What's the purpose behind minimizing the # of different instructions?
Every instruction needs to be implemented and tested. Fewer instructions means less work and faster time to market Mmm, but you'll need to test the higher-level instructions built on top of the lower-level instructions too right? Or will we be leaving that for third parties to figure out what higher-lev instructions they want? The key is that we can start testing the higher level instructions (which already exist) in parallel with development of the lower level. Instead of longer dev time for 28 instructions, we have (presumably) shorter dev time for 1 instruction and while that is being done simultaneously test the higher lever instructions and C compiler and C library functions. Not that most C library functions will be used as it will be expensive, but I figure if relatively complicated C library functions works based on a single opcode (which we can test right now!), then it bodes well for Turing scripts to function properly I am optimizing time to market by splitting the projects into pieces that can be developed in parallel. Maybe not the perfect split, so always open to improvements as measured by quicker expected time to market James Thanks! Besides time to market, what other factors do we need to consider? CfB mentioned memory, I guess execution speed is also a factor? subleq supports (and maybe encourages?) a lot of jumping around in memory, though the operations performed on the memory contents (subtraction, comparison of first bit (i.e <= 0 or not)) are fast. I'm playing around with subleq right now, so I'm curious about all this. EDIT: I might be able to help here. I know something about general computability theory (e.g. Turing completeness, halting problem concerns), algorithmic complexity analysis. But don't have experience with software testing (at industry standards). 10000 NXT bounty for methodical playing around  There should be publicly available testing suites for standard C libraries, hopefully someone can post it. I doubt execution speed will be an issue at all, since it will be small programs most likely to fit in CPU cache. The OSIC machine model seems to be pretty simple, they implemented it on an FPGA and it got some decent performance. Since CfB said the script cant access anything external, I think they will execute as fast as you can load it from the blockchain, unless it infinite loops. Not sure how CfB will deal with that sort of thing, kill -9 James |
|
|
|
Am I the only one that feels a tip bot for nxt is a terrible idea? Copying what worked for doge is like BMW putting giant hamsters in their commercials because it sold cars
for Kia. Isn't Nxt's 2nd gen features and sophistication going to look played down if we "tip" with it? Also, people tip with doge because it was worthless and there were billions
of them. Even with the recent price surge they are still fractions of a penny in value. Nxt is currently valued at 6-7 cents and will be a lot higher soon. Will you tip
with them just because some tween made a funny meme? Ha ha Dos Equis guy said Chuck Norris doesn't do push ups he pushes the earth down. Here's 50 Nxt.
I love each of my Nxt, no wanna tip with. Wanna marry thems. Please explain to me why I am wrong or misunderstand. I obviously am not getting it or missed something. Thanks
DOGE has trained people to tip. It becomes not such a hard thing to do. Maybe tip amounts will be 5 NXT or 2 NXT, the key is to get more people using NXT. That is always good. We can always tip milli-NXT when that is there. In general, if it is not hard to do, it makes sense to copy what has met success elsewhere. Also, the more people that earn NXT by doing projects, the more people we have that are familiar with programming for NXT. That is most definitely a good thing, even if what they are doing might not be as valuable as we hope. James |
|
|
|
Competitive update: etherium was struck a potentially lethal blow today, CfB said "lets do it" to implementing Turing complete scripts. I stumbled onto a single opcode Turing complete machine code (no joke) and it even has C compiler. There is bounty to start the testing of said C compiler and libraries. I reached out to the developers of the subleq project, no word yet emunie is in a heap of trouble: https://bitcointalk.org/index.php?topic=439045.msg4930738#msg4930738I still have not gotten any answers to some pretty important questions that I have asked over and over: https://bitcointalk.org/index.php?topic=439045.msg4943552#msg4943552However, in spite of all these issues, emunie still might actually do some pretty cool things, so need to keep an eye on it. It supposedly can handle 1000TPS, but now that it is a UK ltd company we are looking at ripple without VC backing and no staff. Have some homework to read the mastercoin thread, but my plan for dealing with mastercoin is to help XCP as much as possible. [disclosure: I am guilty of burning BTC] XCP already does what mastercoin promises to do and while XCP is currently where NXT was right after genesis, XCP is where NXT was right after genesis. DOGE probably doesnt have to worry too much about the NXT tipbot yet, but soon someone will take the 30,000+ NXT bounty. Do I need to bother with Protoshares? Last I heard, key guys jumped ship. While technically XCP is competitive to NXT at some level, I really see the XCP community as being most similar to the NXT community. No big corporate games going on, lots of open activity and since they are built on top of BTC blockchain, the overlap is really not that big. I am working on ways where both NXT and XCP can help each other. Even though I had a NXT tagline, they welcomed me graciously and I felt right at home. James P.S. I can really use some help here. Working night and day on some pretty complex stuff and you have to remember I am supposed to have low competence level with tech, so please guys with actual tech skills, step up. If not for the bounties, so that we can all make NXT more valuable |
|
|
|
5000 NXT bonus if we also get facebook tipping I'll match this also. 5k. Hey! It feels like we are playing poker. How many aces do you have? |
|
|
|
What's the purpose behind minimizing the # of different instructions?
