Topic: Tau-Chain and Agoras Official Thread: Generalized P2P Network - page 158. (Read 309768 times)

Thanks for the questions,

1. Whatever can be implemented with "old" coins like BTC, will probably be supported on Agoras, of course with limited features.
2. There will be no "gas" or fees for using tau or agoras. Otherwise, someone will create a clone with no fees, as you pointed out. We do try to come up with a design that is hard to compete with, of course, by all means we find.
3. No one essentialy prevents someone from creating Agoras competitor over tau, but it'll be hard to compete with tau's developers, especially in introducing the features we've been thinking on for over a year. No one even came up with a "small and simple" feature of agoras like Zennet. But again I think this points to a kind of risk that exists in any project, and usually doesn't come realistic, yet we do keep that in mind and try to make it more valuable to join us rather compete with us.

Please let me know if this answers.

I wasn't asking about RDF triples and NQuads: these are just a representation format. What I am asking is the specification of the Tau language itself. OWL2 for instance is the RDF implementation of a logic language called SROIQ(D) formally specified here that is a subset of description logic and its derivation as RDF based ontology is specified in great details here.
We know from the whitepaper that Tau is borrowing to constructive / intuitionist logic. But is it a subset, a variation, or does it implement the full logic? Is there a whitepaper or document covering the formalism? Is there a document that describes the ontology including the core logic and library of built-ins related properties?

Is Tau entirely structured around CWM? What about the builtins that are not already implemented by CWM? For instance DHT related features, RNG etc. Will you contribute them to the original CWM? Are you going to fork CWM to extend it? Or does CWM come with a way to register new builtins and if it does is it statical (compile time) or dynamic (runtime)?

You are using the W3C swap/math and swap/log ontologies.
Are they a subset of the Tau ontology?
Again, that would need to be clarified. What IS valid Tau and what is not?

If I understand correctly your reply I/O originated events and data are being incorporated as new facts that will be (lazy-)evaluated / stated next time the reasoner does a resolution that needs i/o related facts, which in other words means synchronous i/o. I can see how this is going to add a lot of latency, but let me read that Idris paper first and I'll get back with more questions

You gave this example above: "{ ?z a tau:tcpEndpoint } => { ?y tau:tcpConnect ?z }".
Are the built-ins behind tau:tcpEndpoint and tau:tcpConnect already developed or specified in a document?
What will be all the built-ins introduced by Tau?


Let me quote HMC's reply to the question "How does it compare to Ethereum" in the interview on CoinTelegraph
That's the point. So Tau will need to rely on a very self-constrained ultra-modular programming paradigm so that it is possible to prove that every chunk of code is both space and time bounded regardless of its input, while making sure that the proof itself doesn't become more complex than the chunk of code it vouches for...

I'm really curious how that will work in practice. A piece of code that demonstrates this would be more than welcome.

Tau, the first Proof-of-PhD smart contract system

I don't mind finding my way in a 10MB dump. I don't want to waste your time answering questions if the answers are already there somewhere.

Proposed that nicely, how can I refuse? Here are a few more questions:
- Do you have specs about what exactly Agora is going to be. I mean other than the story of Bob and Alice . What kind of assets will Agora allow to trade? Will it features a full fledged distributed exchange like Ripple? Will it support other market paradigms like auctions or prediction markets?
- Have you started working on Agora?
- How long do you estimate you will need to complete Agora?
- Tau ETA?
- Agora ETA?

Maybe to phrase what he is asking differently: after various markets are created over Tau (Agoras that will require the Agora coin to use them)-- what stops one to create competing markets, also over tau (parallel to Agoras) with BTC, ETC, or any other crypto tools. Won't these have an advantage over the Agoras market? (an advantage in the sense that they'll utilize other coins without limits).

OR, even if the above is correct, is your intention with the Agoras to enable features that can utilize any coin but that like in ETHEREUM, will require Agoras an an underlying gas/fee/whatever for every transaction while also enabling them as a "classic coin" within that Inner world?

Accepting BTC or fiat for features elevating tau's unique properties (such as decidability and proofs over code) is of course impossible - such features are impossible to be implement over existing chains, even ethereum's.
As for the risk that someone else will write another and better Agoras, I think such risk exists in every project that comes to mind.

What makes people who trade code asking Agoras in return of their services? Same applies for the other suggested markets, what obliges people to start using Agoras? They could ask, accept and take e.g. BTC, fiat or some other “app currency” instead?

Is there strong enough incentive for early adaptors and early coders to start asking compensations and stick in Agoras and what is that incentive? How sure it is that Agoras will be accepted as “the money” over tau? If not so and there’s no trading volume after IPO, price starts going down and coders easily abandon the Agoras ship…

For now you can do on tau what existing RDF reasoners would do (specifically Euler[EYE]) and at the same syntax, which is Turing complete btw. When we'll finish the typesystem it'll be essentially different, and ofc tau will cease to be Turing complete at that point.

Is it possible to write already apps on some alpha version? I would like to try out ;-)

Agoras is indeed "just another app" over tau. This is important to bear in mind. Everyone can create more currencies over tau, and tau doesn't have a coin as for itself.

But Agoras is of course far beyond a cryptocurrency. It will have markets for trading code for money in a totally secure and trustless way (as in Code and Money blogpost), allow renting computational power with or without verification of execution, renting professionals in video conference and micropayments, smart contracts, pay-for-proof wheter it's math homework or a program to write, decentralized intelligent search engine, and even more.

Can you explain more about this? What can I do, in terms of interfacing with Tau, with Agoras? Is it just another cryptocurrency that is just built using Tau, or does it have some special status or feature within Tau Chain? Why would anyone want to own Agoras over another cryptocurrency within Tau, or outside of Tau for that matter?

