Consider a graph of world food production based on UN Food and Agriculture Organization statistics:
is a reproduction of a jpeg image taken from Freedom21's Meeting Essential Human Needs: Part 2.
Figure 3 Total food production in calories per person per day is more has increased by 25 percent since 1961 and is more than sufficient to feed every person on earth 2800 calories per day. Source: FAOSTAT Agriculture Data: Food Supply, Crops, Primary Equivalent. February, 2004.
http://faostat.fao.org/faostat/collections?subset=agriculture
According to the caption on the source image, the data was drawn from FAOStat. However, the current site only provides per country per capita calorie production. It isn't possible to produce a global graph from those numbers without knowing the population of each country.
A researcher working now can't tell if in 2004 FAOStat provided access to different figures or if the graph was produced from some other source.
This issue is fundamental to the client-server model that underlies . Data reputability flows from organizational reputability by way of the originating server. Derivative works can link back to the source URL, but there is no guarantee that the contents will not change. The only method to guarantee access to source materials is to make a copy, and even then it is not possible to access the underlying data.
is a set of formatting rules for structuring metainformation languages for text data. Source documents are ordered into trees by the addition of balanced "tags."
This paragraph contains italicized text!
]]>This paragraph contains italicized text!
p
) tags which represent semantically that all the contained text is a single paragraph. Similarly, the "italic" tags delineate a section of the string to italicize.
tags are required to be evenly balanced and with a single root tag which means documents are representable as trees. A standard isomorphism named the Document Object Model is defined by the W3C which also defines .
A tree, is ordered and there are several types of nodes, including elements and text.
is a method for transforming trees based on processing a series of messages isomorphic to the document tree rather than building a structure in memory. Events are generated when elements are entered or exited, and when text is reached (text nodes do not have children):
ToDo: Describe -1
is a distributed version control system which uses a novel structure for creating verifiable trees.
ToDo: Draw In-depth explanation with diagrams
is designed for managing versions of filesystem trees. It begins with the files and generates -1 hashes of those. Then it walks up the tree generating a hash of the names and child hashes for each node. When the root is reached, a single hash is left that can only be regenerated if all of the contents of the entire tree are reproduced. This allows for the verification of data hierarchies using a single value.
Cryptographic signatures are a method of verifying identity in a virtual environment. An individual creates a pair of large prime numbers, and computes a function on a bit string whose value allows a party to verify that either of the primes was used. One of the numbers is released along with the signature, and the other, so long as it is kept secret, provides a method of identity verification.
is a type of peer-to-peer where peers can register subscriptions with the network and conditionally receive content entering the network that meets matching criteria.
Freenet is a network which uses opportunistic caching. Once information enters the network it has no fixed origin, rather it is copied and cached according to metrics involving latency, bandwidth, and cache state.
Because Freenet's caching is based on request frequence, response time decreases in response to popularity, as opposed to server-based solutions where there is competition for a single resource.
The same-origin policy is a security measure implemented in all modern browsers. It restricts programmatic access of loaded elements to elements from the same server in order to protect sensitive information.
ToDo: Cross-site scripting attack example
ToDo: Prevented legitimate use example
The underlying data structure for producing a particular series of events does not have to be an document. The structure that I would like to explore for my thesis is generating content by traversing a multi-user graph database.
The graph database contains trees as well as structured data. events are produced by traversing the database, but data authenticity is provided through cryptographic signatures rather than originating server. This means that nodes can be reproduced without diminishing trust.
If the motivating example were prepared in this system, the original data would have been present in the graph. The selector for the desired set might look like:
The graph is then generated from the retrieved dataset in the form of a SVG, which is an format for describing vector graphics:
The root of this tree is linked to in the database and the user is able to add nodes and edges to annotate the diagram:
When the figure is referenced from another document, a link is added to the graph database and the source graph can be referenced directly.
Graphs are added to the database and a -like hash can be calculated allowing the subgraph to be copied through a network.
The end result would be similar to Freenet, with the addition of persistent sources for information. (Freenet does not store the source of data items to protect publisher anonymity.)
Conditional traversal allows for generating custom content based on user preferences.
ToDo: Diagrams…
One particularly interesting deployment scenario for this type of network would be to provision a municipal mesh network…