Every instruction needs to be implemented and tested. Fewer instructions means less work and faster time to market Mmm, but you'll need to test the higher-level instructions built on top of the lower-level instructions too right? Or will we be leaving that for third parties to figure out what higher-lev instructions they want? The key is that we can start testing the higher level instructions (which already exist) in parallel with development of the lower level. Instead of longer dev time for 28 instructions, we have (presumably) shorter dev time for 1 instruction and while that is being done simultaneously test the higher lever instructions and C compiler and C library functions. Not that most C library functions will be used as it will be expensive, but I figure if relatively complicated C library functions works based on a single opcode (which we can test right now!), then it bodes well for Turing scripts to function properly I am optimizing time to market by splitting the projects into pieces that can be developed in parallel. Maybe not the perfect split, so always open to improvements as measured by quicker expected time to market James |
|
|
|
5000 NXT bonus if we also get facebook tipping |
|
|
|
OK, somebody here must be able to use this and weselyh's code to deal with NXT payments and make a quick 20,000+ NXT Anybody? You dont have to cooperate if you dont want to, I would just like to know whenever somebody starts on a project so I can go from recruiting mode to checking status mode for that project James |
|
|
|
If we fit all functionality into 50 instructions then it would be great.
1 instruction
What's the purpose behind minimizing the # of different instructions? Every instruction needs to be implemented and tested. Fewer instructions means less work and faster time to market |
|
|
|
do you have a whitepaper - i would love to read about the economic model. fixing and almost fixing exchange rates is one of the most complicated issues in economics
No details but dan says he has solved this. He will win nobel prize in economics! |
|
|
|
Ah, no I cant defend that. He should have blotted them out. My apologies, I thought you were talking about the material in general.
None of which changes my mind about MrV's principles, as I have already explained.
Your both lines contradict each other. His posting both of a private chat, and the real names of his friends, show that he has no priciples and is a ruthless person. Tell me, do you think Dan's words above are relevant to a future investor?
It shows that he wants to create an alternative to Paypal and Visa, and not Bitcoin. We want to create a payment system that is fast and anonymous. At the same time we have decided to comply with all UK laws and use all our energy to impede any potential illegal activities in advance. This is probably why some founders have left. They have imagined eMunie as an anarchist revolution. The founders that are left see it as a global payment system that will be easy to use for anyone, decentralized, extremely fast, inexpensive (even free transactions are possible), but still with a compliance to the law. Ripple competitor! |
|
|
|
|
Does anybody know where the source code for doge tipbot is
|
|
|
|
James/buybitcoinscanada -- you guys just stepped up in a big way. Thanks!
It is my job to allocate nXTcommunityfunds to promising projects and this fits that category for sure Anybody can propose a project and if it is promising i will certainly consider funding it James |
|
|
|
|
Hey eid,
Sorry things got so messed up, you used to be big emunie guy
To this day my questions about the specific algo used to manage the price has not beem answered
Deadline for whitepaper, nowhere to be found. Very interested as no govt ever suceeded at fixing their fiat price
My fear is that a day or two before fundraising closes, whitepaper comes out so nobody can complain that no details were available before investing. The 99% of people that invest before the whitepaper, caveat emptor
What if white paper is not accurate, no way to verify
On the hatching side, no details on how you can boost hatching. I was hoping things werent made so complicated on purpose so that nobody could predict what will happen, except the one person who actually knows how it works
If people are counting on dan's ability to predict crypto market, keep in mind his debate with bcnext about bcnext claim that nxt will be worth a few millions or he will eat his hat. Bcnext hat is safe despite the pages and pages of furseleers sure sounding denouncements of any possibility of nxt success
I still hope that emunie is not incorporating to shield dan from personal liability, but this thread makes me think that emunie ipo investors will be disappointed
My main problem about price fixing by autohatching to match demand is that by definition all demand is met so the price cant go up. Without a matching buyside demand, NOTHING will stop the crash in price. Theoretically it is cool that instead of getting a higher price, you get more emunie, but my feeling is that you will end up with more emunie that is priced proportionally lower. Of course need math models to get any visibility into this, but no details
The last issue i has was that it looks like anybody that buys emunie is giving up 50% to the hatchers, so a postmine of 50%, which sounds pretty steep, especially if you only end up with three times the emunie whose price is 20% of original price.
