LLUP : Limited Lifetime Ubiquitous Protocol
What is LLUP?
LLUP is a notification system that can (but is not required) to use an XML-based element and attribute structured format that has been predefined to include:
- Sender Information (should match this to the 'author/name' and 'author/uri' elements of Atom):
- the senders name and the base URI the return message should be sent to. The base of this URI must match the base of the URI cotained in the 'link' element.
-
DateTime Information
- a timestamp based on the time the message was sent
- a start date that the content will be available
- an expiration date when it will expire
- Recipient
- a URI-based system, a blog acting as the perfect mechanism as it is already built upon a URI foundation, is
focused on an individual, is easily extended to allow for individuals to belong to as few or as many groups
of related blogging communities based on location, interest, job, etc...; still undecided if using standard
email addresses is something that should both be allowed and encouraged as one possible mechanism of delivering
using XMTP (this is something that needs a lot of attention)
- Region-based Web Servers; the blogXast server is the type of system that would be used to act as an interface
into a particular region of the World... Systems could use GeoIP, or simply just allow users to specify
"I'm in this zone" based on area code, postal code, etc...
- The regional-based system comes is built from three separate and distinct pieces
- Regional governments (citys, counties, states, provinces, countries, etc...) adopt this type of
system, building on top of it an extremely inexpensive, but highly advanced communication system
that allows for pinpoint precision as to the scope, while at the same time simply making the
availability of the information known, and allowing the subscription system already in place
with RSS and Atom to enable the ability to subscribe/unsubscribe
- Possible uses include notification of public hearings
- Community calendar events
- Severe weather early alert system
- Extension of Amber alert mechanism allowing the ability to send blip notifications to each device registered within the scope of the areas in which are considered likely locations the perpatrator could be, based on standard time since crime occured, likely traveling method, etc...
- This is the one area where Government mandated "forced subscription" might be focused on by political leaders. While this is the one area that goes against the "individual choice" foundation, its probably one that we should prepare and be happy to build into the mechanism based on the idea that a tool such as this, when in the hands of those in charge of the Amber Alert system, could literaly change the face of locating and capturing the criminals before there was even enough time for a crime beyond the initial kidnapping could be commited.
- Regional governments (citys, counties, states, provinces, countries, etc...) adopt this type of
system, building on top of it an extremely inexpensive, but highly advanced communication system
that allows for pinpoint precision as to the scope, while at the same time simply making the
availability of the information known, and allowing the subscription system already in place
with RSS and Atom to enable the ability to subscribe/unsubscribe
- Private business sector:
- Existing ping mechanism's such as Technorati, PubSub?, etc... can expand their focus into bringing greater value to site subscribers, fine tuning the categorization of data, delivering combined feeds of information that a user can rely upon as being spam free, focused content.
- Using existing web browsing features of mobile phones, a REST styled query mechanism can be used to search for restaraunts, movie times, etc... While obviously LLUP has nothing to do with the query mechanism or the content that is returned based on these systems, with the reach of the mobile phone market coupled with the "on-the-go" nature they suggest, the marriage between the server that receives the blips, categorizes, and then builds a natural query system and this mobile market is something that can and most likely will be very profitable for the companies that do it right.
- Etc... Etc... Etc...
- The regional-based system comes is built from three separate and distinct pieces
- Public interest sector:
- Everything from neighborhood block-watch programs, to impromptu book club meetings, to whatever else. This is where the social networking side of things becomes much more interesting, building upon a system of the "gotta have/do it now" tendencies that society has become increasingly focused towards.
- Region-based Web Servers; the blogXast server is the type of system that would be used to act as an interface
into a particular region of the World... Systems could use GeoIP, or simply just allow users to specify
"I'm in this zone" based on area code, postal code, etc...
