A new model for building successful and massively scalable applications is emerging. Bitcoin led the way with its open-source, peer-to-peer nature, cryptographically-stored records (block chain), and limited number of tokens that power the use of its features. In the last year dozens of applications are adopting the Bitcoin model in order to succeed. Ethereum, Omni and the SAFE Network are just a few of those “decentralized applications” that use a variety of methods to operate. Some use their own block chain (Ethereum), some use existing blockchains and issue their own tokens (Omni Layer), and others operate at two layers above an existing block chain and issue their own tokens (SAFE Network).
This paper describes why decentralized applications have the potential to be immensely successful, how the different types of decentralized applications can be classified, and introduces terminology that aims to be accurate and helpful to the community. Finally, this paper postulates that these decentralized applications will some day surpass the world’s largest software corporations in utility, user-base, and network valuation due to their superior incentivization structure, flexibility, transparency, resiliency, and distributed nature.
For an application to be considered a Dapp (pronounced Dee-app, similar to Email) it must meet the following criteria:
Satoshi Nakamoto, the creator of Bitcoin described his invention as “A Peer-to-Peer Electronic Cash System[2]”. Bitcoin has been shown to effectively solve the problems that arise from a trust-less and scalable electronic cash system by using a peer-to-peer, distributed ledger, the Bitcoin block chain. In addition to being a peer-to-peer electronic cash system however, Bitcoin is also an application that users can interact with through computer software). But most importantly for the purposes of this paper, based on the criteria outlined above, Bitcoin is a decentralized application. Here is why:
Decentralized applications were initially described as Decentralized Autonomous Corporations, DAC, in an article written by Daniel Larimer, of Invictus Innovations. This papers avoids the term corporation for two reasons.
First, because it carries with it unnecessary preconceptions. For instance, a corporation is established in a jurisdiction, it has shares, a CEO, employees, etc. Dapps, like Bitcoin, have none of these characteristics. In addition, the narrative is very important for the way Dapps are perceived by various nations and jurisdictions. The same way that governments struggle to learn and regulate Bitcoin because the concept of currency is associated with it, governments might be compelled to regulate an open-source computer program that is a decentralized application.
Second, because traditional corporations may engage in several techniques to raise capital (like selling shares of its stock and pay dividends or borrowing against its stock and pay interest) that a Dapp does not need. The concept of a Dapp is so powerful and elegant, because it does not include these traditional corporate techniques. The ownership of the Dapp’s tokens is all that is required for the holder to use the system. It’s that simple. The value of the tokens is determined by how much people value the application. All the incentives, all the monetization, all the ways to raise support are built into this beautifully simple structure. Dapps are not required to recreate the functions that used to be necessary in centralized corporations in order to balance the power of shareholders and offer returns for investors and employees.
There are several characteristics according to which decentralized applications can be classified. For the purposes of this paper, we will classify Dapps based on whether they have their own block chain or they use the block chain of another Dapp. Based on this criterion, there are three types of Dapps.
Type I decentralized applications have their own block chain. Bitcoin is the most famous example of a type I decentralized application but Litecoin and other “alt-coins” are of the same type.
Type II decentralized applications use the block chain of a type I decentralized application. Type II decentralized applications are protocols and have tokens that are necessary for their function. The Omni Protocol is an example of a type II decentralized application.
Type III decentralized applications use the protocol of a type II decentralized application. Type III decentralized applications are protocols and have tokens that are necessary for their function. For example the SAFE Network that uses the Omni Protocol to issue ‘safecoins’ that can be used to acquire distributed file storage is an example of a type III decentralized application.
A useful analogy for a type I Dapp is a computer operating system (like Windows, Mac OS X, Linux, Android, iOS) for a type II Dapp a general purpose software program (like a word processor, a spreadsheet software, a file synchronization system such as Dropbox) and for type III Dapp, a specialized software solution (like a mail-merge tool that uses a word processor, an expense report macro that uses a spreadsheet, or a blogging platform that uses Dropbox.) Using this analogy, it may be expected that due to network effects and the ecosystem surrounding each decentralized application, there will be a few type I Dapps, more type II Dapps and even more type III Dapps.
At this point, it is important to mention that there are currently several excellent open-source projects that leverage type I Dapps. Colored coins and CoinJoin, for example, are based on the Bitcoin block chain and provide useful features to their users. These projects however cannot be classified as type II Dapps, according to our definition, because they don’t issue and manage a token. (The development and operation of these projects depends on donations instead.)