With long open ipo, who is left to buy emunie?
James
|
|
|
|
I will allocate 10000 nxt to be added to this |
|
|
|
This is an interesting concept but I don't expect widespread adoption of XCP. First of all, its use is very limited. I don't think it will win over a single professional trader ever. In this time and age of HFT (high frequency trading) in most markets, especially the very liquid ones, where latencies are counted by the microseconds, the pace of 1 block per 10 minutes is not really going to cut it. This alone already rules out liquid markets like FX, MM, rates equities and commodities, where 99.999% of the trades are. Low Lack liquidity in XCP will translate to huge spreads and slippage from hell, making it unprofitable to trade through XCP, which in turn hurts liquidity. You can't break that vicious circle unless you can compete with the lightning speed and market depth at the exchanges. Unfortunately I don't see that ever happening. And oh, the exchange fees in these very liquid markets are negligible, especially for high volumes, so no advantage for XCP there either. Mind you I'm not dismissing XCP, it's fantastic technology and it will have many novel applications. But I just don't think it's going to have the same level of impact as bitcoin, which was truly a paradigm shift. I hope I'm proven wrong though! Blockchain-based technologies in general, whether BTC, MSC, XCP or Ethereum, will not be used for HFT. The scalability is simply not there for that. OpenTransactions servers, on the other hand, are excellent for HFT. Blockchains are for higher-value transactions and settlement. There is also some sentiment that maybe the value of all the things being traded at millisecond intervals is actually not changing so fast. The change in prices just being created by all the HFT interacting with each other. So, a slower pace might not be such a bad idea. Put best bids/asks in and it gets settled every ten minutes (or minute for some blockchains) James |
|
|
|
Who wants to make 10000 NXT? I dont want to make a mistake on the selection of low level opcode. Part of the task is to actually use the subleq, higher_subleq, C compiler, etc. and to make sure it functions properly. http://arxiv.org/pdf/1106.2593.pdfWe need an objective assessment, so lets compare it with the Oberon compiler or whatever other proposed solutions are made. Limit of three candidates to make sure it doesnt turn into never ending task. We also need a C compiler test suite to assess what percentage of the language is even implemented and what percentage passes the tests. Identifying this test suite is part of the task. We also need a bunch of low level C library source code (no machine dependencies) along with matching test suites to validate library functions. This portion of the task is valid regardless of what low level selection we make. This is why I am announcing bounty so quickly. I want there to be a repository of all the library code that has been verified to work with whatever compiler/language/machine code we end up with. I prefer cooperative approach and this task has many parts. Please quote this post if you are interested in doing this. The clock is ticking, lets get this done! James |
|
|
|
So, please explain why C compiler -> higher_subleq -> subleq is useless academic and not a practical solution. You have to remember I have been proven to be incompetent in technical stuff so I need to have things explained to me, even if it is obvious to the smart guys
James
Hmmm. Where to start with? 1) Yes, it could be done with it. Why, because both languages are turing complete. (That's the stuff on the paper, proven with formulas etc.) 2) Yes, it could be practical. But, the bugs are not on the paper and not in the formulas. But in the source code of the compilers and the VM AND their respective compilers. It's not the mathematical proof that's wrong, it's just that programmers make mistakes as human usually do. 3) So, use proven technologies in the sense that time has proven them to be reliable. Not proven in the sense of formulas. It's more like robust or bulletproof code for VM and compilers. Less bugs, more support, because more people are using them right now. What if the subleq developers discontinue development? And so on. But anyway, yes. You won the contest. 1 Op-Code is cool. Hey, as cool as one opcode is I much prefer to use 28 or 32 or whatever if it gives us a better starting point as far as fully debugged libraries, etc. goes. The less work we have to do to be able to construct useful solutions, the better. That is my metric. Does anybody know of any open source C compiler for the 28 opcode language? Libraries? If I have to choose between subleq and another unproven simple machine language ecosystem, I would choose subleq just because is to kind of cool with one opcode. But only if it is at the equivalent state. So, please anybody know of simplified machine language with open source C compiler and lots and lots of library code and projects verified on it? James ARM-RISC, there is. Compilers and VMs en masse. VM for ARM-RISC.... have to research that. But's time to catch some sleep. order of magnitude more work to implement ARM VM than OSIC VM, OSIC is the CPU they designed using subleq |
|
|
|
I have idea for high level language to put on top of C. Why not some subset of Java? If we can then allow access to a subset of Java (the safe ones) in the NXT core to the script, then it would be most powerful setup: Java -> C -> higher_subleq -> subleq -> subleq VM Since the C compiler for higher_subleq is relatively untested, this would require a project to port and validate the Java (subset) using that compiler. Anybody interested? How much work would it be? Just running a small java program could generate so much NXT! Probably not so practical, but for a tech demonstration on the power of NXT, it would certainly get a lot of PR. Also, depending on how we count transactions, 1000TPS now seems it might not be enough... James P.S. http://arxiv.org/pdf/1106.2593.pdf is the key paper on this |
|
|
|
Libraries/Demo
Libraries for programs written in SUBLEQ assembly, as accepted by sqasm, are available.
