[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


Janitor acceptance emails will be sent out over the coming weeks. Make sure to check your spam folder!


[Advertise on 4chan]


File: T429090.jpg (633 KB, 1385x3198)
633 KB JPG
CURRENT EVENT: archive.today is owned by a jihadist or some shit.

This is a general which is focused on archiving, but also interested in other related topics.

Storage technology and file sharing:
Hardware, software, services, shadow libraries, backups, home server, and networks such as tape drives, HDDs, file systems, archive.today, IPFS, Arweave, BitTorrent, etc.

Development:
E.g.: web archiving is much harder in 2026 compared to 2016. Too many websites are walled off by systems such as Cl0udflare, making it impossible for services such as archive.is to capture their webpages. That's a big chunk of important data that easily disappears with no web archive captures. We have to develop solutions to this, such as using the SingleFile extension and other stuff.

In-depth history:
E.g.: important web history events and future events such as sites closing, or get into the "minutia and trivia" about the history of websites and all the little changes.

Analysis:
E.g.: analyzing files and folders that you obtained from scraping or data hoarding, or, what you're sad was lost and not archiving, what you're glad was archived.

Questions:
Ask whatever questions about any of this.

Previous
>>109032656
>>
>archive.today is owned by a jihadist or some shit.
OP = screenshot of
https://archive.is/2026.06.17-092825/https://web.archive.org/web/20260616092752/https://phabricator.wikimedia.org/T429090

In the past I posted on KF forums that archive.today had some donation or about page which had a photo of a library bookshelf with Arabic text superimposed on it. I seriously kicked myself for not even fuckin screenshot-ing this finding. I haven't found that webpage again. (Search KF fourms for user PonyLover and you may find that post.)

Also, some may remember that years ago all the porn captures by archive.is redirected to
https://en.wikipedia.org/wiki/Pornography

It no longer does that.
>>
File: KiwiFarmsLogo.svg.png (34 KB, 960x362)
34 KB PNG
>>109075434
>I can not reproduce it, but according to this edit summary btdig.com also redirected to Terran Times: https://en.wikipedia.org/w/index.php?title=BTDigg&oldid=1358639447
>btdig.com is also owned by archive.today, see https://infosec.exchange/@iampytest1/116010542909144419 and https://en.wikipedia.org/wiki/Talk:BTDigg#btdig.com_is_not_the_same_as_btdigg.org
>Moreover, archive.today seems to have reverted the change as they no longer redirect to Terran Times when accessed with a Wikipedia referer.
I didn't know that archive.is owned https://btdig.com/ which is a very cool site, that would be way better without the Cuckflare wall. One of their subdomains is this:
https://web.archive.org/web/20251012055635/https://megalodon.jp/2025-1012-1455-53/https://en.btdig.com:443/bfe951d50130113368f0e3f8fff7ee9309585e9e

>>109075449
>KF fourms
by that I mean picrel
>>
>>109075434
Here's that "video proof" from
https://video.infosec.exchange/w/7hhdjQPSVSRziKjgMWEyPn

Wikimedia Meta-Wiki has blocked all archive.today domain names:
https://archive.is/2026.06.17-094501/https://web.archive.org/web/20260617092432/https://meta.wikimedia.org/w/index.php?title=Spam_blacklist&diff=next&oldid=30692621

All blocked except https://archive.closed.social/HpTh0 which is the Chinese or Japanese version of archive.is. Those links are in a spam blacklist so you cannot publish text matching
>\barchive\.(?:today|is|li|fo|ph|vn|md)\b
(regex) to Wikidata, Wikipedia, etc. (Cannot link to or mention archive.ph, archive.vn, etc.)
>>
does undetected chromedriver still work for bypassing cloudflare? used this for ripping a site but haven't looked into it ever since (a long time ago)
>>
File: 0wo4X.png (25 KB, 1024x768)
25 KB PNG
>>109075516
I hope that that Reddit tranny didn't fake that video because he hates that KF forums -- not excluded by archive.is -- has info about trannies that troons tried to memory-hole.

>https://archive.closed.social/HpTh0
Here's a capture of that minus the Cuckflare wall which shows up literally everytime I open that webpage:
https://web.archive.org/web/20260617101140/https://4440667c5a4692.lhr.life/memento/20260617091901/https://archive.closed.social/HpTh0

>>109075555
I've never heard of that. Maybe this will help you:
https://wiki.archiveteam.org/index.php/Dealing_with_Cloudflare
>>
File: puKpp.png (674 KB, 1024x768)
674 KB PNG
Glowniggers were attacking archive.today:

The FBI Is Targeting the Popular Anti-Paywall Site Archive.Today
>The site is still operational, but who knows for how long.
>The subpoena demands personal information for the site's owner.
>https://web.archive.org/web/20260429005136/https://lifehacker.com/tech/the-fbi-is-targeting-popular-anti-paywall-site-archiveis

FBI Tries to Unmask Owner of Infamous Archive.is Site - Democratic Underground Forums
>https://web.archive.org/web/20260617102728/https://the.democraticunderground.org/10143560152

FBI Demands Archive.is Reveal Owner in Legal Showdown
>Nov 9, 2025 - Summary - The FBI is attempting to identify the operator of Archive.is as part of a federal criminal investigation by subpoenaing subscriber information from domain registrar Tucows. - Tucows received a subpoena requiring it to provide customer details for archive.today, with a compliance
>https://digitrendz.blog/newswire/business/81634/fbi-demands-archive-is-reveal-owner-in-legal-showdown/

FBI Subpoenas Tucows in Search for Archive.today’s Creator Amid Privacy and Legal Tensions – Legal News Feed
>https://legalnewsfeed.com/2025/11/07/fbi-subpoenas-tucows-in-search-for-archive-todays-creator-amid-privacy-and-legal-tensions/
>>
Anyone else notice that since like 2 months ago, archive.is has been unable to directly capture upload.wikimedia.org links?