For a complete analysis about why tokens associated with Dapps, and the networks powering the Dapps are valuable, see this paper.
There are two common mechanism by which Dapps can establish consensus: the proof-of-work, POW, mechanism and the proof of stake, POS, mechanism.
With the proof-of-work mechanism, decisions about changes in a Dapp are made based on the amount of work that each stakeholder contributes to the operation of the Dapp. Bitcoin uses that approach for its day-to-day operation. The mechanism for establishing consensus through POW is commonly called mining.
With the proof-of-stake mechanism, decisions about changes in the Dapp are made based on the percent ownership that various stakeholders have over the application. For instance, the vote of a stakeholder who controls 10% of the tokens issued by a Dapp, carries a 10% weight. The Omni Protocol is based on the POS mechanism.
The two mechanisms can be used in parallel, as is the case with Peercoin. Such a combination allows a Dapp to operate with less energy consumption than proof-of-work alone, and allows it to be more resistant to 51% attacks.
There are three common mechanisms by which Dapps can distribute their tokens: mining, fund-raising and development.
With the mining mechanism, tokens are distributed to those who contribute most work to the operation of a Dapp. Taking Bitcoin as an example, bitcoins are distributed through a predetermined algorithm to the miners that verify transactions and maintain the Bitcoin block chain.
With the fund-raising mechanism, tokens are distributed to those who fund the initial development of the DApp. Taking the Master Protocol as an example, Mastercoins were initially distributed to those who sent bitcoins to a given address at the rate of 100 Mastercoins per bitcoin sent. The bitcoins collected were then used to fund the development of applications that promoted the development of the Master Protocol.
With the development mechanism, tokens are generated using a predefined mechanism and are only available for the development of the DA. For example, in addition to its fund-raising mechanism, the Master Protocol used the collaboration mechanism to fund its future development. An additional 10% of the Mastercoins generated through fund-raising was set aside for development of the Master Protocol. Those Mastercoins become available through a pre-determined schedule and are distributed via a community-driven bounty system where decisions are made based on the proof-of-stake mechanism.
To summarize: Tokens of a Dapp that establishes consensus through proof-of-work are distributed by mining, by people buying directly from miners and by trading for goods and services; that is the case with Bitcoin. Tokens of a Dapp that establishes consensus through proof-of-stake are distributed based on the contribution of stakeholders during a fundraiser, by people collaborating on the development of the Dapp and by trading for goods and services; that is the case with the Omni Protocol.
Development of decentralized applications takes place in three steps.
Step 1: A whitepaper is published describing the Dapp and its features
As in the case of Bitcoin, the most common way by which a Dapp takes form is by the public release of a whitepaper that describes the protocol, its features, and its implementation. After the public release, feedback from the community is necessary for the further development of the DA.
Step 2: Initial tokens are distributed
If the Dapp is using the mining mechanism to distribute its tokens, a reference software program is released so that it can be used for mining. In the case of Bitcoin, a reference software program was released and the initial transaction block was created.
If the Dapp is using the fund-raising mechanism, a wallet software becomes available to the stakeholders of the Dapp, so that they can exchange the tokens of the DA. In the case of Mastercoin, an Exodus fund-raising address and a wallet script were publicly released.
If the Dapp is using the development mechanism, a bounty system is put in place that allows the suggestion of tasks to be performed, the tracking of the people who are working on those tasks and the criteria by which bounties can be awarded.
Step 3: The ownership stake of the Dapp is spread
As tokens from mining, fund-raising and collaboration are distributed to a greater number of participants, the ownership of the Dapp becomes less and less centralized and participants that held a majority stake at earlier have less and less control. As the Dapp matures, participants with more diverse skills are incentivized to make valuable contributions, and the ownership of the Dapp is distributed further. Through market forces the tokens of a Dapp are transferred to those who value it the most. Those individuals then can contribute to the development of the Dapp in the areas that they have an expertise.
The case of Bitcoin illustrates the point. By some estimates, Satoshi Nakamoto mined many of the first 1,000,000 bitcoins. As developers contributed code to Bitcoin and miners contributed computational power to the Bitcoin network, the market began to value bitcoins more highly. As the system matured even more, people with diverse skills started valuing Bitcoin and contributing to its development. Now that more than 12 million bitcoins are in circulation and Satoshi Nakamoto’s high original ownership stake has been diluted.
