PING Wallet · Topic 03Open the wallet →

The X-DOC

You write the X-DOC in the wallet's markdown editor and dress the page in a skin — instrument, take, typewriter, and so on. A skin is clothing: the same document in several looks, without touching the content.

The document is yours and it lives on the devices of the parties — not on a server, not in the federation. When you record it, it gets a seal: the hash of the content, the time in UTC and the link to the previous version. After the seal, the text no longer changes — but the skin does: changing the clothing creates no revision and does not break the seal.

It travels along with the memo, encrypted end to end: it leaves your device and arrives at the other party's. The federation nodes are pack mules — they carry a sealed package, and what they keep is not the document, it is the anchor: the proof of the time, without a word of the content.

Correcting is not editing: it is recording the next version. The previous one stays in the book, and the seal of the new one includes the seal of the old one. That is why the versions form a chain.

And it was built to hold up outside your own house: the seal can be asserted against third parties, and there are acts that give the document force in court — two witnesses turn it into an enforceable instrument, and initialling an arbitration clause makes that clause effective. That is in when the signature of the parties is not enough.

In this topic
  1. Recording always means sealing
  2. The boxes in the wallet
  3. The seed: why nobody writes in anybody else's book
  4. What the seal proves — and what it does not
  5. Trust leaves the counter and moves to mathematics
  6. The seal is perpetual

Recording always means sealing

There is no such thing as an "open document" and a "sealed document" as two kinds of paper. Every X-DOC is born sealed — sealed, immutable. What changes is what you ask for when you send it to someone.

There are three possible requests, and therefore three doors on the + button in the conversation. The door chooses the request; the document is the same.

the content travels sealed

The X-DOC is text attached to the memo — same envelope, same secrecy. Only your device and the other party's read what is written; the federation nodes carry a package they cannot open.

the seal is public

O hash is a short number computed from the text — change one comma and it becomes another, and from it there is no way back to the text. It reveals not a word of the content, yet it proves that this content is that content. That is why a third party — a court, an auditor, someone who has never seen the document — verifies without reading.

Secrecy and proof are usually a trade-off; here they are not. This is the difference between sealing a document and making it public: keeping a secret normally costs you the ability to prove, and showing it in order to prove costs you the secret. The seal undoes the dilemma — it is the proof of the content without the content.

Circulation is P2P

An X-DOC is private communication: it goes to no wall, it enters no public list, and nobody browses somebody else's document. It travels from device to device — from one person to another, inside the memo, encrypted end to end, like a transaction between two wallets. Each copy stays stored on the device of whoever received it, and verifies on its own. And it is a non-fungible: object — each document has an identity of its own and a chain of revisions that belongs to it alone.

the document goes from device to device · the node only carries

YOUR DEVICE NODE · PACK MULE THE OTHER PARTY'S DEVICE ANCHORTHE TIME ONLY BOTH SIDES · SAME SEAL
The document never stays at the node — it passes through. And what the federation keeps is not the content: it is the anchor, the proof that this seal already existed at that time.

What each door does to the chain

Here is the difference between an X-DOC and a token that changes owner. Sending a document does not take it away from you — and each of the three doors treats the chain differently.

sending is not transferring · the document stays yours

READ · the chain stays with whoever drew it up SEED · forks SIGN · no new link copy new chain, in the wallet of whoever received it the original one stays intact signaturesignature the same seal
Read copies, seed forks, sign creates no link at all — the consensus is two signatures over the same seal. None of the three takes the document away from the sender.

Why we do not call this a token. A token changes owner: it leaves one and reaches the other. Here, sending for reading leaves the document on both sides, and the seed opens a new book without erasing the old one. It is a copy with provenance, not a transfer of possession — and it is what lets the same document exist in several wallets without anyone having to ask a public registry who the owner is.

The document's statement

Like every object in the wallet, the X-DOC has movements — and they read like a bank statement. Each line is an act that happened to that document, with time and author; the seal sits at the foot, where it always was.

the same document · the line of what happened to it

14:0211:3709:14yesterday signed by both partiesread by the other partysent in the conversationrecorded — rev 1 rigor: electronic signaturereceipt stamped with the timedoor: for signingthe seal is born here SEAL
A document has no balance — it has a backing of acts. It is this statement that makes the X-DOC weigh as an asset in the wallet without becoming a token: what it accumulates is history, and history is verifiable.
the exception To be forwarded

The three requests above assume an author on the page. When the author cannot be there, the document still has a seal, a time and verification — and it does not compromise whoever wrote it. It is not a fourth operation: it is the fourth box, where the wallet gathers what you recorded anonymously and what reached you to be passed along.

read the topic →

The boxes in the wallet

Since there are three requests in the send, there are three boxes in your wallet — one per operation — plus a fourth one, for the exception: the anonymous ones. They sit in the main list, next to your conversations, and they show up even when empty, telling you what lands there.

They are not people. Where the list shows the round disc of a member or the square plate of a brand, an X-DOC box shows the whole sheet of paper, with no frame — that is how you read, at a glance, that this line belongs to the system and not to a person.