For example,
>https://upload.wikimedia.org/wikipedia/commons/thumb/9/9a/Don%27t_feed_bread_to_bandicoots%2C_Armadale%2C_WA%2C_Australia.jpg/500px-Don%27t_feed_bread_to_bandicoots%2C_Armadale%2C_WA%2C_Australia.jpg
gets forever stuck at
>https://archive.ph/submit/?url=https%3A%2F%2Fupload.wikimedia.org%2Fwikipedia%2Fcommons%2Fthumb%2F9%2F9a%2FDon%2527t_feed_bread_to_bandicoots%252C_Armadale%252C_WA%252C_Australia.jpg%2F500px-Don%2527t_feed_bread_to_bandicoots%252C_Armadale%252C_WA%252C_Australia.jpg
>>
File: OIP-2791843826.jpg (56 KB, 474x592)
56 KB JPG
(Something learned from the previous /asdiq/ thread: 4chan threads are bumped off of the board at page 10 and not page 11; if the bottom right "replies / images / page" has "10" for "page", you must push the thread up.)

>>109075516
They talk about that archive.today addition to the blacklist here:
>https://archive.is/2026.06.17-104330/https://web.archive.org/web/20260617092516/https://meta.wikimedia.org/wiki/Talk:Spam_blacklist
>Thanks for doing this, @XXBlackburnXx. We support your decision, and think this is a good step in protecting readers and other users from unexpected behavior from this site owner. EMill-WMF (talk) 23:52, 15 June 2026 (UTC) Reply
Cringe handholding. Accept chaos.

>>109075449
>kicked myself for not even fuckin screenshot-ing this finding
picrel
>>
File: FeGOD.png (126 KB, 1024x768)
126 KB PNG
>>109032658
Saw another related dead general:
>https://desuarchive.org/g/thread/87296636/
>What happened to /CUMG/? It used to be a good general before people started with that third eye shit. Now where am I supposed to discuss my porn hoarding addiction?
regarding
>https://desuarchive.org/g/thread/85293073/
>In this thread we discuss and create technology and software for cooming, data-hoarding, scripts, and more.
>>
If a per-torrent share ratio of more than 15.00 isn't good enough than what is?

See >>>/t/1400510 where an anon comes back to a 1-GB torrent that he shared to a ratio of >15.00 one year later and it's dead. Restarting qBittorrent sometimes fixes the problem of no peers with 100% of a torrent.
>>
File: rsPlL.png (45 KB, 1024x768)
45 KB PNG
>>109075434
SingleFile has this thing, which sounds kinda interesting:
>https://gildas-lormeau.github.io/singlefile-woleet/index.html
>Verify SingleFile Proof of Existence
>If you chose to generate a bitcoin proof of exitence in SingleFile preferences, it is possible a few hours later to verify the blockchain timestamp here by choosing or dropping the generated file. You can then download the proof receipt and keep it as an undisputable proof that can be verified independently.
>>
File: 7GVDR.png (114 KB, 1024x768)
114 KB PNG
"The most difficult thing about history is being there."

This is a quote from some archiving website (I forget the domain name). Part of archiving is catching things that would otherwise fall through the cracks or become lost media / lost forever.

>>109075928
A thread in yet another related dead general mentions /cumg/ multiple times:
https://web.archive.org/web/20260617115006/https://desuarchive.org/g/thread/99489610/
>>
File: 9r7z664wm1471.png (387 KB, 843x635)
387 KB PNG
>>109076058
An example of that is in the album "Family Guy: Live in Vegas". Characters say:
>https://web.archive.org/web/20260217181538/https://genius.com/Family-guy-pubertys-gonna-get-me-lyrics
>Lois: {gasps} Peter, you got me a ring! Oh, it's beautiful
>Peter: Eh, it's nothin'. I mean they charged me up the asshole for it - I had to sell my Richard Marx record collection of my 7th Heaven blooper reel where Stephen Collins says "Jesus Christ, I just stubbed my goddamn toe." Boy, did the WB try to keep that one under wraps!
>Lois: Oh...

