Create an Account
username: password:
 
  MemeStreams Logo

RE: Yahoo! News - BitTorrent and RSS Create Disruptive Revolution

search

Rattle
Picture of Rattle
Rattle's Pics
My Blog
My Profile
My Audience
My Sources
Send Me a Message

sponsored links

Rattle's topics
Arts
  Literature
   Sci-Fi/Fantasy Literature
  Movies
  Music
Business
  Tech Industry
  Telecom Industry
Games
Health and Wellness
Holidays
Miscellaneous
  Humor
  MemeStreams
   Using MemeStreams
Current Events
  War on Terrorism
  Elections
Recreation
  Travel
Local Information
  SF Bay Area
   SF Bay Area News
Science
  Biology
  History
  Nano Tech
  Physics
  Space
Society
  Economics
  Futurism
  International Relations
  Politics and Law
   Civil Liberties
    Internet Civil Liberties
    Surveillance
   Intellectual Property
  Media
   Blogging
  Military
  Security
Sports
Technology
  Biotechnology
  Computers
   Computer Security
    Cryptography
   Cyber-Culture
   PC Hardware
   Computer Networking
   Macintosh
   Linux
   Software Development
    Open Source Development
    Perl Programming
    PHP Programming
   Spam
   Web Design
  Military Technology
  High Tech Developments

support us

Get MemeStreams Stuff!


 
RE: Yahoo! News - BitTorrent and RSS Create Disruptive Revolution
Topic: Technology 8:32 pm EST, Dec 16, 2003

Decius wrote:
] Having said that, I'm recommending this article because these
] are fairly reasonable ideas. Yes your RSS aggregator should be
] signing up to receive a message anytime something new pops up,
] instead of going around to 1000 machines and asking every few
] minutes. As for using bit-torrent as an RSS feed distribution
] mechanism, yes, but no. Bit-torrent shouldn't know anything
] about RSS or vice-versa. My take is that bit-torrent should be
] pushed down below the applications, and it ought to cache any
] web page you hit if the owner marks it as cachable. Any url
] you try to pull should get pulled from bit-torrent first, if
] available.

I agree with Decius..

Better ways of tracking when content has changed are in order. Thats the key problem. Not the clients fetching the data, its the client's checking the data.. There are multiple systems out there already in a position to solve that problem, and they are likely going to be the ones to do it. Technorati, Newsisfree, Syndic8, etc.. A new pure distributed model isn't really necessary when the natural market place creates multiple sites that can play the role of content checker and knowledge cache. Its already distributed.. The solution to the problem lies in someone knowing what content you care about, and telling you when its changed.. That's could happen via some new great piece of distributed vaporware that's just gonna work and get widely adopted overnight, but its not likely. Its probably going to happen by subverting another key persistent connection the user has already had for many years that preforms a very, very, very, similar function: Instant Messaging.. [ think the presence mechanism.. not geting IMs.. -ed ]

Bittorrent like distribution models only makes sense for rich content. Something much larger then the average 10k - 30k size of most text only RSS documents. There is a certain overhead that would be created by such a network, and I don't see that as necessary unless the data being distributed warranted it. And wiring that into the syndication format is a bad idea.. The syndication format, as far as any given author/weblogger is concerned, is just their authoritative source what's available/new/changed.. It points to content associated with entries and the various methods of retrieving them, and the client can make a descision about how/if to get it based on that.. That is the point where bittorrent like models fit in, once there is content that is sizable enough to warrant the overhead of a p2p swarm/cloud kind of thing.

Even the current semantics in play enforce this way of looking at it. Its called a syndication format, not distribution format.

Don't read this as me totally dismissing the idea of distributed feeds.. Things like Konspire/kast highly interest me. I think they will be critical at some point when it comes to distributing content.. However, being informed content has changed or new content exists, and actually being given content are different things.. Its easy to lump it all together in the context of RSS when its just a few k of hypertext being pushed around, but that's going to change when more pictures and video start moving around. Those problems will not stay lumped together for long.

Right now we are still figuring out (and arguing) how the authoritative sources are going to work in what's already a pretty distributed environment.. The solutions that handle scaling and distributing the content will be built on top of that kind of thing, not eliminate or replace it.

Also worth noting, many people seem to be unaware that unless the RSS document has changed, the HTTP connection that checks is very small. If anyone is thinking that every one of those hits was the full size of the file, its not. Only if its changed. Hence, I worry this is leading people to do some very bad math in estimating their current overhead, and using it to justify that of a p2p network.. Granted, there are a number of sites that do not properly support cashing (we are still currently one of them), but that's their problem and their bandwidth. I think all the popular news readers get it right at least..

RE: Yahoo! News - BitTorrent and RSS Create Disruptive Revolution



 
 
Powered By Industrial Memetics
RSS2.0