Tau is completely independent in Agoras or its coins. Agoras, on the other hand, is an application over tau. Everyone can write apps over tau - that's the very intention. So formally, tau as for itself is not funded. Also it does not come with a coin, but coins can be programmed over it.
Since after all one has to develop economical part over tau, using tau's unique features, we're going to do it as Agoras. And we do need to distribute the coins somehow, hence the presale.


No, the 42M intermediate tokens represent 100% of all future coins, no matter if they'll be 147B or any other number, and the prices refer to those intermediate tokens.

This looks like a strong idea even though I don't understand much of the white paper details. Is Tau-Chain purely valued through this Agoras? If later someone wants to use Tau-Chain to solve some problem they buy Agoras (from market?) and pay for the usage of Tau-Chain network (how pricing the usage?). I understood inside Tau-Chain own rules can be set for that so can they make a rule to let someone use lot of computing power without charging anything? Or maybe I don't understand the logic where the value of Agoras might come..

If one Agoras is around 0.0006 and market cap is 7M USD or 30000 BTC (where to see the market cap?), then little over 50M Agoras is owned by people? And there is 147B Agoras to be shared. So 0.034% of the total amount is shared at the moment? Are my conclusion/numbers correct?

Also I encourage you to play with http://attempto.ifi.uzh.ch/race/

The syntax of the language itself is not new. The RDF family is basically all syntax-sugaring to NQuads http://www.w3.org/TR/n-quads/ so all RDF languages can be converted to NQuads, while plenty of converters out there. Current tau's code can read NQuads, JSON-LD, and (yet partial) Notation3.

Supporting network features etc. is done via builtins, which is also a common mechanism among RDF reasoners. See for example http://www.w3.org/2000/10/swap/doc/CwmBuiltins


Let me use the cwm builtins I just linked to demonstrate printing numbers from 1 to 10:

1 log:outputString 1.
{ ?x log:outputString ?x. (1 ?x) math:sum ?y. ?y math:lessThan 11 } => { ?y log:outputString ?y }.

see http://pastebin.com/H2gqLUzC


Except I won't promise now we'll use exact SPARQL syntax, then sure. Tau can be equivalently described as a decentralized triplestore with dependent types.


At the syntax level, IO is just another builtin predicate, like { ?z a tau:tcpEndpoint } => { ?y tau:tcpConnect ?z }. At the typesystem level, we of course have Effect types/Monads https://en.wikipedia.org/wiki/Monad_(functional_programming)
The reasoning flow does not change at all due to IO (in general), but types originated from IO are tagged as effect types and of course aren't treated as "pure truth". A good reference is how Idris use side effects http://eb.host.cs.st-andrews.ac.uk/drafts/eff-tutorial.pdf


Apologies but I didn't really understand the question.


This goes right back to decidability, and to the point that one shouldn't expect the autoprover to prove all math by itself instantly.

Assume I write a tau or idris program that halts on the first zero of Riemann's Zeta which is not on the critical line. Of course, no one expect the prover to determine quickly if this program ever halt. If I knew how to do that magic, I'd already take $6M by solving the 6 math open problems of Clay institute

On the other hand, since the logic is decidable, the provable statements are exactly all correct statements and vice versa. Since every program has some runtime complexity, it is apriorily possible to give a proof for a runtime of any program.

So assume I wrote some simpler code with runtime complexity that I know how to compute and prove, then I can specify my proof or parts of it, and let the prover finish the missing details.


This is exactly what I work on these days, so I'll have to finish first.


I don't think there's anything one can easily track. ~10MB of text won't help anyone. So for now one can only ask and get answers, and with time we'll document everything.


Please let me know if I missed something or if other questions arise.

I have a few questions:
- Do you have a specification (ontology, BNF grammar, spec document...) of the Tau language as it is now and/or as it is supposed to become once all the networking and crypto features are built in?
- Where can I find a Tau code example that emulates simple procedural behavior like a loop for instance?
- Will the Tau node to be a full fledged RDF triple store and SPARQL end-point?
- How do you plan to make the transition from knowledge (decision of spawning a new socket for instance) to action (making a system call to create a new socket): will each i/o module poll the knowledge base periodically for relevant action? Will you place control points at key steps of the inference process to intercept propositions with a known action semantic and take action?
- Will input from i/o module trigger asynchronous reasonning or will it just add new propositions in the knowledge base for some synchronous process to reason with at next iteration?
- Is the interface between the knowledge base and the system level functionalities already developed or specified? Where can I find further details?
- How can Tau guarantee that it can determine the upper bound of when a program can halt in less time than it would take to run the program? What if the program is already the shortest possible program to solve the problem it represents? At most I could imagine how code analysis could help determining a big-o estimate of the program complexity but then that won't help much to make sure that the program can run within the time boundaries of a given blockhain block as was suggested in the cointelegraph interview.
- Where can I find more (code / documents) about the JIT compilation of the Tau language?
- Are there archives of IRC discussions / forum discussions / email discussions etc that can help understanding decisions made around the design?

A special offer to wholesales in september was announced: http://www.idni.org/blog/sep-special

have a look at the code github.com/naturalog/tauchain

Don't you have the majority of TAU? If you consider TAU money, then you have it all. Now do some code! LOL

It never went but is only one among many other features in Agoras, most of them (which I try to put emphasis on) are possible only thanks to tau's unique features.

I like the original distributed computer idea. Wonder where that went?

I like the idea also. It sucks I have to miss out because its way over priced.