Relatedly, I have an archive of the comments on deleted YouTube video "Family Guy - Stewie tries to kill Peter 1080p" as of 2020-05-07:
https://www.youtube.com/watch?v=atUzYVA4H_A
https://g10.vnar.xyz/raw/_F-CA5YTPGR6pA3eDfhZ-UpLiU7aP3e7lmh2WgVG1qc

What did the users say? Any signs of intelligent life? Let's look into this "atUzYVA4H_A.csv" file which was deleted off of archive.org/details/ by someone other than the uploader; some of the comments:
>why does it non stop keep zooming in
>Why the heck does it have to zoom in like this?
Shows that the video was modified to try to avoid YouTube detecting it as a vid containing copyrighted material.
>I like how the title has almost nothing to do with the episode
Commentary on clickbait in YouTube.
>I'm honestly surprised Louis didn't think her arms were clutter and cut them off
Tells about the plot of the episode seen in the YT vid.
>This episode describes the never ending hunger women have for more resources. Its not going to be enough and thats why they need logical people around them to say no or they train themselves and those women are amazing and usually not baby crazy anymore and post menopause....so we got crazy resource machines until they calm down and then its still hard but managable haha
Social commentary.
>Peter running on the bridge is the funniest visual gag I’ve seen on this show
TV commentary.
>>
I remember decades ago I was looking at a computer screen and said something like "Computers need to learn to speak English." I was thinking that what I was seeing was complicated or non-intuitive. Now, in current year, we have AI which means that "Computers can speak English." AI allows people to more easily do certain things. The controversial part is the quality of it and the ethics involved (as this one guy was saying).

Obviously, LLMs fail when it comes to specific archiving-related information. Here's some proof which is related to >>109076248
>Me: What's the title of YouTube video ID atUzYVA4H_A?
>duck.ai: [searching the web] I can't find that video ID — either it's incorrect, the video is private/removed, or search couldn't locate it. Check the ID and try again (it should be the 11-character string after "v=" in the YouTube URL).

And even if it could interact with whatever snapshotting website like web.archive.org and get the title, it can't tell me the comments on that video. That information is too esoteric. I imagine that in the future, large language models will get better at knowing or storing very specific info.
>>
I rediscovered the link to the thing that converts IPFS CIDs (Qm..., baf..., etc.) to the .data filename:
https://web.archive.org/web/20241125010048/https://go.dev/play/p/WOuGq_mJWbU

Now I need a thing to convert it the other way around, and I'd like to be able to run it offline. The .data file path looks like this:
~/.ipfs/blocks/25/CIQBRBDTGLSE6VUAJUZZXYBWRVQ2W4OE3M3YR5O7YOA7UDP3B5IJ25A.data
>>
How much redundant storage do you wish you had?

I know of some collections which I want. They amount to roughly 500 terabytes. So, I "can't see myself needing more than" 2 or 3 petabytes. Of course, storage power in this range is expensive.
>>
>>109075928
>>109076058
QRD: /cumg/ was completely banned because posters were embedding loli shit in PNG files and eventually posting illegal material using Leto the pedo's proxy. KF has a thread on him with an entire section on /cumg/ if you need more info.
>>
>>109079597
There's the type of fictional cp which is allowed by clearweb sites gelbooru.com and ATF booru. I'm fine with that. Then there's nonfictional cp which is illegal in America and like every other country. I don't like this. It creates problems. Pedos putting their shit where it doesn't belong is especially bad.

Reminds me of some pedo tactic that I read about years ago: they fill up a CD or DVD or flash drive with vile shit then physically leave it in some public space, waiting for someone to pick it up. I've seen weird stuff get left in public and not-so-public physical spaces (on the sidewalk or something), but nothing as blatantly illegal as that.

I never heard of "Leto the pedo's proxy" before.
>>
>>109077008
>CIQBRBDTGLSE6VUAJUZZXYBWRVQ2W4OE3M3YR5O7YOA7UDP3B5IJ25A.data
>https://archive.is/https://ipfs-02.hypha.coop/ipfs/QmPzM*
This is some media from the book "FREEDOM!" by Adam Charles Kokesh. It's some libertarian manifesto or something. It's licensed under the public domain:
https://en.wikisource.org/wiki/FREEDOM!

This .pdf book was in archive.org/details/ but it was deleted by someone other than the uploader (fuck IA).

As for converting, the CID to a datastore key and back again, so far I've cloned these repos:
https://github.com/ipfs/boxo
https://github.com/ipfs/go-cid
>>
>>109077008
>run it offline
Done. How-to: first create an empty folder, cd into it, then
>$ go mod init example.com/dskey
>$ vim main.go # paste in the code, change the "cid.MustParse" to a different CID
>$ go mod tidy # download things
>$ go build -o dskey # create the binary
>$ ./dskey # run it

Didn't do the "convert datastore key to CID" thing.

>>109080128
>https://upload.wikimedia.org/wikipedia/commons/5/55/Adam_Kokesh_-_FREEDOM%21.pdf?utm_source=en.wikisource.org&utm_campaign=index&utm_content=original
I read the "Racism" chapter in that book. It seems that he supports USA's civil rights due to ideas of "force" and "non-aggression principle":
>Using any judgement about a person to justify an act of force against them is wrong. Racism happens to be a very common justification and one of the most vile because it denies people their individuality in the mind of the racist who sees them only as a member of a collective.
Would the book's author be opposed to a "Whites only" store? Seems like it, unless I misunderstand.
>>
>>109080270
>Didn't do the "convert datastore key to CID" thing.
Done:

>https://github.com/ipfs/boxo/blob/main/datastore/dshelp/key.go
>// MultihashToDsKey creates a Key from the given Multihash.
>func MultihashToDsKey(k mh.Multihash) datastore.Key {
and
>// DsKeyToMultihash converts a dsKey to the corresponding Multihash.
>func DsKeyToMultihash(dsKey datastore.Key) (mh.Multihash, error) {

What file "main.go" was:
package main

import (
"fmt"

"github.com/ipfs/boxo/datastore/dshelp"
"github.com/ipfs/go-cid"
)

func main() {
c := cid.MustParse("bafy...bjou")
fmt.Println(dshelp.MultihashToDsKey(c.Hash()))
}


Changed to this thanks to slop:
package main

import (
"fmt"
"log"

"github.com/ipfs/boxo/datastore/dshelp"
"github.com/ipfs/go-cid"
"github.com/ipfs/go-datastore"
)

func main() {
dsKey := datastore.NewKey("/CIQETO7LBTHYWN4WGO47F2L276WRRAJD6YVS7OTBRXUG54ETEI2CS5I")

mh, err := dshelp.DsKeyToMultihash(dsKey)
if err != nil {
log.Fatal(err)
}

// Reconstruct CIDv1 using the same codec as you want (use cid.Raw for raw data)
c := cid.NewCidV1(cid.Raw, mh)
fmt.Println(c.String())
}
>>
>>109080490
>Changed to this thanks to slop:
Nope, this is stupid. It takes the data and creates a new IPFS CID from it. I need it to find the CID just from looking at the datastore key. More on this in a next post, I guess.

duck.ai reply:
>Understood — the datastore key encodes only the multihash, not the original CID codec/version. You can reconstruct a CID only if you also know (or assume) the original codec and CID version. Common approach: try sensible defaults and attempt converting to CIDv0 when possible.
>...
>There is no reliable way to recover the original codec/version from the ds key alone; this code tries CIDv0 when possible and otherwise emits a CIDv1.Raw. Adjust codec assumption if you know the original (e.g., cid.DagProtobuf).
me:
>but what about this function: func DsKeyToMultihash...
duck.ai:
>[blah blah blah]
me:
>[...]
and so on

>>109073945
>>109074518
I sycned a Linux video games torrent over ssh:
>$ h="Linux"; utc; rsync --partial -a --info=progress2 [src]/$h/ [dest]/$h/"
> 2,605,961,155 81% 1.06MB/s 0:08:58 rsync: [sender] read errors mapping "[src]/Linux/SiN/Sin.iso.bz2": Input/output error (5)
> 2,968,278,986 93% 1.04MB/s 0:03:27 xfr#22, to-chk=2/31)
> 3,627,638,995 113% 1.17MB/s 0:49:28 (xfr#25, to-chk=0/31)
>rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1852) [generator=3.4.1]
>$ # done: 2026-06-18T00:18:10.043873024Z

Oddly, even with that I/O error, the torrent checked to this while offline:
>Pieces: 761 x 4.0 MiB (have 761)
which is 100%.

So what were the I/O errors? Maybe errors with the metadata, or unsureness about the integrity of the file.
>>
>>109080665
>convert datastore key to CID
I had the LLM change the code to be better. However, it only worked in certain cases. Some of the CIDs it came up with from said key resulted in this error:
>$ ipfs dag export $str >/dev/null
>Error: root blk load failed: protobuf: (PBNode) invalid wireType, expected 2, got 7
>$ # trying CIDv0 or non-raw when it wasn't

Long story short, here's the better code by an LLM, minor corrections by me:
package main

import (
"fmt"
"log"

"github.com/ipfs/boxo/datastore/dshelp"
"github.com/ipfs/go-cid"
"github.com/ipfs/go-datastore"
)

func main() {
dsKey := datastore.NewKey("/CIQB46MJBZUVA2EWELIXMOJ2D7L7ZQRHTWMGPB5OTA2OKASUAPP52DQ")

mhBytes, err := dshelp.DsKeyToMultihash(dsKey)
if err != nil {
log.Fatal(err)
}

// try casting to a CID (may return v0)
if c, err := cid.Cast(mhBytes); err == nil {
if c.Version() == 0 {
// build equivalent CIDv1 using same codec and multihash
c1 := cid.NewCidV1(c.Type(), c.Hash())
// also build CIDv1 with Raw codec
c1raw := cid.NewCidV1(cid.Raw, c.Hash())
fmt.Println(c.String()) // CIDv0 (Qm...)
fmt.Println(c1.String()) // CIDv1 (bafy...)
fmt.Println(c1raw.String()) // CIDv1 with Raw codec (bafk...)
return
}
// already CIDv1: print original and also Raw variant
fmt.Println(c.String())
fmt.Println(cid.NewCidV1(cid.Raw, c.Hash()).String())
return
}

// fallback: construct CIDv1 with Raw codec
c1 := cid.NewCidV1(cid.Raw, mhBytes)
fmt.Println(c1.String())
}


This image is the CID resulting from running ./dskey in the past. It's an Angry Video Game Nerd (AVGN) pic.
>>
>>109077008
>>[other posts]
>>109080889
So what does any of this matter?

