Channel Multiplication

(Not sure if there are other pages that are named completely different and already meet this topic)

With ChannelMultiplication one can describe a phenomenon that can be wittnessed in many OpenSource software projects. Traditionally simple email, mailing lists and or the usenet where the primary channels to communicate.

Now we have:

And also the traditional channels have increased. A new channel like a new channel list is quickly created but hard to remove. The GNOME project now has 187 mailing lists alone. If you then also take the increase of wiki pages it becomes more and more hard to read and communicate in all that channels. This is not so much a matter of volume of communication but who really wants to subscribe to 187 mailing lists just to be informed?

In WikiWiki there is the WikiRefactoring as an important technique to handle the mess and clean up. But still what can be done with the increase of channels on all levels? (ThiloPfennig)

An increase in number of channels isn't something that should be avoided. To the contrary, any reasoned split should be embraced. Consider the issue from a different perspective: People are producing info-content - lots of it. In fact, the total content produced by people tends to grow linearly with the total number of people, and both the number of people (e.g. users of a product) and the number of projects (e.g. sub-projects for GNOME) tend to grow exponentially. If all communications were in one massive channel, one big deluge, it would be completely beyond any person's ability to comprehend or organize. Different channels, however, constitute different organizational units in the flow of information. They allow consumers of information to be reasonably selective.

The only real problem with such a large number of channels is that the channels themselves can be lost... a person with information that is best suited for Channel X knows not of Channel X and so creates a new channel, or sticks it on a channel he or she is more comfortable with. Your little comment at the top of the page is a fine example of this phenomenon: (Not sure if there are other pages that are named completely different and already meet this topic). As the number of channels grow, they, themselves, will need better organizational discipline, allowing people to more quickly discover them. Fortunately, companies like Google and Yahoo have stepped up to this massive task by working on search and discovery not just for channels, but for subtopics on channels. And, while their products aren't perfect, they're a huge step better than trying to find Blogs and Wikis or discover mailing-lists through other channels (friend-of-friend, telephone, e-mail, etc.). There is even more that can be accomplished on their end, such as real information processing rather than just Bayesian lumping of pages with similar words, and progress in this direction continues.

You said: who really wants to subscribe to 187 mailing lists just to be informed?. That is, perhaps, the wrong question. Would you really be better informed if you subscribed to 187 mailing lists? Or would you be just wasting extra time sorting through channels that aren't of interest to you? As a person who has some 100 Cable channels but only watches a few of them, and just a few series from those channels, I know my answer.

. I think you did not get my point. About the 187 lists: Many of them are dead but its hard for people to guess is they are. Same is true for IRC channels. Its not that all these channels exist at the same level. So you can end up asking questions in dozens of channels and getting no answer just because they are DeadChannels?. Not sure how a channel becomes dead but I encountered that often enough. I did not want to suggest that subsribing to 187 lists is a real choice. I rather wanted to say: Having the ability to really access all the information to the lets say 50 channels with content is what many people want - but its futile to do it with some channels. WikiWiki OTOH allows us to put all information in one big channel. There is still the problem of too many pages but still a page is more easy to access than the channels i talked above --ThiloPfennig

I understand your point; it's irritating to ask questions and receive no answers because the channel is 'dead'. But I also believe you're only reviewing one half of the problem, Thilo. There are processors on the receiving end of every channel, and they can only accept so much volume of input before they fry... and that applies to both human brains and computers. If communication channels DON'T increase at a rate at least linear to number of information-producers (e.g. people on the internet), it would become near impossible to communicate.

The problem, thus, isn't ChannelMultiplication. The problem is finding the right channel(s) - a search problem. For searches, unlike communication, it is best that they have all the information on all the channels so they can find you the right channels on which to perform a search. And if an automated search can't get you into the exact right area, things like hyperlinks provide some ability to perform further navigational searching on your own.

WikiWiki is a fine example of ChannelMultiplication: each page is effectively a communications channel, as is RecentChanges. You say that "there is still the problem of too many pages", but you're wrong. If people tried to communicate what they want to communicate on fewer pages, the pages themselves would simply become much larger and more difficult to follow. Hell, even as is, a great many of those pages become near-illegible ThreadMesses. But it would be far worse if you attempted to use fewer pages... fewer 'channels'. However, WikiWiki also represents a fine approach to the search problem. WikiWiki makes link generation and navigation easier, and it provides a search feature for finding pages. Of course, WikiWiki briefly predated the creation of a bunch of other Wikis, which then become 'meta-channels' in their own right (since there isn't much cross-wiki linkage or searching). Since the WikiWiki search doesn't possess information on all other wikis and such, it simply can't ever match a global system like google. (Of course, WikiWiki's FindPage is awful for many other reasons compared to state-of-the-art... even for searching inside WikiWiki, consider using Google's domain-limited search instead).

So, while I'm fairly sure I "get your point", I simply disagree with it. ChannelMultiplication is mathematically necessary for communication, and is, therefore, not a bad thing. But if you wanted, say, the ability to issue a search of the form: "Where can I ask <XYZ> and get an answer in less than an hour?", I'd be with you there. Search has a long way to go.

View edit of July 9, 2010 or FindPage with title or text search