- a URI-based system, a blog acting as the perfect mechanism as it is already built upon a URI foundation, is
focused on an individual, is easily extended to allow for individuals to belong to as few or as many groups
of related blogging communities based on location, interest, job, etc...; still undecided if using standard
email addresses is something that should both be allowed and encouraged as one possible mechanism of delivering
using XMTP (this is something that needs a lot of attention)
- Link
- Must point to a web-based resource; This base URI of this resource must match the base URI specified in the with the sender information. While that resource can obviously link to other places, automatica redirects are not allowed and if detected the blip message MUST be deleted, the base URI reported to a yet to be developed "watch system" that will implement a series of tests to determine the nature of the redirect, where it goe/what it points at, and making a decision as to whether or not this URI should be blacklisted based on historical data showcasing a history of attempted abuse. [NOTE: While this scares me too, Sean McGrath recently posted this > http://www.cse.ucsd.edu/users/tkohno/papers/PDF/< describing the ability to create digitial fingerprints of a device, using this for computer forensics etc... could also be used as criteria to determine a bonafied spammer, not matter how many URI's he/she attempts to hide behind]
- Summary
- A limited-character count, summarizing the content of the post at the other end of the linked URI
- Keyword(s) [I propose we change 'keywords/keyword' to singular top level Atom 1.0 element 'category']
- After thinking about your reference to RPV's, I realized that atom now implements the 'category' element which allows for the 'term' 'label' and 'scheme' attributes, without definition for what these must be. I used channelxml.org as the basis in the sample below, but obviously anything could be used, such as existing structures like DMOZ, or customized to allow for the dynamic changes in general day-to-day occurences (this is where ChannelXML could really shine...). Generally speaking these would be highly specialized keywords such as an individual (an actor or politician, each falling under some sort of DMOZ or ChannelXML categorization, but obviously not something that can be used as a general category encompassing many possible sub categories -- that is unless you thought of it from the perspective that an actor, politician, or even yourself, speaking in terms of the things that any of us might be interested in, the areas of ChannelXML you might subscribe to, etc... are going to be further back-up the heirearchy -- given this perspective then obviously any given entity within this system would need to be though of in more of a 'spatial' context, allowing for any entity to connect to any other entity, without traversing back up and down the heirarchal tree as necessary. Yikes! That one hell of a deep subject to think about...
I think its best that we get some input from Uche before trying to tackle anything from that arena ;)
While the sample below is not very well thought through(in terms of the category[@label|@term|@scheme] standpoint), its a start, and gives us a think or two to play with as we move forward in this space.
NOTE: Its almost eerie how well the category element and term. label, and scheme attributes fit into the RPV space. Hmmm... ya think Tim and company are trying to lay some groundwork in the RDF space, without coming right out and saying it (none of the attributes specify what they must contain... Which definitely seems like a "lets see who can figure this out first" kind of easter egg to me... Any thoughts?
Another note worth pointing out is that there really is no need for the top level 'recipients' element as well as 'categories' (what would have been 'keywords'. Theres no added benefit from an implied understanding as to what it means if there is a pluralized parent element or not, so why even have it? This is in direct response to the fact that the 'category' has no 'categories' parent within the Atom 1.0 spec... obviously they recognized the benefit was null. Following there example, and keeping as close to the original spec as possible, heres what I came up with.
Sample Atom Notification with integrated LLUP elements and attributes:
POST /blip-notification HTTP/1.1 Host: blip.ws Content-Type: application/atom+xml Content-Length: nnnn <notification action="updated"> <!-- NOTE: possible values for @action = created, updated, deleted. subscribe, unsubsribe would need to be added. I think. Russ, can you verify based on the development work you did if these are necessary @action values? --> <entry xmlns="http://www.w3.org/2005/Atom" xmlns:pub="..." xmlns:llup="http://www.x2x2x.org/2005/LLUP"> <id>tag:iblog.name/m.david/public/groups/comp/lang/lisp:1</id> <link rel="self" href="https://secured.iblog.name/m.david/public/groups/comp/lang/lisp/HelloLLUP.xml" /> <title>Hello LLUP</title> <published llup:expires="2005-09-15T00:00:00Z">2005-09-08T00:00:00Z</published> <updated>2005-09-08T00:00:00Z</updated> <author> <name>M. David Peterson</name> <uri>https://secured.iblog.name/m.david/blip-response</uri> </author> <summary>Heres a quick 'Hello World' sample for all you lisp weenies out there.</summary> <category label="computers" term="programming" scheme='http://channelxml.org/channels/top-level' /> <category label="languages" term="lisp" scheme='http://channelxml.org/channels/mid-level' /> <category label="sample" term="HelloWorld" scheme='http://channelxml.org/channels/specific' /> <llup:recipient href="nntp://comp.lang.lisp" /> </entry> </notification>