1. CID to datastore key. Take an IPFS CID ("Qm...", "baf...") and see where it would be stored in an Kubo IPFS repo. For example: ~/.ipfs/blocks/LC/CIQM4GPYRN7M3U3P45QWGMNBJIICW7OHUUEDEXPS5DY2W4BWCDBGLCQ.data = notice how the third to last and second to last characters, "LC", are the name of the folder in "blocks/"

2. Datastore key to CID. Maybe you ran "find $IPFS_PATH -type f" to see many ./.ipfs/blocks/ paths. What are the CIDs of those? The code at >>109080889 helps in finding that; it presents 3 possibilities for each "CIQ...". There's probably some other thing in the codebase or data which maps it one-to-one and not one-to-three where only one of the three is the correct one. Maybe something in ~/.ipfs/datastore/
>>
>>109060687
>>109060732
So a single 256-TB NVMe SSD cost between 46,080 and 256,000 USD. But then how long would it last? 5 years? 10 years? We can think of the meaningfulness of that amount of years in terms of the average human lifespan. But what about the meaningfulness of the data on the drive? And what happens to the data in the far future? Maybe storing a lot of the Arweave blockchain on it would be a good choice. Stats on that as of today -> Network Size: 20.6 PiB; Weave Size: 352.86 TiB.

Not that I would buy that SSD anytime soon, so call this daydreaming.
>>
File: videoframe_7051.jpg (1.14 MB, 1920x1080)
1.14 MB JPG
I have this artful nudity stage performance video titled "EROPOLIS rennes 2015 2 shows glamours" (topless only) from
https://www.youtube.com/watch?v=iXk6g33TFPU

This 1080p vid was deleted off of archive.org/details/ by someone other than the uploader.

I have uploaded it here to gain an amount of independence from archive.org:
https://files.catbox.moe/cc5lkn.mp4
>>
>>109081744
(Speaks to how much money you'd have to make per year as you'd have to keep buying more storage hardware after the drives die every 3 to 5 years.)

>>109082375
>deleted YouTube video which is also a deleted archive.org video at files.catbox.moe
Maybe catbox.moe will exist for decades. It has funding from both donations and a paid tier where you can upload larger files and poison the AI or something.
>>
File: BEbTK.png (16 KB, 1024x768)
16 KB PNG
Myrient was a website which contained <s>roughly 250 terabytes</s> more than 390 terabytes of non-current-gen video game data.

Myrient shut down on 31 March 2026.
>There are several reasons for the closure:
>- Insufficient funding
>- Paywalled download managers
>- Rising RAM, SSD, and HDD prices
>There were many other smaller reasons as well however they weren't disclosed.
>In short, they could no longer afford to run the site.

Sources:
https://minerva-archive.org/browse/
https://minerva-archive.org/faq/
>>
File: 1702504809919.png (94 KB, 225x225)
94 KB PNG
>>109083652
That's a screenshot of this page from a while ago:
https://minerva-archive.org/memorial/

It now says: "1,379,986 people have said thank you!"

>>109076058
>/fac/ File Archives and Compression General
A couple posts in that thread:
>> >7z
>> >the sleek carefully licensed opensource wunderkind
>> >recommended by 9/10 githubbers
>> >large storage archive completely dies if one bit is corrupted
>> >rar
>> >397928 million years old
>> >seen as outdated weird proprietary shareware hacker format from the beige box win9x era
>> >extremely resilient to damaged archives, only file under the corrupted area is affected
>>
>> I ain't gotta say anything more, eh?
> 1. rar still has the best speed/compression ratio, especially when it comes to extracting
> 2. rar has recovery records. 7z doesnt
> 3. the 7z dev is an incompetent Russian retard who thought it would be a good idea to fill the cipher IV with zeroes. After he was made aware of this, that bydlo fuck even had the nerve to claim that this wouldn't be a big issue.

I agree with the part about .7z not being fault tolerant. I've had experience or problems with that in the past.
>>
>>109083652
>Myrient was a website which contained more than 390 terabytes of non-current-generation video game data.
200 terabytes of that was the data that redump hashes refer to. http://redump.org/ doesn't have hashes of all editions of all video games. For example: the game "ODS Air Explorer" by Ohio Distinctive Software. redump.org has hashes of the Teacher's Edition but not the Student Edition (AKA Standard Edition); at least we have a torrent of both editions, where the data for the non-Teacher's Edition is non-redump: >>>/t/1383139

We know that there's multiple editions of "ODS Air Explorer" based on what the CDs physically look like, plus the data that they contain. We have photos of 2 of the 3 CDs. (Wish we had photos of all of them.) Looking at http://redump.org/disc/60920/ and http://redump.org/disc/60703/ these records are seen:
>Mastering SID Code = L806. Mastering Code (laser branded/etched) = SKY-CA CMCA 7983 AIREXPL U00505-26 CT-01-0306 CARNIVAL
>Mould SID Code = IFPI 9B78. Mastering Code (laser branded/etched) = METATEC R5RX203CC

Pretty sure that the mastering codes are all laser branded or etched, so they're text with that reflective property on the data side of the optical disc near the center. If it's a mould SID code then that's the tiny text imprinted/impressed on the plastic in the center of the disc. The mould SID code isn't printed (as in ink) on the disc. I'm glad that redump has that info; it's something of a record of the discs' physical appearance, and it can be helpful to differentiate or identify discs.