Operating under open-source licenses allows Dapps to be open for innovation without restrictions of copyright or patent. In addition, by being completely open-source, decentralized applications can operate under the legal model of open-source software. Bitcoin, for example, uses the MIT open-source software license. The Master Protocol similarly, requires all code that is based on it to be open-source and available to the community.
From a technical perspective, those issuing tokens as part of a crowd-sale are selling access to software for the users of that software. The private keys associated with the tokens that the users purchase are literally the passwords that the users need to access a Dapp’s software. From a tax perspective, those holding tokens are holding digital property. If the tokens have no market value outside of their use in the Dapp, it is hard to determine their actual value.
Because very few jurisdictions have publicly given guidance on how tokens issued by Dapps will be treated from a regulatory and tax perspective, legal expert in the particular jurisdictions should be consulted.
There are no legal entities required for a Dapp to operate because it is not a company. Owners of tokens do not need to be represented by a corporation and contributor do not need any specific legal entity either. However, sometimes tokens are issued by a non-profit organization that will never receive financial benefits from the Dapp. Such an organization will have the following responsibilities:
Ideally, the non-profit organization will make decisions in a decentralized manner, using a “proof of stake” voting mechanism for any decision.
The purpose of a token is to allow access to a computer application. For example, an individual must own a number of bitcoins in order to be able to perform any transaction on the Bitcoin network. Tokens in Dapps do not represent any underlying asset, they do not give rights to a dividend, and no equity is represented through them. Although the value of a Dapp token may increase or decrease over time, tokens are not equity securities.
There are several ways by which the tokens of a Dapp may be distributed:
To develop a Dapp it is advised to follow these steps:
The model allows contributors to get involved with the project as purchasers of tokens, as project contributors or as providers of resources to the network. All of these contributors benefit from the exchange of the tokens and from the possible appreciation of their value.
A user-behavior reward is given to contributors that provide utility to the network. The whitepaper should outline what constitutes utility for the DApp. (For example, hashing power is utility on the Bitcoin network and it is rewarded.) Utility should be measurable, like in the case of a data storage Dapp, amount of storage is measurable.
One mechanism by which type II Dapps can leverage the block chain of type I Dapps is by embedding additional data to the transactions taking place in the type I DA. The Master Protocol, for example, embeds additional data on the transactions of the Bitcoin network. Although currently (February 2014) Mastercoin embeds its additional data in an ad-hoc way into the Bitcoin block chain, the release of the 0.9 version of the Bitcoin reference client will provide a standard method for that embedding. By using the methodology of “provably prune-able outputs,” type II decentralized applications that are based on Bitcoin will be able to embed data in a systematic way and Bitcoin miners will have the option to prune those data.
Given this development, several type III Dapps are in various stages of development. They include:
Dapps have the potential to become self-sustaining because they empower their stakeholders to invest in the development of the Dapp. Because of that, it is conceivable that Dapps for payments, data storage, bandwidth and cloud computing may one day surpass the valuation of multinational corporations like Visa, Dropbox, Comcast, and Amazon that are are currently active in the space.
It would be beneficial to have a well-grounded and easily accessible metaphor for Dapps. Such a metaphor would ideally have the virtue of simplexity, so that it could be used for human-computer interfaces.
Such a metaphor could be a zygote. A zygote is the point where one biological cell generation ends and the next one begins. A zygote acclimates, and it responds to the outside world without changing its genes, it cannot be regulated, it is stuck with its own genes and its recursive. The zygote is autonomous because it is stuck with its own genes, it is an application because it is a cell, it is distributed, and it is authorized to act as a single entity from other other cells; it shares, in other words many of the characteristics of a decentralized application.
Terms that could created out of the term zygote include zyprotocol, the zygotic protocol, zapp, the zygotic application, zen, the zygotic entity, and zybit, the zygotic bit.
“Everything that can be decentralized, will be decentralized”. David A. Johnston
Based on the economic and efficiency advantages of decentralized applications its clear that existing centralized services will be displaced over time by decentralized alternatives. This shift is likely to come most quickly for services in which the network effect advantages of Metcalfe’s Law are most critical to the success of the service provider.
Continue to Part 2
Follow me as I write about Machine Learning , Data Science and Blockchain . You can follow me and my code on Github
Have a nice day !!!