[brudgers]

The problem with the utopian vision of the linked web is that links to meaningful content elsewhere on the web die.

In November of 1994, I created a personal web page. I linked the one image (the logo of the university where I intended to go to go to grad school). By March, the link was broken. The school had redone its website.

The world wide web broke the social contract implicit in Gopher. Geocities is no longer online.

[PaulHoule]

Linkrot is a real problem, but it's nowhere near as harmful as PageRank was.

Once links have commercial value, people don't want to give them away for free. Serious publishers quit making links to outside sites and soon other people got out of the linking habit.

[vog]

This problem has been solved 7 years ago. In 2005, Google introduced nofollow:

  <a href="..." rel="nofollow">...</a>

I believe that nofollow was mainly introduced to fight against guest book spam, so links in comments wouldn't be accounted by the PageRank algorithm, thus removing the incentive for that nasty spam.

However, you can also use nofollow to link to your competitors without increasing their page rank.

[PaulHoule]

that's indirection and psychological warefare on the part of Google.

in the big picture you give them data and they process it as they wish; "nofollow" is just a suggestion. for instance, they can compute metrics of it such as the ratio of follow/nofollow links pointing to a site and feed it into their scoring system

good links do get passed around in comments (this is the blog post you really should read about this) that shouldn't be nofollowed, while other webmasters are stingy and nofollow everything they link. so who knows what Google or bing does with this.

What is clear is that if you play it safe and give no link, Google won't give them any credit.

[brudgers]

I don't think page rank really plays as much of a role as the fact that it is easier to replace a webpage than to remodel it. See the "More" button at the bottom of this page.

[vog]

This reminds me of the words of Tim Berners-Lee: Cool URIs don't change!

http://www.w3.org/Provider/Style/URI.html

[progrock]

And at the same time info hoarding for the sake of it is pretty awful too. How do you wade through the vast amount of dirge that is out there in the sea of bits? How do you avoid feedback loops that occur from search engine lookups? How can I actually find a bit of useful information these days? I'm sure it used to be faster than it is now... At least we used to have website circles, that would filter some of the wheat from the chaff.

[_delirium]

That was one of the major challenges Ted Nelson's 70s/80s Xanadu project though would need to be addressed in order to produce a deployable hypertext system. It seems like they were wrong on that front: the web won out by just deploying without addressing it.

[acuozzo]

Ah, good ol' WorseIsBetter.

[davedx]

This is what HTTP redirects are for:

http://support.google.com/webmasters/bin/answer.py?hl=en&...

[chalst]

That isn't very useful if you are not sysop of the site the old content links to.

[davedx]

Of course. The responsibility for maintaining a link lies with the linked to site. My point is there are mechanisms that allow the links on the web to change without breaking it.

[protomyth]

The problems with hyperlinks are not a bad way to explain dynamic linking, static linking, and app bundles to a non or beginning programmer. Particularly when you throw in the example of a linked website replacing the image you're linking to with another image.

[edit] style sheets are the easiest since they can be internal to page, linked to same site, linked to external site.