Attached pic is "Drivers and Utilities: Dell Dimension ResourceCD (P/N 9G705 Rev. A00)" from May 2001 with
- Mastering Code (laser branded/etched): RCD-405 Dimensions MADE IN USA BY Q-MEDIA QM
- Mastering SID Code: IFPI LK79
- a source scanner photo at https://02.aoar.io.vn/raw/6Y6lpdl-JubHWg-8Yeu23RKCajWc1CnSchQDZ-itW3o

>>109075961
Like 12 or 24 hours later a wonderful person showed up with 100% of that torrent.
>>
File: SPINDISK.gif (7 KB, 48x48)
7 KB GIF
>>109084049
>Mastering SID Code: IFPI LK79
These aren't unique to the disc. I used a search engine and found these discs which have that exact same SID Code:
>https://archive.is/2026.06.18-142516/https://archive.org/details/boole_pheromones
>https://web.archive.org/web/20230419001413/https://www.discogs.com/release/15467254-Johnny-Gimble-And-Texas-Swing-Just-For-Fun
>https://megalodon.jp/2026-0618-2327-08/https://www.ebay.com:443/itm/165775382826
and other discs.

>Drivers and Utilities: Dell Dimension ResourceCD (P/N 9G705 Rev. A00)
History of that in IA (fuck archive.org):
- September 11, 2020: uploaded at https://archive.org/details/resourcecd_pn_9g705_rev_a00
- March 31, 2024: same disc seemingly uploaded at https://archive.is/2026.06.18-133829/https://archive.org/details/dell-dimension-drivers-and-utilities-cd-inmage and compared to the other item the ISO/BIN has a significantly different filesize
- July 25, 2024: IA item resourcecd_pn_9g705_rev_a00 wasn't deleted yet
- June 22, 2025: that item (resourcecd_pn_9g705_rev_a00) was seen as deleted by someone other than the uploader

I have the larger .iso/.bin in IPFS in a ZFS mirror pool at
CIQA4RBWVXLW326JO25NDTYL5F6HTDB7RCACNK5KKUOLIETN7ECFYMA.data

Attached is an animated GIF in it.
>>
exist
>>
File: 1362841402193.gif (480 KB, 237x237)
480 KB GIF
This is a 1-terabyte torrent of full images from 4chan; it contains 1,105,578 files when unpacked:
>>109048515

For 15 days it sat there with no peers showing up with 100%, but today someone did. Cool! Here's an image from it. This GIF has a Unix Epoch of 1362841402.193 which is Saturday 2013-03-09 15:03:22 UTC.
>>
File: 1691025791761.jpg (116 KB, 1012x1033)
116 KB JPG
>>109086567
This has the time the files were uploaded to the server, but not the time the files were downloaded from some server. For a web archive, at least 3 things are needed:
- the URL
- the second it was downloaded (example: 20130309200322)
- the file (webpage or binary file)

Here's another image in that.
>>
File: 1594257896579.jpg (299 KB, 1600x1094)
299 KB JPG
I checked back on the Filecoin deals I made in 2025-05. I remember that they were all stuck on "deal status: Adding to Sector" for month(s). Now they all say "Error: deal proposal must be proven on chain [...]":
>$ export FULLNODE_API_INFO=https://api.node.glif.io; boost deal-status --deal-uuid=c05b707e-4528-4762-9672-4a1edb7441d8 --provider=f03312989; boost deal-status --deal-uuid=fd50e3fc-a948-4890-84d6-63329f581dcf --provider=f03312989; boost deal-status --deal-uuid=1349ef48-8e10-480e-8da1-bf5687cbceca --provider=f03312989 # 2025-05-20 to 2025-05-26 # boost provider retrieval-transports f03312989
>got deal status response
> [...]
> deal status: Error: deal proposal must be proven on chain by deal proposal start epoch 4992361, but it has expired: current chain height: 5344090
> [...]
> chain deal id: 116468127
>
>[...]
> deal status: Error: deal proposal must be proven on chain by deal proposal start epoch 4992372, but it has expired: current chain height: 5344090
> [...]
> chain deal id: 116468128
>
>[...]
> deal status: Error: deal proposal must be proven on chain by deal proposal start epoch 4992410, but it has expired: current chain height: 5344090
> [...]
> chain deal id: 116468129

I paid for each of the folders to exist in Filecoin for 3.5 years. I don't know what it's on about now, nor do I really care.

Filecoin has always been an annoying-to-work with and useless data storage blockchain in my experience. Arweave (and IPFS) on the other hand, have "always been great".

>>109086636
And another
>>
File: 6lgyr.png (274 KB, 1024x768)
274 KB PNG
>>109086778
Here's an academic paper on that:
>Degrees of Decentralized Freedom: Comparing Modern Decentralized Storage Platforms
>https://web.archive.org/web/20250722024316if_/https://tma.ifip.org/2025/wp-content/uploads/sites/14/2025/06/tma2025_paper16.pdf

It says:

>Filecoin only allows storing and retrieving files of size 512MB or greater.
Maybe that explains why I could never retrieve anything from this shit system.

