A.N.N.E. — The Alpha Testing Begins
The alpha test of the ANNE protocol has just been released to a small group. The following is the help document on the site.
This is A.N.N.E. search, one example of how the ANNE protocol creates interconnected usable data, on BSV.
A.N.N.E. is a protocol standard for storing various data on-chain via OP return (and later op pushdata). It loosely emulates the human brain in that its basic building block is a neuron and those neurons interconnect to form circuits. Using the standard creates a “relon”. A relon is a conceptually encoded standardized indexed triplet. Everything and anything can be encoded with ANNE by breaking it down into its most basic concepts and wiring them back together as a circuit. It doesn’t matter how complex or small the circuit is. It is what it is. That circuit can then be used in greater circuits because they are all made out of the same “stuff”.
Using this standard produces an ANNE relon. With neurons and relons we can form brain-emulating circuits.
Thinking in terms of neural circuits will require a change of mind-set for some. Instead of 1000s of custom tables for every “thing”, with differing sets of columns in each table for properties of those things, we have only two indexes and two “virtual” tables. Both are native to Bitcoin. Bitcoin Addresses and Bitcoin Transactions. I see a lot of excitement and creativity about using TXs but without leveraging the power of bitcoin addresses those TXs have little native meaning. It is a brain full of connections that fly off in every direction but never land at another neuron. Translation: They all require special interpretation that takes energy, resources and time to make happen. We have both TXs AND Bitcoin addresses to work with. ANNE uses them both with equal measures of importance. They are both foundational. Every concept gets its own bitcoin address. Every would-be column name, every class, every instance, every verb, every word, every thing. It is indexed and is a uniquely identifiable neuron. Bitcoin allows for this on a scale previously not possible. We get this automatically by using the gift of Bitcoin with intent knowing what it has to offer in this area.
Note: While I do not plan to break words down into circuits of their letters, this is entirely within reach of ANNE and accounted for. It can be done at a later date, if needed, or wanted. Tech like visual recognition and corpus-free (training-less) language parsers (learning by observation) would use such a thing.
Every TX that follows the ANNE standard is essentially a “row” in a virtual table. The key difference with this virtual table is that it has a set amount of columns (a schema) that will NEVER change. An unchanging table which can store anything? I know, to many it sounds impossible… yet it is not. Every TX fits within the same virtual table as properties are themselves defined by ANNE TXs. New properties that are invented tomorrow get added as new neurons and then can be used as the “relationship” of future TXs. This performs the same function as creating a new custom column in a custom table but it doesn’t break everything that looks to that table nor does it require custom tables at all. For those in the DB world, think of a standard DB table turned 90 degrees on its side. One of those tables for all neurons (bitcoin addresses). Another for all ANNE TXs (relationships connecting neurons). Instead of a column for properties, like car color, and make, and model (and that’s just the cars table), we have properties such as color, make and model defined in the neuron table (which give them a bitcoin address) and then that address is used as the relationship of the ANNE TX in the ANNE TX virtual table. There are no actual tables, which is why I say “virtual”. These properties are all indexed as bitcoin addresses so the color of a car is defined by a standardized TX. Car->color->Red, but the Car is a bitcoin address of the instance of the car, the color relationship is the bitcoin address for the class neuron of “color” and red is the bitcoin address of the instance neuron for the color red. 3 bitcoin addresses wired together in a single BSV TX to specify the information that a particular car was the color red. The make of the car is defined by a standardized TX. Tire size. Engine type. Etc. We do not need 1000 custom tables or 1000 custom protocols to store cars, actors, presidents, countries, or anything you might invent today or tomorrow. And once you go to production but then say “oops”, we forgot something…. The future property that your project “absolutely needs to add” is simply added by creating a new neuron and using it …. in a standardized TX.
Any db tables can be generated, as outputs, of the ANNE protocol. Any format. Those decision need not be made, and locked in, and limit the capabilities. They are output choices not make-or-break storage decisions that come with trade-offs.
I have used this protocol in various incarnations to save corporations millions. I have made personal projects that I have enjoyed which offered me utility that no other apps offered. I had less to work with when shoe-horning this protocol into a relationship DB and forcing the DB to be something it was not meant to be. Graph DBs came along but were text centric and required custom code that worked with extra layers which changed with the wind. I grew tired of things breaking. This is in part why the notion of “lock the protocol” down spoke so loudly to me. It is why I loath the shifting sands that some *devs* create to ensure their job security but throttle the user. Now we have bitcoin.And with Bitcoin we now can encode information — and anything can be created with this standard. What you want to build can be encoded, registered, interconnected, built, adjusted, referenced and used within the ANNE protocol. Worlds await your inventive genius, your passion, your creativity. It is with neurons and relons that we can “tinker”. Any outside non-ANNE protocol can still be defined and indexed in ANNE and then referenced by ANNE. ANNE can still work with other worlds. Brains are built to be flexible.
