Tau-Chain and Agoras Official Thread: Generalized P2P Network

[zooitje]

What do you guys think of http://www.tezos.com?

Is it comparable to Tau-Chain and Agoras? Once released, will Tau-Chain and Tezos compete, complement each other or neither?

Thanks. Keep up the good work. I really appreciate and admire the work you all do.

[Sam123]

Very interesting. Just finish reading the white paper.
Thanks for sharing 

[Hueristic]

When this is ready will I be able to install and run an operating system on it?

[ohad]

When this is ready will I be able to install and run an operating system on it?

good q. we want to reach there too. thats a good example why it must be of high quality, well tested, minimalistic, and as fast as possible

[Hueristic]

When this is ready will I be able to install and run an operating system on it?

good q. we want to reach there too. thats a good example why it must be of high quality, well tested, minimalistic, and as fast as possible

This is a goal I can get behind! Now the question is will the project be able to use cpu extensions when it can conceivable be comprised of cpu's that will not have those extensions built in? Will those extensions be virtualized and if so will a licensing fee need to be paid for that capability? I'm aware of a russian os that runs on a cpu without extensions that does virtualize x86 instructions but I am not sure of the legalities involved.

I'm not sure of what you mean by the vague term "high quality".

[ohad]

When this is ready will I be able to install and run an operating system on it?

good q. we want to reach there too. thats a good example why it must be of high quality, well tested, minimalistic, and as fast as possible

This is a goal I can get behind! Now the question is will the project be able to use cpu extensions when it can conceivable be comprised of cpu's that will not have those extensions built in? Will those extensions be virtualized and if so will a licensing fee need to be paid for that capability? I'm aware of a russian os that runs on a cpu without extensions that does virtualize x86 instructions but I am not sure of the legalities involved.

that's part of being minimalistic: hopefully we'll use a very restricted instruction set. we also plan to be somehow multiarch on release. recall we build a jit compiler, a real one, i.e. a one that in real time emits machine code.

I'm not sure of what you mean by the vague term "high quality".

it has to be competitive to the state of the art of programming languages.

[Hueristic]

When this is ready will I be able to install and run an operating system on it?

good q. we want to reach there too. thats a good example why it must be of high quality, well tested, minimalistic, and as fast as possible

This is a goal I can get behind! Now the question is will the project be able to use cpu extensions when it can conceivable be comprised of cpu's that will not have those extensions built in? Will those extensions be virtualized and if so will a licensing fee need to be paid for that capability? I'm aware of a russian os that runs on a cpu without extensions that does virtualize x86 instructions but I am not sure of the legalities involved.

that's part of being minimalistic: hopefully we'll use a very restricted instruction set. we also plan to be somehow multiarch on release. recall we build a jit compiler, a real one, i.e. a one that in real time emits machine code.

I'm not sure of what you mean by the vague term "high quality".

it has to be competitive to the state of the art of programming languages.

OK, when you state your building a JIT do you mean the network will compile any code on the fly? The only JIT I know of is java but I don't program in many 5th level languages. I'm a little confused on how JIt will in any way allow the running of an os (what I am really interested in) It sounds like you are building a virtual machine that will run scripting on the fly. Which of course is not required for what I am interested in as it is all precompiled. I guess it could compile vanilla linux kernels but what system can't? Can you clarify this for me.

[ohad]

OK, when you state your building a JIT do you mean the network will compile any code on the fly? The only JIT I know of is java but I don't program in many 5th level languages. I'm a little confused on how JIt will in any way allow the running of an os (what I am really interested in) It sounds like you are building a virtual machine that will run scripting on the fly. Which of course is not required for what I am interested in as it is all precompiled. I guess it could compile vanilla linux kernels but what system can't? Can you clarify this for me.

the language consists of facts, rules, queries, and side effects (eg disk/network).
the runtime process mainly includes reasoning over the rules to resolve the query and by that possibly encounter side effects.
this is actually an automated theorem prover, like idris/coq/agda.
all this logic better happen very fast, as the whole point is to rely on it. so the whole process is incrementally translated to machine code in memory, and called to be executed by the cpu.
specifically for operating systems, as they're written mostly in C, at the same way they can be written with tau.
jit compilation is a technology used in many languages eg java you mentioned. it indeed proved itself as a high performance way to translate a high level programming language to the machine.

[Hueristic]

OK, when you state your building a JIT do you mean the network will compile any code on the fly? The only JIT I know of is java but I don't program in many 5th level languages. I'm a little confused on how JIt will in any way allow the running of an os (what I am really interested in) It sounds like you are building a virtual machine that will run scripting on the fly. Which of course is not required for what I am interested in as it is all precompiled. I guess it could compile vanilla linux kernels but what system can't? Can you clarify this for me.

the language consists of facts, rules, queries, and side effects (eg disk/network).
the runtime process mainly includes reasoning over the rules to resolve the query and by that possibly encounter side effects.
this is actually an automated theorem prover, like idris/coq/agda.
all this logic better happen very fast, as the whole point is to rely on it. so the whole process is incrementally translated to machine code in memory, and called to be executed by the cpu.
specifically for operating systems, as they're written mostly in C, at the same way they can be written with tau.
jit compilation is a technology used in many languages eg java you mentioned. it indeed proved itself as a high performance way to translate a high level programming language to the machine.

