|Russ Lipton Documents Radio
simplex veri sigillum
What Is Publish and Subscribe?
When you create a website (any website) and host it (or have it hosted for you) on the Internet, you have published it.
(Weblogs, in the best sense, are 'just' a different kind of website).
The notion of 'subscribing' to information is hardly new. Have you read your daily newspaper today? Delivering this in software - simply - has been a challenge for many years.
What It Is and Why We Need It
Without a publish and subscribe model, the Web suffers from forcing users to make explicit proactive decisions to go to a website of interest. If I want to read the New York Times online, I go there. It doesn't come to me.
This is one reason both discussion groups and mailing lists have been so popular on the Web. Mailing lists (always) and discussion groups (often) provide methods for notifying us that something of interest awaits.
Radio's delivery of weblog subscription feeds to us through the news aggregator is one of those 'simple' features that (will) revolutionize actual behavior on the Web. It works like this:
... A weblog author chooses to have their site packaged (automatically) so its posts can be passed to other weblogs in a standard, readable format.
... A weblog reader (me) asks the weblog of interest to send their stuff to me.
... When the weblog author makes a new post, the post appears not only on their weblog but within my personal weblog site.
(I can then post it to my public weblog but I don't have to).
Here is a fearless prediction:
Within 24 months, 75% of all new Internet websites will offer a subscription feature, probably using the very same protocol as Radio.
(I haven't yet but will add my documentation on this feature when I get good-and-ready. That is, when I can find the time ;-).
Discussion groups and email lists aren't going away. Why need they? (Instant messaging is another huge subject which I also reserve for another occasion.)
However, weblog publish-and-subscribe offers benefits that neither of these models provide.
Invitation-only discussion groups support near real-time back-and-forth that assemble conversations about the same subject in a single place. Publish-and-subscribe between weblogs cannot achieve this.
However, invitation-only and/or moderated discussion groups become limited and self-referential very quickly. They are best employed within professional groups to suit the purposes of a specific project. Otherwise, they wither over time as the subject matter is played out.
Unhappily, public discussion groups nearly always attract flaming over time. This drives away rational participants and kills conversation. Flaming flourishes because discussion group members can choose anonymous handles that eliminate accountability.
Weblogs, by definition, produced by individuals. People are less likely to write inflammatory materials whose only purpose is to wound another person when they can be identified.
Even more germane, a weblog is not 'in my face' the way a message is on a discussion group to which I am subscribed. I don't have to read your flames. I can subscribe to your weblog but I don't have to.
Nothing is permanent on the Internet (though the Wayback machine makes our postings more permanent than we may wish sometimes). Still, we can render weblog postings permanent more easily than discussion group or mail list items.
Cutting-pasting, clipping, storing, finding and republishing items from a discussion group or a mail list is a do-once-then-never-again phenomenon.
Because the weblog calendaring concept lends itself so neatly to archiving, a weblog's author (and, to a great extent, readers) can find earlier material that is useful at some point in the future. This will improve as weblogs offer more sophisticated search-and-retrieval of old posts.
You say 'huh'?
Both the strength and the weakness of instant messaging, discussion groups and mail lists are their spontaneity. It is very easy to say something. Consequently, it is very easy to say anything.
Because a weblog post persists over time, weblog authors are more inclined to take greater pains over their writing.
When responding to other weblogs - speaking metaphorically only, this resembles a slow-motion multi-weblog discussion group - there is a sense of .... publishing thoughtfully .... rather than reacting impulsively.
This is enhanced by the editable nature of weblogs. Nothing prevents me from editing this post, slightly or significantly, over time.
(Interestingly, professional journalists often mock this aspect of weblogs as though electronic bits on a screen must be frozen in concrete once displayed. Webloggers, by contrast, treat electronic texts much as vendors treat versions of software. Replace (not supplement) version 1.0 with version 1.1 when ready).
I will edit this story over time. Like, later today. Or next month. Or next year. Or never.
Managing Publish and Subscribe
Subscription features are still embryonic. Radio makes it a no-brainer to set this up but not necessarily a no-brainer to manage.
Ideally, I would like to subscribe to hundreds of Internet sources.
Unfortunately, this imposes some performance constraints (does the word 'slow' mean anything to you?) as well as information management challenges. I don't yet have good tools for filtering, sorting, extracting and managing my subscriptions.
Without those tools, I reach a limiting point where checking through my subscriptions is almost as tedious as proactively going to websites used to be.
In a nutshell, what we still need are ...
.... Faster Internet connections and super-duper algorithms for transmitting subscription feeds. Coming. Coming.
.... 'Smarter' ways to tell the news aggregator which items within a feed are likely to be of most interest to me so they announce themselves appropriately when they arrive.
(Caveat: this imposes a major classification burden on weblog authors as well as packing-unpacking burdens on the software news aggregators themselves. The jury is out on whether this burden can be reduced satisfactorily by more sophisticated authoring tools).
.... Better tools for extracting and archiving subscription items much as I can archive my own weblog posts.
But, hey, who's complaining?
My wish list is primarily just 'a small matter of programming'. The really hard part has been evangelizing a new way to look at the Web. Even that isn't done, but the snowball is gathering momentum.
Do you recall my prediction above?
I could be wrong. If publish-and-subscribe goes viral, 75% of new Internet sites will have this feature in twelve months. The fulfillment of my wish list won't be far behind.