So, in ANNE, it should be clear now that ALL concepts get their own Bitcoin address and with their own address they are address-able. If they are addressable they can be individually defined, individually linked to, used in relons and individually rated by users to provide new search indexes. They are anchors in the BSV graph. Permanent references. Resources for the world. Interwoven in that fabric is a hierarchy of interconnected classes. A taxonomy of sorts. These classes are defined with specifications. Cars have tires. Dogs have tails. Instances of these classes inherit the properties and the class definitions inform us of what we expect to see with an instance of this class.
Functionally, at the highest level there are 3 types of neurons:
- TEXT — a simple text string neuron. ie, the string “dog”. CLASS and INST neurons point to TEXT neurons to get their text)
- CLASS — defines a class of information, can be a sub-class of a higher class in the hierarchy, ie, dog)
- INST — a unique instance of a class, ie, your dog FIDO. (which overrides the class definition with a TX to specify he has no tail!)
Neurons (addresses) that do not connect to other Neurons (addresses) have NO inherent meaning. It is for this reason ANNE encodes information by connecting neurons and NOT by jamming in custom information into TXs that require custom ways to pull out each specific type of custom information. Neurons are made to connect and form circuits that provide meaning and value. We have the power to do this, now.
A TX provides meaning by connecting to other Neurons. All nomenclature aside, this is a simple standard on BSV using Bitcoin as the backbone for data structures that enable new downstream utility and a potential for intelligent services that are within our reach.
To the User:
Over the coming weeks and months we will be releasing features that enable you a whole new way to search, discover, and connect. The central theme is empowering the user. It is your identity. You choose what to supply. The data goes with you as you control the keys to your address. With ANNE we can enable a much more granular level of encoded feedback. I say encoded since I’m not referring to a text box of typing and sentiment analysis done on keywords in the text. We’ve been given thumbs, and stars, and hearts… but what do you REALLY feel? Why not just give you a direct way to specify it? In ANNE, that’s what FEELZ are. Being “interested” is not the same as “loving”. Feeling “amazed” is not the same as being “impressed”, yet many rating systems would bucket those all under the same response type. With ANNE, any feedback system can be created an used, but I have opted to first chase and lead with “feelz”. “feelz” will allow a simple positive (+) , negative (-) or neutral (/) or it will accommodate diving deeper into the type of feelz. They are bucketed by type (backed by research in the field) and it is up to the user to express what they want. With a new realm of expression there will be new indexes. You will be able to find content based on how specific population groups are feeling about the content. New utility and services will be available to those who provide this data that is a silver platter for intelligent services to work with.
New AI, recommendation, prediction, and matching services will become rapidly available.
To the Company or Business minded:
ANNE will provide an ecosystem of opportunity to build apps, websites and to build services for other apps and websites to use. Validation, AI, recommendation, matching… Reputation, cred, Q scores… These are few services that will be in demand that you can build, market, and offer your users in any way you see fit.
With worldviews you can build your own private (encrypted or not) world of truths that can be used in your business, apps, and sites. In the coming weeks and months we will release more about worldview and our vision of its role in ANNE.
To the Developers:
Over the coming weeks and months we will be releasing tools that help developers build sites, apps, and index their content in ANNE. We will release comprehensive protocol documentation (which you can already view some of in the code of this site). Once your data is indexed in ANNE as neurons and relons (addresses and TXs) all features of ANNE will be available to your users. Want to capture feedback? Want to provide advanced search? Content recommendation? Want to invent a new protocol and use it to build an app? All services built for ANNE data will be available to those using ANNE data.
Easily use content in ANNE to power drop-downs and content on your site. Use ANNE to connect your tweets, memos, articles, blogs, Bitcoin Assets or anything you can dream up. Connect them into the graph as apart of the circuit and many features reserved for big corps and well funded shops becomes available to your ideas.
We wire the brain to learn something new. Based on 2017 research by The Harvard University Center on the Developing Child it is now believed that infants wire 1 MILLION NEW CONNECTIONS per second. Those numbers are hard to comprehend but it clear that we do not need a new brain to learn something new. We form new connections between existing and new neurons. In order for us to do that in Bitcoin — we need a standard that creates neurons and relons. Building blocks.
I did what I could, up to this point, in code, to show with proof-of-work what can be done. I will be doing more… but for this idea to work… to REALLY work.. there needs to be an “us”. I need you. ANNE needs you, to see what this could be and become a part of using this new way of doing things. We will get to Metanet. Metanet is reality… and ANNE….
ANNE is Metanet’s little sister.