>Similarly, Filecoin [9] is being used to store all data [10] for Internet Archive [11] and SETI Institute [12].
>[...]
>Notable adopters include the Internet Archive, which stores over 1 petabyte of cultural heritage data [11]
>[11] W. Hanamura, “What information should we be preserving in filecoin? — internet archive blogs,” Oct. 2020. [Online]. Available: https://blog.archive.org/2020/10/22/what-information-should-we-be-preserving-in-filecoin/
Idiots who wrote this paper have this contradictory part. Internet Archive stores only 1 PB of their ~70 PB total in Filecoin.
>>
File: 1695305238386.gif (605 KB, 220x182)
605 KB GIF
>>109086778
>Filecoin has always been an annoying-to-work with and useless data storage blockchain in my experience.
Only true when it came to deal with my files that I uploaded there. I did find other person(s)' data which was fun and/or great; I downloaded the CAR files from some storage providers' nodes (HTTP server at some IP address or domain name).

>>109087028
(Rather, "academic paper related to that".) Other quotes:
>IPFS, with moderate resource requirements and free-to-use model, maintains the largest network (23,000 peers). In contrast, Filecoin's high resource/participation requirements limit participation to only 1,700 peers. In terms of network availability, cryptocurrency incentive plays a crucial role. Platforms without financial rewards showed 50% of peers online for less than 4 days, while incentivized platforms exhibited median uptimes of 40-48 days. We studied the availability patterns between 2024 and 2025; our data shows that IPFS's peer availability declined from 60% to 40%, highlighting the sustainability challenges of systems lacking economic incentives. Filecoin and Swarm also showed decreasing network sizes as cryptocurrency prices fell.
>[...]
>When FIL price fell from $200 to $6, storage capacity decreased as providers exited the market. Storj, however, leverages a low entry barrier model that produces the opposite effect. As STORJ token value declined from $3 to $0.5, network capacity grew from 20PB to 100PB as providers deployed more storage to maintain earnings. These economic structures translate directly to pricing differences. Filecoin offers extremely cheap storage at $1/TB/month due to its low crypto coin price and high storage capacity. On the other hand, Swarm commands $335/TB/month because of the low supply. This demonstrates how incentive design shapes both provider sustainability and client affordability.
>978-3-903176-74-4 ©2025 IFIP
>>
>>109084215
>floppy disk
Makes me think of how about one week ago I accidentally bent a 5.25-inch floppy disk. I was grabbing or moving a keyboard and the QWERTY keyboard bent the disk by roughly one inch.

It's a black floppy diskette which has no labels or text. It only has the envelope it's in which looks like picrel.

>>109086778
For the record, those Filecoin deal details can be seen here (https://filfox.info/en/message/... webpage):
https://web.archive.org/web/20260618231137/https://6d353d3842e285.lhr.life/memento/20260618230759/https://archive.is/ow8xs
>>
>>109087331
Now, if only I had some device which could read that 5.25-in floppy disk. The data on it is maybe corrupted or bit-rotted by now. Or, it's a blank disk. I meet some tech bros IRL recently. I should have asked if they had the hardware to read this.
>>
File: denuvo.png (20 KB, 484x306)
20 KB PNG
I like to archive games that don't need to be installed and run by clicking the .exe, steam games are like these after you crack the steam dll.
I was wondering if I install a gog game, can I archive the installed folder the same way? I'm wondering if the .exe plays in any machine without installing, the reason why I'm asking is because there are several games on gog that don't have denuvo while they do have denuvo on steam and while denuvo was recently bypassed, it still hogs cpu resources for no reason so I'd rather archive the gog version and not the steam version.
hopefully someone would be able to answer this.
picrel are the games that have denuvo on steam but dont have it on gog
>>
>>109088187
>I have rom collections for ps3, ps4, xbox360 and switch already, they follow a different rule from windows .exe games
No way you have every PS3 video game: that's around 36 terabytes for a complete all regions redump collection. I assume you mean you have some ps3 games (not devaluing your efforts BTW).
>>
>>109088179
There's some GOG.com torrents which are terabytes in size: >>>/t/1389864

I can check if it has those games.

>>109088685
I made this post. It's maybe dumb. I think I misread >>109088187 as saying "romset collections" and not "rom collections".
>>
>>109088685
I dont do romsets, I curate a list of exclusives and get those
>>
Posting before thread gets beleted
>>
>>109088179
>>109088704
>I was wondering if I install a gog game, can I archive the installed folder the same way? I'm wondering if the .exe plays in any machine without installing
I don't know how GOG games work. I don't do that much gaming. If you have some GOG game downloaded, go ahead and see what happens. Run the executable and see if you see some installation wizard or files added to Windows folder "C:\Program files (x86)" or "C:\Program files".

In fact, I have some GOG games downloaded and can check around by using wine in gnu/linux.

>Total War: Something Something; Mad Max; Middle-earth: Shadow of War; Batman: Arkham Knight; Yakuza: Like a Dragon; Breath of Fire IV
>2019, 2023, 2015, 2017, 2015, 2020, 2025
I used exiftool and grep to search those two .torrent files for those titles. They have no filenames matching those titles.