OK, as you say "All this better happen pretty fast" well nothing can happen fast on a decentralized system in comparison to say a soc. So I see no way this can run a modern OS faster than a watch could (actually I'm sure it could not). You state it is equivalent to idris/coq/agda and they are languages so what you are saying is this will be communicated to by code so I do not see how it can run an OS to begin with, it sounds like it will be a system that can be sent jobs in a que type system, correct? There is nothing wrong with that and it is a sound plan but I cannot equate this model being able to run a precompiled OS. Maybe my understanding of what you are saying is off?

[ohad]

OK, as you say "All this better happen pretty fast" well nothing can happen fast on a decentralized system in comparison to say a soc.

everything i speak about at this scope is local. to run the programming language on a single machine. the logic should happen fast as the runtime may grow exponentially with the length of the code, so we want to push the limits to the max.

So I see no way this can run a modern OS faster than a watch could (actually I'm sure it could not).

a tau based os won't be faster just for the reason it's using our compiler. but once one is able to reason over the code, in a consistently decidable manner, the picture changes. you can get an apriori proof of correctness, security etc all up to formal definitions. you can also reason about performance issues.

You state it is equivalent to idris/coq/agda and they are languages so what you are saying is this will be communicated to by code so I do not see how it can run an OS to begin with, it sounds like it will be a system that can be sent jobs in a que type system, correct? There is nothing wrong with that and it is a sound plan but I cannot equate this model being able to run a precompiled OS. Maybe my understanding of what you are saying is off?

like idris/coq/agda only from logical aspect. they don't have a jit compiler architecture, for example. on tau we try to take the logic as closer to the bare metals.

that all aside the blockchain and dht parts. which of course can be referred from inside the language too.

[Hueristic]

OK, as you say "All this better happen pretty fast" well nothing can happen fast on a decentralized system in comparison to say a soc.

everything i speak about at this scope is local. to run the programming language on a single machine. the logic should happen fast as the runtime may grow exponentially with the length of the code, so we want to push the limits to the max.

So I see no way this can run a modern OS faster than a watch could (actually I'm sure it could not).

a tau based os won't be faster just for the reason it's using our compiler. but once one is able to reason over the code, in a consistently decidable manner, the picture changes. you can get an apriori proof of correctness, security etc all up to formal definitions. you can also reason about performance issues.

You state it is equivalent to idris/coq/agda and they are languages so what you are saying is this will be communicated to by code so I do not see how it can run an OS to begin with, it sounds like it will be a system that can be sent jobs in a que type system, correct? There is nothing wrong with that and it is a sound plan but I cannot equate this model being able to run a precompiled OS. Maybe my understanding of what you are saying is off?

like idris/coq/agda only from logical aspect. they don't have a jit compiler architecture, for example. on tau we try to take the logic as closer to the bare metals.

that all aside the blockchain and dht parts. which of course can be referred from inside the language too.

OK, so just to clarify. As I understand it this system will be able to compute (in a virtualized manner) a OS (probably with a wrapper I would assume that gets compiled on the fly : not sure everything has to be though) but the tradoffs will force such lag that in a real world application there is no way you could use this system as even a desktop? I do understand that for specifically written large scale computations this would be a very viable alternative to a supercomputer. I am trying to nail you down On one thing and I have mentioned that in every post.

Will this be able to run a Linux or M$ OS at an acceptable rate to say write an office Document? I would think it will tkae an hour to move a mouse across the screen even if you don't get kernel exceptions everywhere.

[ohad]

OK, as you say "All this better happen pretty fast" well nothing can happen fast on a decentralized system in comparison to say a soc.

everything i speak about at this scope is local. to run the programming language on a single machine. the logic should happen fast as the runtime may grow exponentially with the length of the code, so we want to push the limits to the max.

So I see no way this can run a modern OS faster than a watch could (actually I'm sure it could not).

a tau based os won't be faster just for the reason it's using our compiler. but once one is able to reason over the code, in a consistently decidable manner, the picture changes. you can get an apriori proof of correctness, security etc all up to formal definitions. you can also reason about performance issues.

You state it is equivalent to idris/coq/agda and they are languages so what you are saying is this will be communicated to by code so I do not see how it can run an OS to begin with, it sounds like it will be a system that can be sent jobs in a que type system, correct? There is nothing wrong with that and it is a sound plan but I cannot equate this model being able to run a precompiled OS. Maybe my understanding of what you are saying is off?

like idris/coq/agda only from logical aspect. they don't have a jit compiler architecture, for example. on tau we try to take the logic as closer to the bare metals.

that all aside the blockchain and dht parts. which of course can be referred from inside the language too.

