When hypertext experts first saw the World Wide Web, one of the main reactions was criticism that it was too simple. In particular, the fact that links were made one-way, that there was no co-ordination between the site linked from and linked to, was considered a flaw. It would allow broken links.
Sure, that proved to be the case. Many links are broken on the web. But, as has been observed, this web works despite (or maybe because) of being slightly broken.
One clear advantage is that linking is easy. In particular linking is cheap because there is no co-ordination cost between the linker and the linked. It's cheap because the links are one-way.
A network which only depends on one way links is always going to be easier to build than a network based on two-way linking, that incurs a co-ordination cost between the two ends of the link. And because of this, such networks can be built quicker or cheaper or with one way links which wouldn't exist as two way links.
We have a web based on a one way system, and it grew very rapidly. Now we have ways to follow links back (eg. Google) but it's possible we'd never have had a web (or a Google) at all, if the whole thing hadn't been so cheap to begin with.
In general, when planning to build anything with a network architecture, ask whether it has two-way or one-way links; and if two-way, whether it couldn't be recast with one-way links.
- the invention of money turned the trade network from one composed of two-way links (barter) into one-way links.
the original ModelViewController pattern in User Interface programming. Each layer looked into (And had to co-ordinate with) the lower ones, but the lower ones didn't have to look into or co-ordinate with the higher one.
Perhaps keeping links 1-way is a heuristic for keeping a StupidNetwork. That also explains why a "bit broken" (WabiSabi) is a good sign in a network. It shows that the network hasn't been trying to be too smart by policing the connections.
Is there a conflict with NewMediaIsSymmetricalConversation?
Great article on Ning : Their storage engine is interesting. Every object has ID, app, user, tag, and type as metadata, and holds arbitrary key-value pairs. That's all the storage engine does, so it's very web loose-coupled. If your appointment data points to my address data, I can delete my address data and your appointment's left pointing to nothing. It's up to you to deal with references if it matters to you. Diego reckons it turns out to be not as big a pain as you would think, though that's what MySQL people said about the lack of transactions and eventually they had to buckle and add transactions. ( ) http://radar.oreilly.com/archives/2006/07/a_week_in_the_valley_ning.html )
: Interesting point. -- PhilJones
About two-way links, the ghost of Xanadu : (TedNelson) http://www.disenchanted.com/dis/technology/xanadu.html?ref=12 (TedNelson) which can also be joined with StigmergicSystems.
In fact, via the automatic BackLink detection of the above article, I also found more on two-way links from WeblogKitchen : http://weblogkitchen.com/wiki.cgi?BiDirectionalLinks
DaveWiner on the difference between one and two way linked SocialSoftware
See also :