Why
external linking is a really
bad idea, in several easy steps; or,
things we don't always think about when dreaming up schemes for E2. Please leave these
writeups intact, this is a common idea.
Philosophical
- So external linking is really a problem for the purpose of creating self-standing content. We'd have a lot of people link to certain sites to pull in external content into our database. E2 doesn't depend on any external forces to work. We don't require anything else to retrieve our content. E2 is supposed to be an entanglement web of ideas and things, and compositions and people.
- Even linking to something as simple as our good friend Slashdot would encourage people to discuss /. posts for XP. This as with many other things in fortune cookies is true. Google? Nope, Google, can pull off stuff from a cache, so that's virtually every web page. Again, that violates the main point, in that people would refer to heavily to external sites to pull in node material.
- Screening of URLs would be a horrible process, especially with CGIs and such. With so many legal concerns over what is cool and not cool to "link" to, do we really want to pull our precious E2 into any sort of legal hot water? If someone C&Ps work here, then we can catch them, remove it, and apologize. We do our best here to remove offending materials. There's little way to police the links.
- People would link to more than just HTML files. It really wouldn't help much, but people would be doing things like pulling in graphs, and pictures and diagrams into their pages. While at first not such a bad idea, this could lead to widespread abuse, and would change the flavor of our db too much. It wouldn't be as solid and independent as it is now.
- External linking wouldn't encourage people to add anything here. We'd be a giant bookmark hive, and that really isn't what we are about. Small holes would be filled with external sites. Even if people mostly used them to claim their sources, there is little restraint on them as to how they would be used.
Technical
- We currently parse tags with the (I believe) parseTags() function somewhere in the nodebase, and it takes a setting and rips all but those HTML tags out of the nodeage. This is kind of suck, but that's how it works. It keeps people from adding shit like <table> tags to writeups when they shouldn't be there. Now that doesn't stop us from doing anything revolutionary, but there is no easy jive that sort of approach with our current system, but it still causes a problem.
- In parseLinks, there is an external link type, that I don't think ever gets used anywhere on E2. This is for the best. We use stuff like urlGen() and just flat string dumps to put together our href= strings for the like two external links we have here (in the footer and the like). This is really for the best. There will be more problems with people trying to figure out the new link type than there would be in its gain. Plus, we'd end up in a situation like eBay with a myriad of different URL and page tricks being pulled if we lightened up on the control of those tags.
All in all, these ideas about linking have come up a bunch of times, and more often than not, it's determined to be a rather bad idea. It "breaks" what we do here in general with
writeups and overall
content. Now there is a question for
linking on thing like non-standard nodes, such as user nodes (
homenodes). That is a completely other
debate based around what we feel people should have, and what they would do with them. For writeups however, I am most certainly against any sort of
addition.