OK, so just to clarify. As I understand it this system will be able to compute (in a virtualized manner) a OS (probably with a wrapper I would assume that gets compiled on the fly : not sure everything has to be though) but the tradoffs will force such lag that in a real world application there is no way you could use this system as even a desktop? I do understand that for specifically written large scale computations this would be a very viable alternative to a supercomputer. I am trying to nail you down On one thing and I have mentioned that in every post.

Will this be able to run a Linux or M$ OS at an acceptable rate to say write an office Document? I would think it will tkae an hour to move a mouse across the screen even if you don't get kernel exceptions everywhere.

still very soon to tell to which extent tau will be used to write an os. tau is only a programming language. complex things like os will take relatively significant time until people will begin to design it to the details.
for now we know that the more we push forward the quality of the product, the more areas it'll be able to cover. i hope it'll reach os too.
that all said,
tau has no problem calling native code (eg dll).
also it can always be used as a component as is written in c/c++ so can always be used from other programs.

[Hueristic]

OK, as you say "All this better happen pretty fast" well nothing can happen fast on a decentralized system in comparison to say a soc.

everything i speak about at this scope is local. to run the programming language on a single machine. the logic should happen fast as the runtime may grow exponentially with the length of the code, so we want to push the limits to the max.

So I see no way this can run a modern OS faster than a watch could (actually I'm sure it could not).

a tau based os won't be faster just for the reason it's using our compiler. but once one is able to reason over the code, in a consistently decidable manner, the picture changes. you can get an apriori proof of correctness, security etc all up to formal definitions. you can also reason about performance issues.

You state it is equivalent to idris/coq/agda and they are languages so what you are saying is this will be communicated to by code so I do not see how it can run an OS to begin with, it sounds like it will be a system that can be sent jobs in a que type system, correct? There is nothing wrong with that and it is a sound plan but I cannot equate this model being able to run a precompiled OS. Maybe my understanding of what you are saying is off?

like idris/coq/agda only from logical aspect. they don't have a jit compiler architecture, for example. on tau we try to take the logic as closer to the bare metals.

that all aside the blockchain and dht parts. which of course can be referred from inside the language too.

OK, so just to clarify. As I understand it this system will be able to compute (in a virtualized manner) a OS (probably with a wrapper I would assume that gets compiled on the fly : not sure everything has to be though) but the tradoffs will force such lag that in a real world application there is no way you could use this system as even a desktop? I do understand that for specifically written large scale computations this would be a very viable alternative to a supercomputer. I am trying to nail you down On one thing and I have mentioned that in every post.

Will this be able to run a Linux or M$ OS at an acceptable rate to say write an office Document? I would think it will tkae an hour to move a mouse across the screen even if you don't get kernel exceptions everywhere.

still very soon to tell to which extent tau will be used to write an os. tau is only a programming language. complex things like os will take relatively significant time until people will begin to design it to the details.
for now we know that the more we push forward the quality of the product, the more areas it'll be able to cover. i hope it'll reach os too.
that all said,
tau has no problem calling native code (eg dll).
also it can always be used as a component as is written in c/c++ so can always be used from other programs.

Am I correct in assuming the network backbone will equate to a system bus?

[ohad]

Am I correct in assuming the network backbone will equate to a system bus?

in the sense of being self defining, namely, everyone run the root code.
but the main usage is not to use only the root but to use networks and application over tau. which can define any (collaborative) behavior they want.

[Hueristic]

Am I correct in assuming the network backbone will equate to a system bus?

in the sense of being self defining, namely, everyone run the root code.
but the main usage is not to use only the root but to use networks and application over tau. which can define any (collaborative) behavior they want.

IC, Gl. Looking forward to release.

[maco144]

What do you guys think of http://www.tezos.com?

Is it comparable to Tau-Chain and Agoras? Once released, will Tau-Chain and Tezos compete, complement each other or neither?

Thanks. Keep up the good work. I really appreciate and admire the work you all do.

So, this is the same project as Tau?

[ohad]

What do you guys think of http://www.tezos.com?

Is it comparable to Tau-Chain and Agoras? Once released, will Tau-Chain and Tezos compete, complement each other or neither?

Thanks. Keep up the good work. I really appreciate and admire the work you all do.

So, this is the same project as Tau?

not at all. tau isn't a smart contract language. it's a general purpose language, including being a programming language, contracts, information, know-hows, processes, math, everything.
tauchain is a shared collaborative network to store them all, not only store, but also understand. tau is able to reason over information given in its language.

[kjn311]

Can I run Tau on Ethereum?

I believe the answer is yes. Why not just start off in that direction and give this project more exposure?

[ohad]

Can I run Tau on Ethereum?

I believe the answer is yes. Why not just start off in that direction and give this project more exposure?

because.. eth is an inherently unsafe and low-capacity platform.

[ICOcountdown.com]

Can I run Tau on Ethereum?

I believe the answer is yes. Why not just start off in that direction and give this project more exposure?

Tau is the opposite realm of Ethereum, this is why I keep calling it a "hedge". Tau isn't turing complete it uses decidable language instead of undecidable language.