If RSS becomes any more popular, it should be possibly to do pure "push" protocols without large changes on the client side (possibly by inserting a proxy along the network). Note that one of the great bandwidth savers is weblogs.com, which almost all MT-blogs ping to notify that changes are available. And DW controls that data.

As for the forking of syndication formats: DW made his own bed here. He never "gave away" RSS control to Harvard until Atom had already picked up speed.

And on the subject, any more progres with Jaeger?

[I am still working at it. I am sure I am so thick fingered at these things that I am perhaps not the best evaluator - but I am continuing.]

You are right that the one stop update point would save a lot of this my doomsday prediction. The trouble is there is nothing stopping the blogger, like me from setting up a separate RSS feed for every post and a million aggregators from grabbing each feed as a separate column on their aggregators. I alone could create 561 million requests across the internet every 30 minutes in that way.

You're the ideal evaluator -- I'm not aiming at techies. If there's problems, please tell!

And then there is this idea - which will soon have its own RSS so you can keep on top of whose feeds are added to the RSS list of feeds in common with the object of desire feed.

I'm planning to integrate them all into Jaeger in the same when Technorati is. I'm going to make a "Tools" submenu.

I am still trying to get my head around the RSS doomsday to see if my gut reaction has any validity. Put in context, according to the Internet Retailer website on 19 January 2004, in December 2003, the U.S. Internet population totaled 152.1 million users who spent an average of 27.6 hours for the month online. According to November 2003 stats, US web users average about 1 page per minute and are on-line for 3 hours for an average of 180 pages requests a day. That is a total of about 27.36 billion page requests a day. If, on average, aggregators make requests every 30 minutes, that is 48 requesting events per day. If the average aggregator requests updates from 20 feeds each 30 minutes, that is 1448 requests per day or 8 times the current average, were all US users to move to aggregation at that average. However, with both the proliferation and increased specification of RSS feed capability (my site now gives off about 20). It is far more likely that this traffic activity will be higher. This, of course, does not take into account the packet size which, for a bare request, might be expected to be much lower than average packet size. I would be interested to be directed to anyone with a better handle on the scale of traffic increase RSS success might be projected to create.

PS - my math skills usually suck worse than my spelling.

Keep in mind that if servers implements If-Modified-Since and ETags the amount of data transmitted for requests of non-changed data is neglible. Probably less than 1K in total.

Your webblog doesn't support this, by the way. Nor does blogger. Almost all MT weblogs do.

What I would really like to know is the capacity of the internet at any given moment, the volume of RSS feed requests it would take to fill that capacity (and as you say, not all requests are equal) and then divide that by the number of feeding sites or aggregating users to determine the average feed per site buy-in required to crash the whole thing. Is it an average of 1000 feeds per site or 0.0001? I have no idea.

Do a google for "death of usenet film at 11". The capacity of the Internet is infinite for text, both because text is small and the capacity of the Internet is expanding.

I do understand what you're getting at, but if the number of users of the Internet is U, and the number of feeds available is F, and T is the number of times/day one pulls data, the number of pulls is not U * F * T, but rather U * N * T, where N is the average number of feeds a person reads.

Now consider the case of _me_, and I'm a high end outlyer. U = 1, N = 200, T = 50: I'm pulling 1000 feeds a day. Now, let's say each of those pulls are 50K -- they aren't, the average will be more like 500 _bytes_ -- but let's say it's 50K just for kicks and giggles.

This means I'm downloading 50,000K of feeds of day: 50Mb.

Now, I'm on 1.5Mbit/s link, so I have the daily capacity of about 150K/s, or about 24 * 60 * 60 = 12,960,000,000. That 13Gbyte/day capacity on my cable line. So I'm using 0.3% of my line doing RSS downloads.

Except they're not really 50,000 bytes each, they're more like 500 bytes on average, so I'm only using 0.003% of _my_ capacity on RSS.

Well then why is there this idea of only checking feeds every 30 minutes?

Server load. I might only check 200 feeds, buy there'll probably be 1,000,000 people checking instapundit every 30 minutes by the next of next year.

Google phrase: power law distribution

Oh well, to hell with them. If your popularity is based on the technology buy the tools to service the populace.

I've just posted a new version of Jaeger. You should find substantially (I hope) easier to use. After you are finished reading weblogs, press "Recently Updated" again and everything you've read will be wiped. Double clicking it will wipe everything.

Please let me know if this makes it easier to use! (and that you got this message - there's something wrong with my e-mail)

So what does this mean? Any technology that relies on people "playing by the rules" is a pretty crappy technology. Rules in this sense only means limitations which should be pushed, broken and corrected - unless you are a person who has a faith-based relationship with the technology.

I just realized that syndication is like car radio re-sets. You never hear what you do not know already, you are comforted with the familiar.