  • Open X-DOCs — the seeds that arrived for you to edit, and the chains you have already started from them.
  • Sealed X-DOCs — what arrived for reading only. It opens straight onto the page: with no invitation to edit, there is no choice to make.
  • Signed X-DOCs — what asks or has asked for your signature, including what is still missing. The summary counts it: "4 documents · 1 awaiting you".
  • Anonymous X-DOCs — the exception box: what you recorded with no author in the seal and what reached you to be forwarded, together. It is not a fourth request in the send — it is where the wallet keeps what has no author on the page.

State changes, and the box follows: a document that arrived for reading and later starts asking for a signature moves box. That is why the stable address of a document is never the box — it is the chain. The box is a view organised by what is expected of you, the same way the tape organises by what is left to decide.

The seed: why nobody writes in anybody else's book

When someone sends you a document to edit, your copy does not continue their chain. It opens a chain of your own, and the link between the two is recorded under the hash.

It looks like a technical detail and it is the central decision of the product. The alternative — a single chain passed from hand to hand — would require that, at every instant, exactly one person had permission to write, and that every device agreed on who that person is. With no server in the middle, two offline devices would draw up the next version at the same time, and the book would start asserting two incompatible things.

Notaries solved this centuries ago, and not with a lock: each scrivener has their own book. A new book opens with a transcript of the previous one, and the bond between books is a cross-entry. Nobody has to deny anybody else's writing, because nobody writes in anybody else's book.

Hence three consequences you see on screen. The sender does not stop working while you write — their chain goes on. The numbering belongs to whoever writes: your version 2 is your second act, not the second version of a document made by four hands. And the agreed deadline is a work target, not a lock: once past, it hands nothing back on its own — it is merely recorded that the work came after what was agreed.

What the seal proves — and what it does not

The seal is a hash: a fingerprint of the content. Changing one comma changes the whole hash. With it, anyone verifies on their own device whether the text in their hands is the same one that was recorded.

The seal proves integrity (this content has not changed), priority in time (it existed at this hour) and ancestry (this version continues that one). A signature is another matter: it proves will over the content. A sealed and unsigned document is dated text with authorship — which is already a great deal, and is not an agreement.

And there is a limit worth stating plainly: the chain proves itself backwards. From any version you reach the first one. The reverse does not hold: whoever recorded a document cannot enumerate the copies born from it across the network — they only know the ones somebody tells them about.

Trust leaves the counter and moves to mathematics

At a notary's office, what guarantees a document is the public faith of an official: you trust because the State delegated to that person the power to assert. In the X-DOC what guarantees it is SHA-256 — the same cryptographic function that underpins Bitcoin. The practical difference is this: the proof does not ask you to trust us; it asks you to check.

And checking depends on nobody. Anyone recomputes the seal of the content in their hands, on their own device, offline, with any tool — no wallet, no X-ID account, no permission to ask for. If the result matches, the text is the same one that was recorded; if it does not, the chain points exactly where it stopped checking out.

And we are not consulted in that calculation, on purpose: a system where only the manufacturer knows how to verify proves nothing, it merely promises.

What mathematics settles on its own
  • Integrity. The content has not changed since the seal — and that is verified with no intermediary, forever, even if X-ID ceases to exist.
  • Ancestry. Each version carries the hash of the previous one; each seed carries the hash of the link that originated it. Lineage is proved, not alleged.
  • Authorship of the act. Each link is signed by whoever drew it up. There is no record without an author.
  • Independence from the channel. The seal is born on the device, with no network. Reading, recording and checking all work offline.
What still needs an anchor
  • Identity. The seal proves that a key drew that up. That the key belongs to Bia Duarte is attested by the federation — that is trust, not calculation.
  • Exact time. The stamp comes from the device's clock. With no external anchor, "when" is a declaration; what the chain proves rigorously is the order between the versions.
  • Global ordering. There is no distributed consensus: nobody else validates, there is no replicated book. That is the price of working offline — and it is why the chain proves ancestry, not a single queue for the whole world.
  • Public faith. Where the law requires a public deed or notarised signature, the X-DOC does not replace the notary. It proves the content and the act; legal form remains the State's.

Why this is worth it for you: a notary records isolated acts — you bring in a finished document and pay for each stamp. The X-DOC records the evolution of a text and its propagation across your network: who wrote which version, from which copy each branch was born, who read it, who signed it. This is not a cheaper notary service; it is something the notarial book does not model at all.

And forgery, here, is not prevented by a guard — it is detectable by anyone. Whoever claims a false origin produces a hash that does not close, and a hash that does not close does not close in anybody's hands. It is the same old cross-entry logic, with the checking within reach of whoever received the paper instead of locked behind a counter.

The two counterparts

Everyone learned contracts on paper, and paper taught a choreography: I print two counterparts, sign both, send them by post; you sign both, keep yours and return mine. When mine comes back, the agreement stands.

The model is right about what has to happen in the end — and wrong about one thing only: believing that the sequence constitutes the agreement.

