i2p.i2p/apps/syndie/doc/intro.sml

Syndie is a new effort to build a user friendly secure blogging tool, exploiting the capabilities offered by anonymity and security systems such as [link schema="web" location="http://www.i2p.net/"]I2P[/link], [link schema="web" location="http://tor.eff.org/"]TOR[/link], [link schema="web" location="http://www.freenetproject.org/"]Freenet[/link], [link schema="web" location="http://www.mnetproject.org/"]MNet[/link], and others.  Abstracting away the content distribution side, Syndie allows people to [b]build content and communities[/b] that span technologies rather than tying oneself down to the ups and downs of any particular network.

[cut][/cut]Syndie is working to take the technologies of the security, anonymity, and cryptography worlds and merge them with the simplicity and user focus of the blogging world.  From the user's standpoint, you could perhaps view Syndie as a distributed [link schema="web" location="http://www.livejournal.com"]LiveJournal[/link], while technically Syndie is much, much simpler.

[b]How Syndie works[/b][hr][/hr]The [i]magic[/i] behind Syndie's abstraction is to ignore any content distribution issues and merely assume data moves around as necessary.  Each Syndie instance runs agains the filesystem, verifying and indexing blogs and offering up what it knows to the user through a web interface.  The core idea in Syndie, therefore, is the [b]archive[/b]- a collection of blogs categorized and ready for consumption.  

Whenever someone reads or posts to a Syndie instance, it is working with the [b]local archive[/b].  However, as Syndie's development progresses, people will be able to read [b]remote archives[/b] - pulling the archive summary from an I2P [i]eepsite[/i], TOR [i]hosted service[/i], Freenet [i]Freesite[/i], MNet [i]key[/i], or (with a little less glamor) usenet, filesharing apps, or the web.  The first thing Syndie needs to use a remote archive is the archive's index - a plain text file summarizing what the archive contains ([attachment id="0"]an example[/attachment]).  From that, Syndie will let the user browse through the blogs, pulling the individual blog posts into the local archive when necessary.

[b]Posting[/b][hr][/hr]Creating and posting to blogs with Syndie is trivial - simply log in to Syndie, click on the [i]Post[/i] button, and fill out the form offered.  Syndie handles all of the encryption and formatting details - packaging up the post with any attached files into a single signed, compressed, and potentially encrypted bundle, storing it in the local archive and capable of being shared with other Syndie users.  Every blog is identified by its public key behind the scenes, so there is no need for a central authority to require that your blogs are all named uniquely or any other such thing.

While each blog is run by a single author, they can in turn allow other authors to post to the blog while still letting readers know that the post is authorized (though created by a different author).  Of course, if multiple people wanted to run a single blog and make it look like only one person wrote it, they could share the blog's private keys.

[b]Tags[/b][hr][/hr]Following the lessons from the last few years, every Syndie entry has any number of tags associated with it by the author, allowing trivial categorization and filtering.  

[b]Hosting[/b][hr][/hr]While in many scenarios it is best for people to run Syndie locally on their machine, Syndie is a fully multiuser system so anyone can be a Syndie hosting provider by simply exposing the web interface to the public.  The Syndie host's operator can password protect the blog registration interface so only authorized people can create a blog, and the operator can technically go through and delete blog posts or even entire blogs from their local archive.  A public Syndie host can be a general purpose blog repository, letting anyone sign up (following the blogger and geocities path), be a more community oriented blog repository, requiring people to introduce you to the host to sign up (following the livejournal/orkut path), be a more focused blog repository, requiring posts to stay within certain guidelines (following the indymedia path), or even to fit specialized needs by picking and choosing among the best blogs and posts out there, offering the operator's editorial flare into a comprehensive collection.

