heh yup it was probably the CFC era link changing tricks that helped me figure out the atomic one as thankfully all the codes stayed the same! It was just the whole date backwards thing I failed to notice until Stromling came along and saved the day. And yes like you I usually check each lost download against the CFC and Strategy super backups there, but started to realise that they didn't include atomic user personal uploads which many Civ3 & 4 mod uploads appeared to be sadly, so thank goodness we've got a way to find them now! I'm planning on using this trick to get some of the Civ3 mods I've previously rescued using fan copies of (& compare) and I'm also going to use it to see if I can locate all the atomic hosted Civ4 mods @The_J
is asking for over in this thread:
I went through the Napoleonic one step-by-step tonight and now follow how it works. Good sleuthing by both of you! Definitely not obvious all the way through, but once you know the steps... I'll be uploading that one with a friendly name and restoring the link for 1.03.
Cool to see the Civ4 effort too, I think I saw that in passing a few months ago. My motivation has definitely been boosted by now having the ability to restore links, and not having to ping someone else for each one. As well as that there seems to be an upsurge of community involvement in restoring scenarios lately.
Hmm yes that ImageShack one looks like quite the challenge, especially as it appears to not let you see into these archives online which is annoying (as who wants to download all that lol). If you've got some example dead links maybe there's a nice code in them that might help find the file if we know the date of upload.
I do have some dead links! Back in the mid-late 2000s, most of the Stories & Tales used either Photobucket or ImageShack to host their images. One of the ImageShack ones is Aabra-cadaabra: Demigod Iroquois
. The first image is supposed to be right after where it reads, "After several restarts, I got this & decided that it was worth a shot". Its URL is http://img365.imageshack.us/img365/368/01beginnings1mq0.jpg
. In the spoiler is the "full image", whose URL is http://img110.imageshack.us/img110/9248/01beginningsfk0.jpg
The thread was started on August 24, 2008, so the images would have been uploaded then, or relatively shortly ahead of that time.
One of the caveats is that I don't have the actual images for probably any of the threads that used ImageShack (I used Photobucket, which has had its own issues, but not as bad as simply disappearing). So I can't say "that's the right one", although the context will likely make it clear enough. It's very likely a Civ3 starting area.
That collection lives at https://archive.org/details/imageshack.us
. A sample listing (chosen randomly) of the 40 items is https://archive.org/download/imageshack.us-20120918-180952
, and there are two problems. One, the larger files are generally not available for download. Two, I haven't figured out how to use the metadata files to figure out which one I would want (or even which metadata one I would want).
Some of them are recent too, e.g. from 2022. I don't know if those are new images from the new ImageShack or new uploads of old images. There is one hint I have; I'll use https://archive.org/download/imageshack.us-20220405-045657
as a sample. If I download the SQLite file (https://archive.org/download/imageshack.us-20220405-045657/imageshack.us-20220405-045657_meta.sqlite
), and open it in SQLite Studio, there is a "s3api_per_key_metadata" table with columns "s3key", "headers", "live_date", and "old_version_of", all of which except the last have data. Headers has a lot of data. So it could
be that it's possible to download them via an S3 call. I don't really know much of anything about S3, but the "host" in the headers is s3.us.archive.org, which sounds plausible. There's still a lot of questions about how to map a file URL to an archive, but maybe that's a way to see the data at least.
For that 20220405 link, the SQLite has 16 records in that table, a fillsource.txt and 15 which match up with the .warc.gz files listed on the web listing I shared above, which are listed as "not available for download". So maybe that is how it works, you have to use S3 rather than a web API to access them.
It definitely helps to have someone else express interest, writing this out has given me a bit more of a hint than I had the last time I looked at it!
Also possibly of interest to you, back in 2018 I made an archive
of some of the popular stories and tales, after seeing how many were being lost or rendered difficult to follow due to image hosts going down. The priority was making backups of Photobucket ones at that time, as it was still up but becoming increasingly less user-friendly (such as adding watermarks), but a concern was it might be the next to go offline. Figuring out the ImageShack archives, if they really do contain many of those images, would be the holy grail of Civ3 story restoration.
oh wow.. that's a very VERY interesting offer mate. I could have used your help a year ago when I did my huuuuuuge lost civ2 websites cataloging link list project
as heaps of them were geocities sites and some of those were only either partially or not on webarchive at all, and even those that were on there often had dead scenarios downloads not backed up sadly. Using the Geocities 'area codes' I was able to sometimes find alternative geocities backups on oocties.org , geocities.ws , geocities.restorativland.org , geocitiesarchive.org, and reocities.com for some fan sites, but in many cases there was still nothing on them either!
Too long ago now to remember anything specific but I could probably put together a list of the most 'stuffed' geocities sites taken from that catalog link list project that I'd like to take a closer look at. However as for missing scenarios, I think I've found elsewhere nearly every geocities based Civ2 scenario I was after thankfully. Although I've just checked the remaining entries on my super lost Civ2 scenario wish list
and can see there is a lost Civ2 Alpha Centauri mod hosted on geocities here:
The backup of his site is not great with plenty missing but if you follow the download links you end up at a geocities hosted download that sadly isn't in the webarchive backup here:
If you've got any ideas on that one I'm all ears mate!
Well, if it makes you feel any better, I didn't figure any of the GeoCities stuff out until December or maybe early January, i.e. after your project! The gist is that while the whole torrent is 2/3 of a TB, you can download it in sections. Then you can unarchive it through a slightly less than intuitive process, and then browse the various cities and sites you've downloaded.
This will be a great example for me to work out tomorrow and to finish writing that blog post! I can't remember what part I hadn't written out yet - maybe more screenshots? - but I'll try it on that URL tomorrow, polish up the blog post, and reply back with a link.
I should note (and the authors of the torrent also noted) that it is not 100% comprehensive, as they didn't have a 100% comprehensive list of all GeoCities sites when they were archiving it. My best guess is they had some sort of crawler/spider index to work with. The first site I was looking for was not present in it. But the one you listed might be.
I also had mixed luck with the online backups you listed. My best guess is that the torrent itself is as comprehensive as any of those, as I believe they were all started after 2009 (although I could be mistaken about that), and likely based on parts of the torrent. In theory one might think that one or more of them might have 100% of the torrent's contents available, but giving how spotty they seemed to be in my admittedly limited testing last winter, I'm not sure if that's the case. I do know that Jason Scott (of Internet Archive fame) was involved in creating the GeoCities torrent, so in theory they might have the best archive. But I also know that the first version of the torrent had some technical issues that made it nearly impossible to fully download, and that it was a challenge even for the Internet Archive (being an order of magnitude larger than other torrents at the time). The folks at the GeoCities Institute
helped them fix that, and it's their version that's recommended. They also have a pretty cool blog, where they interview people about the GeoCities sites they created, and share cool ones they've found while exploring the archives.
Also, fair warning, as a comprehensive archive, the GeoCities Archive contains both really cool stuff and not-so-cool stuff, notably no one ran a virus scanner on it while creating the archive. So I probably wouldn't go digging around it randomly with Internet Explorer 3 on Windows 98, even if that would be an authentic browser/OS combination for GeoCities.