Hacker News Reader

Comment by amelius | Hacker News Reader

> The average lifespan of a web page is 100 days. Remember GeoCities? The web doesn't anymore. It's not good enough for the primary medium of our era to be so fragile.

> IPFS provides historic versioning (like git) and makes it simple to set up resilient networks for mirroring of data.

View on HN

IPFS is good and useful, but it only retains what people choose to retain.

If geocities.com/aoeu isn't popular, then IPFS won't store it unless someone bothered to pin it. And as soon as they stop, it'll disappear.

You need a dedicated host (like archive.org) to retain it, or volunteers willing to coordinate and commit their resources. Otherwise it's just more resilient (a good thing), but not permanent.

> Otherwise it's just more resilient (a good thing), but not permanent.

It's not "just" more resilient, It's also much more elegant and convenient: with the current web you need to go find some archive version of the dead link you found, while with IPFS the link can simply work, even after the creator stops hosting it.

If literally no one in the world bothers to keep a piece of content, not even archive organizations, what do you suppose could possibly work?

Hello fellow dvorakist!

IPFS is not a real solution at the moment. It's hard to use it, the default daemon is so agressive Hetzner nearly blocked my server due to it's scanning, and your site needs to be relative-url based to be put on IPFS.

On the other hand, nobody is talking about the problem of domains: yes, linkrot is a thing, but many are due to dead domains and dead blogging/content silos.

I've had to deal with Hetzner and IPFS too -- my conclusion is that it's Hetzner who are aggressive here. In of the cases I had fixed the dialing-local-networks behaviour, and then Hetzner still continued to block the server for about a week. They blocked it on 25-Dec and released it on 31-Dec.

> Remember GeoCities? The web doesn't anymore.

Bad example:

http://reocities.com/

Mixed example. Only a small fraction of Geocities pages / content have been preserved. Most of it is lost permanently.

Rather the opposite, a very large fraction of the original pages have been preserved. Many of them were deleted long before by the owners themselves and/or rehosted elsewhere.

Original Geocities content is probably the oldest large body of internet content that will be preserved for many years to come and it will look roughly how it looked on Geocities because it wasn't relying on much of anything other than basic HTML and some images.

For whatever reason, my site (Hollywood/Academy/4430) disappeared from Geocities before the mass archival happened. I certainly didn't delete it.

I search for it once a year, hope springs eternal.

Ai. I searched a bit for you, as well in the original crawls but to no avail. Most likely there were very few inbound links to that site which meant it wasn't discovered before it got wiped. I was still archiving stuff long after Geocities had officially shut down, for weeks new stuff turned up. 4428 is there, the next one after that is 4460, I suspect that if it had been archived you'd have found it by now.

Have you checked Jason Scott's torrent of all things geocities as well?

LOL, same here.

Yosemite/6022

Looks like I even got some of the images on disk here, but no html.

Jacques, I never knew it was you behind reocities, thanks for saving a large part of the old content geocities! The story on what you did at creating reocities alone is a worthwhile read. [1]

[1] http://reocities.com/newhome/makingof.html

edit: Just found again (and remembered) web.archive.org has 2 of its pages, so some of it is still out there.

I similarly look for the first page I ever created. I was maybe 14 years old and created a pretty thorough Goldeneye 007 fan site, with walkthroughs etc, on angelfire.

I distinctly remember a misspelling in my url--- /liscensetokill --but I can't be sure of my parent category anymore. I think /mi5/

Haven't seen it in over 15 years, though.

I was in geocities/South Beach/lounge but I can't remember the number and it annoys me. I know for a fact archive.org indexed it but they didn't at all after geocities moved to ~ urls. (That was after the Yahoo buyout IIRC)

Do you remember any text fragment? I can grep for it.

Either that or you can spend an hour clicking :) :

http://reocities.com/southbeach/lounge/

Thanks for the link! It looks like reocities didn't pick it up. (I downloaded all 250 links and grepped myself)

Clever. :( Sorry!

Reply on news.ycombinator.com