Look at the reason for the back and forth: ink cannot be copied. There is one sheet bearing my signature, and for yours to land on the same sheet, the sheet has to travel. And why two counterparts? Because the original cannot be duplicated. The two copies and the two journeys are not a legal requirement — they are a tax charged by the physics of ink.

What legally happens is simpler than the choreography: I express my will over the content, you express your will over the same content, and the contract binds when both wills exist. The paper is the bearer of those two acts, not their cause. A contract is no more signed because somebody went first.

one seal · two signatures · two identical counterparts

THE CONTENT, SEALED ONCE ONE PARTY'S COUNTERPART THE OTHER'S COUNTERPART signature Asignature B the same seal any order · in parallel
Both signatures point to the same seal — which is why neither depends on who signed first. And the two counterparts are identical by construction: on paper they are equal by trust; here, by mathematics.

Why the signature does not go into the content. The temptation is to chain it: I sign, that generates a new hash, you sign the new one, that generates another. It looks like the post — and it breaks.

Look at the end: I signed the first state, you signed the second, and the document left over is the third — which nobody signed. Worse: your signature would cover mine and mine would not cover yours, and the final hash would change depending on who clicked first. That is why signatures are detached, all over the same seal — which is how PAdES, XAdES and PGP's detached signature solve the same problem.

What digital removes is the tax: the signatures happen in parallel, in any order, and the agreement closes when the last one lands. What it preserves is the part that mattered on paper — each party ends up with a complete object in hand, standing on its own, asking nothing of any server.

And there is a flaw in the serial flow that custom hides: between my sending and your signing, I am bound and you are not. That is why a signature offer has a deadline — the deadline is what gives symmetry back to an act that is not simultaneous.

When the signature of the parties is not enough

There are cases in which the law asks for more than the will of the parties over the content — and almost always for a good reason: to protect whoever has less power in the relationship, or to give the document a force it would not have on its own. Which cases, and with what effect, changes from country to country, so this page does not list statutes: it lists the four questions to ask about your own jurisdiction. The doctrine holds anywhere; the statutes do not.

  • Is there an enforceable-instrument figure? In some jurisdictions a private document signed by the debtor and by witnesses can be enforced directly, with no full lawsuit first. Where that exists, it is usually the biggest change a couple of extra signatures can make — the difference between years and months.
  • Does the law require a signature from someone who is not a party? Spousal consent for a guarantee, a co-obligor, a legal representative. It is a signature by somebody who is not in the deal, and a document that has no place for it simply cannot receive it.
  • Is there a clause that asks for prominence in the wording instead of a separate signature? Consumer statutes often demand that a limiting clause be immediately understandable rather than initialled. Promising a gesture the law does not ask for is as bad as skipping one it does.
  • Are there levels of signature with different evidentiary weight? Simple, advanced and qualified is a common ladder, and acceptance outside the national PKI is often allowed when the party against whom it is asserted accepts it. That is where the four rigors of the X-DOC fit in: a rigor is not decoration, it is the level of proof that act requires.

Ask a lawyer in your own jurisdiction which of the four apply and with what effect. The question this chapter teaches is the same everywhere: what does the law require beyond the will of the parties? The Portuguese edition of this page carries the worked example for Brazilian law.

And one ritual the seal retires: initialling every page. It was never a legal requirement — it is custom, and it exists to stop someone swapping a page in the middle. With a hash seal that is impossible by construction: swapping a page changes the whole seal, and the swap shows up on checking. It is one of the few places where digital does not imitate paper — it makes the gesture unnecessary.

So that these reinforcements can be declared inside the document, instead of agreed outside it, the X-DOC has three markers in the editor — and they mark a place, never content:

:::sign
Where the parties signThe signature block on the page. The author says where it goes; the system is what fills in name, time and seal, at sealing time.
:::witness
Where the witnesses signThe section that can turn the document into an enforceable instrument. Kept apart from the parties on purpose: a witness is not a party, and the page has to show that.
:::initial
The initialling of one clauseThe gesture pointed at that passage — what the arbitration law requires. It generates a signature of its own, bound to the document's seal e to the clause identifier.

Why the author marks the place and never the signature. If it were possible to type the signature into the text, the content would become the place where one declares that somebody signed — and a document that can assert about itself what did not happen proves nothing. Marking a position is layout; writing a signature would be forgery. That is why the three markers are empty: they reserve the space, and the seal fills it.

The tokens are in English because the document is not. :::sign, :::witness e :::initial are the same word in every language of the federation — what changes is the label printed on the page, which follows the document's language. A translated marker would break every X-DOC on a language change, and the seal is perpetual.

The seal is perpetual

Nothing is erased to fix the past. Taking a decision back, refusing a signature, returning a piece of work — all of that is a new act recorded on top, with a new time, never in place of the previous one.

It holds even for what goes wrong. Work delivered after the deadline is not blocked: it arrives marked as late. A refused signature does not vanish: it stays as a recorded refusal, because a refusal is an act too. A notary does not prevent — it records.

Removing a document from your vault takes your copy off this device, and nothing more: whoever received it still has theirs, and the seal still checks out. Nobody has the power to undo a record in somebody else's wallet.