[b]Syndication[/b][hr][/hr]By itself, Syndie is a nice blogging community system, but its real strength as a tool for individual and community empowerment comes when blogs are shared.  While Syndie does not aim to be a content distribution network, it does want to exploit them to allow those who require their message to get out to do so.  By design, syndicating Syndie can be done with some of the most basic tools - simply pass around the self authenticating files written to the archive and you're done.  The archive itself is organized so that you can expose it as an indexed directory in some webserver and let people wget against it, picking to pull individual posts, all posts within a blog, all posts since a given date, or all posts in all blogs.  With a very small shell script, you could parse the plain text archive summary to pull posts by size and tag as well.  People could offer up their archives as rsync repositories or package up tarballs/zipfiles of blogs or entries - simply grabbing them and extracting them into your local Syndie archive would instantly give you access to all of the content therein.

Of course, manual syndication as described above has... limits.  When appropriate, Syndie will tie in to content syndication systems such as [link schema="eep" location="http://feedspace.i2p/"]Feedspace[/link] (or even good ol' Usenet) to automatically import (and export) posts.  Integration with content distribution networks like Freenet and MNet will allow the user to periodically grab a published archive index and pull down blogs as necessary.  Posting archives and blogs to those networks will be done trivially as well, though they do still depend upon a polling paradigm.

[b]SML[/b][hr][/hr]Syndie is meant to work securely with any browser regardless of the browser's security.  Blog entries are written in [b]SML[/b] [i](Syndie or Secure Markup Language)[/i] with a bbcode-linke syntax, extended to exploit some of Syndie's capabilities and context.  In addition to the SML content in a blog entry, there can be any number of attachments, references to other blogs/posts/tags, nym<->public key mappings (useful for I2P host distribution), references to archives of blogs (on eepsites, freesites, etc), links to various resources, and more.  

[b]Future[/b][hr][/hr]Down the road, there are lots of things to improve with Syndie.  The interface, of course, is critical, as are tools for SML authoring and improvements to SML itself to offer a more engaging user experience.  Integration with a search engine like Lucene would allow full text search through entire archives, and Atom/RSS interfaces would allow trivial import and export to existing clients.  Even further, blogs could be transparently encrypted, allowing only authorized users (those with the key) to read entries posted to them (or even know what attachments are included).  Integration with existing blogging services (such as [link schema="web" location="http://www.anonyblog.com"]anonyblog[/link], [link schema="web" location="http://blo.gs"]blo.gs[/link], and [link schema="web" location="http://livejournal.com"]livejournal[/link]) may also be explored.  Of course, bundling with I2P and other anonymity, security, and community systems will be pursued.

[b]Who/where/when/why[/b][hr][/hr]The base Syndie system was written in a few days by [blog name="jrandom" bloghash="ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=" archive0="eep://dev.i2p/~jrandom" archive1="http://dev.i2p.net/~jrandom" archive2="mailto://jrandom@i2p.net"][/blog], though comes out of discussions with [link schema="eep" location="http://frosk.i2p"]Frosk[/link] and many others in the I2P community.  Yes, this is an incarnation of [b]MyI2P[/b] (or for those who remember jrand0m's flog, [b]Flogger[/b]).

All of the Syndie code is of course open source and released into the public domain (the [i]real[/i] "free as in freedom"), though it does use some BSD licensed cryptographic routines and an Apache licensed file upload component.  Contributions of code are very much welcome - the source is located within the [link schema="web" location="http://www.i2p.net/cvs"]I2P codebase[/link].  Of course, those who cannot or choose not to contribute code are encouraged to [b]use[/b] Syndie - create a blog, create some content, read some content!  For those who really want to though, financial contributions to the Syndie development effort can be channeled through the [link schema="web" location="http://www.i2p.net/donate"]I2P fund[/link] (donations for Syndie are distributed to Syndie developers from time to time).

The "why" of Syndie is a much bigger question, though is hopefully self-evident.  We need kickass anonymity-aware client applications so that we can get better anonymity (since without kickass clients, we don't have many users).  We also need kickass tools for safe blogging, since there are limits to the strength offered by low latency anonymity systems like I2P and TOR - Syndie goes beyond them to offer an interface to mid and high latency anonymous systems while exploiting their capabilities for fast and efficient syndication.

Oh, and jrandom also lost his blog's private key, so needed something to blog with again.