They provide several common functions:
getint (read integer from input)
gets (read string from input)
putint (write integer to output)
puts (write string to output)
Several useful procedures are also available:
bubblesort (sort a string of characters)
calc (perform a given operation on two integers)
factorial (calculate the factorial of a positive integer)
primes (generate and print a list of primes)
Usage information and equivalent C code (where appropriate) is provided in the comments of each of these files.
A menu-driven program demonstrating the above libraries is also provided. Simply call make run in the project root directory to run the demo program.
Interpreter
A very minimal (only 222 bytes in size) SUBLEQ interpreter written in C is available:
_____________
What everyday operations do you have in mind?
You really like it, right?  It's okay. Nevermind. It's just that I trust technology like that only after many years of production. That's all. It's not that I do not think they weren't capable of producing a good piece of compiler and libs. It's just that: nobody's perfect, especially no programmers. In the end, only time can 'prove' correctness. Well if we can use proven solution, then we wouldnt have to develop our own. Do you want to wait until etherium is proven and then use that? I am confused again. I think we can have a path to Turing complete script with C compiler and a small amount of functions. The alternative is to implement 28 or 32 opcodes, which is a lot more work as each opcode needs to be implemented and tested, etc. If CfB signs off on subleq VM, then we can start in parallel fleshing out and TESTING more and more C libraries based on subleq C compiler. I like being able to develop a project in separate pieces. As soon as we have subleq scripting, we can then hook up all the higher level stuff that has been ported and tested. Does that make sense? Anon, Alias, CIYAM, other smart guys. Please tell me if I am being incompetent again, or did I get lucky? James P.S. If CfB signs off on subleq VM, I will authorize bounties for getting the higher level functions done. It is clear to me that we need a layer of code that abstracts the details of script, AM, etc and allows higher level development. With a C compiler available, maybe one of the bounties will be to port a higher level language to subleq VM Ah yes. And if you have to pay per op, then you will get poor when using subleq. It's really more expensive. Hmmm. maximizes transaction fees, unfortunate isn't it. I imagine much lower min fee would be needed as soon as we launch Turing complete scripts |
|
|
|
So, please explain why C compiler -> higher_subleq -> subleq is useless academic and not a practical solution. You have to remember I have been proven to be incompetent in technical stuff so I need to have things explained to me, even if it is obvious to the smart guys
James
Hmmm. Where to start with? 1) Yes, it could be done with it. Why, because both languages are turing complete. (That's the stuff on the paper, proven with formulas etc.) 2) Yes, it could be practical. But, the bugs are not on the paper and not in the formulas. But in the source code of the compilers and the VM AND their respective compilers. It's not the mathematical proof that's wrong, it's just that programmers make mistakes as human usually do. 3) So, use proven technologies in the sense that time has proven them to be reliable. Not proven in the sense of formulas. It's more like robust or bulletproof code for VM and compilers. Less bugs, more support, because more people are using them right now. What if the subleq developers discontinue development? And so on. But anyway, yes. You won the contest. 1 Op-Code is cool. Hey, as cool as one opcode is I much prefer to use 28 or 32 or whatever if it gives us a better starting point as far as fully debugged libraries, etc. goes. The less work we have to do to be able to construct useful solutions, the better. That is my metric. Does anybody know of any open source C compiler for the 28 opcode language? Libraries? If I have to choose between subleq and another unproven simple machine language ecosystem, I would choose subleq just because is to kind of cool with one opcode. But only if it is at the equivalent state. So, please anybody know of simplified machine language with open source C compiler and lots and lots of library code and projects verified on it? James |
|
|
|
|
Show Posts
