Geocities To Be Made Available As a 900GB Torrent

An anonymous reader writes "Felt a shortage of the blink tag in your life lately? Well, have no fear. One year after Geocities was shut down in a cost-cutting move by Yahoo, a group self-styled as 'The Archive Team' have announced they will be releasing a ~900GB torrent file archive. It doesn't have every single site, but they believe they got most of it. The team believes that it's important to not just delete our digital culture, and as crazy as Geocities may have been, it was an important cultural milestone in the history of showing that anyone could create content online."

Comments

might be interesting to host it?

[phyrexianshaw.ca]

I'd be interested to see who would host something like that.

if I had the bandwidth for a good price, I'd consider it.

Re:might be interesting to host it?

[Anonymous Coward]

Considering that I'm already seeding a certain 790 GiB torrent this shouldn't be a big deal.

Re:might be interesting to host it?

[EkriirkE]

http://www.geocities.ws/ has been doing it for a while already.

Re:might be interesting to host it?

[Kjella]

However, its architecture is very well suited to doing exactly what you want, having content intelligently hosted by a large number of independent nodes.

Seriously, have you tried it? If you don't need anonymity, then Freenet is extremely underperforming and extremely dumb. A much better solution for this use case would simply be to create a special torrent client that would only store say 5% of the torrent, because you have global statistics at the tracker each new peer will download the rarest parts so in total you'd have a full seed. It could be trivially adjusted to work across many torrents, so that it'd continuously choose the least populated torrents and slowly rotate that content in.

Re:might be interesting to host it?

[Kjella]

Ok, so Freenet is not dumb it's blind. It's got plenty intelligence in trying to navigate in the dark, but it doesn't come close to a tracker that keeps an overview over who and how many have each piece. All the guesswork you do on Freenet can basically be replaced with "Part 252 exists in 12 copies. The IPs that have it are: aaa.bbb.ccc.ddd, ...." The performance would be just like a normal torrent because it would make direct connections, not via tons of other nodes. You could download the whole thing with a normal torrent client because they're all just act like torrent peers, they just start "seeding" after reaching their wanted percentage.

Here's a rough draft of the changes:
1. User sets a size limit, say 10 GB.
2. We calculate sum(size of torrents) = 1000 GB.
3. We download (10/1000) = 1% of each torrent
4. A timer will delete the 5% most redundant pieces
5. The client redownloads from 0.95% to 1.00%

Granted that doesn't actually balance across torrents, it gives each torrent a fixed pool relative to the total size so it could be smarter. But that's a first iteration that could probably be up in a day.

Re:Well, crap

[jlechem]

I actually had a few geocities sites I used back in college. It would be interesting to see if they're in the list. Sadly it will take me 4 months to download this with Comcasts 250Gb monthly cap. There needs to be an index of the sites so people can search through it w/o downloading the entire thing.

Its true. its digital history.

[unity100]

like it or not, its digital history. moreover, most of the early game cheat sites, content sites, predecessors of a lot of now-small-scale publishing operations and even some services started at blinky pages in geocities. have some respect. its like a 1902 model car in 1928 : it may look decrepit to you now, but when more time passes, the people who will come after you will see its vintage value. you cant, because it was just 1-2 decades ago for you.

Internet Archive should pick this up.

[Animats]

Good.

The Internet Archive should pick this up, if they haven't already. I'll talk to some people.

Archiving is getting easier. I had a minor part in preserving the archives of the Stanford AI lab. That required weeks of loading 6250bpi 2400 foot open reel tapes.

Re:On the contrary, the web must forget

[0123456]

Oh, and and for an example in the digital realm: the Apollo Guidance Computer is now flying Apollo spacecraft around the Earth and Moon in simulators, which was possible primarily because packrats kept old software listings in their basement for decades which were scanned, OCR-ed and then hand-fixed where required in order to recreate the original binaries to run in an emulator. The Saturn guidance computer which put the Apollo spacecraft into orbit is not, because there were no packrats to keep the software and IBM appears to have thrown it away or lost it.

Re:On the contrary, the web must forget

[SuricouRaven]

They would all be of interest to some historians. In the year 2500, a historian of the 20th century may ask the significent question of, say, "To what extent did the period of 2008-2012 see the revival of hobbyist mnufacturing projects, and upon their decline which activities replaced them?' Such a question could be answered by taking millions of those to-do lists, analysing them and looking at them statistically. Remembering to account for the inherent bias that only people quite organised will keep a to-do list.