. it is not clearly written down on how to use the software after successful compilation. e.g. two binary files (bts_server, bts_wallet) are generated, but users do not understand what bts_server stands for unless reading the source code. A brief introduction on how to use the software is preferred.
3. in bts_wallet, I find it impossible to get my private key.
4. although I can generate as many new addresses as I want, I cannot view these addresses again from bts_wallet. Please advise on how to view these address if I want to use it again.

Until now I have not received any test BTS yet. my addresses: HD98kN9Y3Hhep7ysGa1drLdR7ft

clout

  • Guest
And moreover the added difficulty and uncertainty in acquiring alt coins like pts. This promotional scheme would only be employed for a limited amount of time but would but both assuage concerns that bitshares is a ponzi scheme - i don't believe this but i have seen this argued- as well as help with adoption in a place like china where there is a fervent crypto community and even more fervently opposed government.

Sent from my RM-917_nam_usa_100 using Tapatalk

clout

  • Guest
Im new to bitshares but am very enthusiastic about its potential as well as the potential of subsequent DACs. I'd like to invest however Im uncertain about the relative development stages of bitshares and its future competitors mastercoin and nxt. I particularly like the transparency of this company in comparison to the development of the other teams. I think that the idea of limiting 16 assets to a block makes bitshares more scalable than mastercoin which might be trying to do to much with the bit coin blockchain itself. It is also a good selling point for investors in pts and ags because they get a stake in each block chain. So im pretty much sold on bitshares over mastercoin, although i do like the idea of overlaying protocols in the same way as the internet. But im not entirely sure how nxt and bitshares differ. Nxt seems to provide the same services as invictus. Since i don't fully understand the technical aspects of transparent forging i am convinced that its either a great innovation that helps with scalability or its an equally great flaw the threatens the security of the block chain. Mr. Larimer, I know that you have said that you do not want to provide estimates on the timeframe of development but could you at all speculate on your companies relative stage of development when compared to your competitors. Additionally, I would like some clarity on the networks ability to store value, the fundamental problem with bitcoin. If i have read correctly, in order to lend bitassets into existence the client must have twice as much collateral in bitshares. If that is the case, ones money only gains half the stability of whatever their preferred assets might be while still maintaining half the volatility of bitshares. I imagine that bitshares could be more stable than bitcoin but its price is still a rather unpredictable variable. If i have indeed misunderstood and the collateral is only held for short positions then how can you guarantee that each long is paired to a short as i believe i have seen you state?  Lastly this might be a stupid idea but in order to facilitate the adoption of bitshares is it possible to distinguish each new node so as to allocate bitshares amongst new users. I understand that you have switched to proof of stake and do not want to dilute ags and pts holders stake in the DAC  but for the purposes of adoption, particularly because of the difficulty  in acquiring bitcoin

Sent from my RM-917_nam_usa_100 using Tapatalk

Offline aasl

  • Jr. Member
  • **
  • Posts: 45
    • View Profile
is sending testing BTS terminated already? I have not received anything yet.

Offline denkhaus

  • Newbie
  • *
  • Posts: 18
    • View Profile
Please, can anybody send me some playing-BTS for testing?

GrKE9fvsiWrEJQX6UCAZaXParzu

Thanks!
« Last Edit: January 22, 2014, 04:51:11 pm by denkhaus »