(About said software: to see the version, run "exiftool -ver"; this is non-standard as it's usually "-v" or "--version". exiftool version 12.76 couldn't read those 2 .torrent files, but exiftool version 13.25 could.)
>>
>>109088179
>>109088817
I've looked into it and have seen a bunch of setup*.exe files from my download of that "GOG Collection" torrent. So the games must be installed and have an install wizard.

I don't know if you'll be able to find or create portable EXEs for GOG games.

I checked these:
>$ wine "./GOG Collection/rollercoaster_tycoon_2/setup_rollercoaster_tycoon2_2.0.0.6.exe"
>$ wine "./GOG Collection/rollercoaster_tycoon_3/setup_rollercoaster_tycoon3_2.1.0.16.exe"
>$ wine "./GOG Collection/rollercoaster_tycoon_deluxe/setup_rollercoaster_tycoon_deluxe_2.1.0.18.exe"
>>
>>109088880
So setup.exe for the game creates a game.exe in "C:\Program files". I know that sometimes those resulting EXE files are standalone and portable. I guess not for some video games. I haven't used Windows in more than a decade.

Unrelated to games, what would be the Linux replacement for JPEGCrops? I know I can run
>$ wine ~/Software/JPEGCrops0.7.5b.exe
or
>$ wine ~/.wine/drive_c/Program\ Files\ \(x86\)/JPEGCrops/JPEGCrops.exe
but can ImageMagick crop a JPG and only decrease the filesize?

JPEGCrops crops JPGs and does not re-encode the image. It's like how ffmpeg can run "-c:v copy", so ffmpeg can do whatever edits to a video but keep the same video stream; you don't have to re-encode the video stream every time you edit a video with -c:v copy.
>>
Is this whole thread just one autist? Is anyone actually reading his posts?
>>
>>109088915
>I know that sometimes those resulting EXE files are standalone and portable.
that's exactly the point of my question, I guess I'll download one and find out
>>
File: Fuck_IA.png (30 KB, 1024x768)
30 KB PNG
>>109088915
>what would be the Linux replacement for JPEGCrops?
duck.ai reply:
>What JPEGCrops does is lossless JPEG cropping: it crops on MCU (minimum coded unit) boundaries without re‑encoding the JPEG compressed data. On Linux the usual replacements are:
> jpegtran (libjpeg / libjpeg-turbo) — lossless transform/crop/rotate on MCU boundaries. Example: jpegtran -crop WxH+X+Y -copy none -optimize in.jpg > out.jpg. Note: crop will be rounded to MCU boundaries; exact pixel crop may not be possible without re-encoding.
> jpegcrop (from libjpeg-tools) — convenience wrapper around jpegtran for cropping.
> mozjpeg's jpegtran (or libjpeg-turbo's jpegtran) — same API, often faster and smaller output when using -optimize.
But is there a GUI JPEGCrops replacement in Linux? duck.ai reply:
>Yes — there are several Linux GUI tools that provide lossless JPEG cropping/rotation (i.e., operate like JPEGCrops by using jpegtran under the hood or similar). Options:
> gThumb (GNOME): Has "Transform" tools including lossless rotate and crop that use JPEG lossless transforms when possible.
> XnView MP: Cross‑platform viewer/editor with lossless JPEG rotate/crop options (uses internal/third‑party libs). GUI and fast.
gthumb installation failed in Arch Linux (guess I have to fix broken dependencies or whatever). XnView MP ain't GPL'd.

I have archived said version of JPEGCrops:
https://jembutkucing.online/raw/o_O67f-ffpcQRPtUjI_D_dJCaxtxb8DImCAVQHUTzeU

That file did exist in archive.org/details/ but the item (JPEGCrops0.7.5b_exe) was deleted by someone other than the uploader.
>>
>>109088952
>Is this whole thread just one autist?
Yeah dude, that's me. 90% of the posts in this thread are me.
>>
File: me.jpg (27 KB, 480x360)
27 KB JPG
>>109089002
>90% of the posts in this thread are me.
Actually 86.2745% of the posts above this post.
>>
File: supertard.jpg (469 KB, 1920x1440)
469 KB JPG
>>109089011
That's a classic Internet image, but I remember it being bigger. I used this AI image upscaler -- https://imgupscaler.com/ -- but the results aren't great.
>>
qBittorrent needs to have better information on torrents. I have 99.9% of this
magnet:?xt=urn:btih:4250507de6f96dfcfdb417e991cc656b50b18cd4
and it says "Pieces: 5533 x 1.0 MiB (have 5531)"

OK, but at what byte offsets are those piece missing? I could more easily restore the data from backup if I knew.
>>
*pieces missing
>>
>>109089149
qBittorrent has a visual indicator which results in the user having a general ideas as to what pieces of the torrent are or aren't missing. The details are in the .fastresume file, but there should should be an easy way of finding out the offsets of the missing pieces.

I'm glad that I had multiple backups of that dead torrent. For some reason my copy of that .iso in an internal drive matches to all but two pieces, and my copy of it in IPFS matches to all of the pieces. So now it's back to 100% thanks to my IPFS copy at
/CIQHB7YA4DSWZ7M37PQNNQPB76OGEIRXWXZAHL44VABUVB57YX7JURQ
>>
File: 1499706792707.jpg (45 KB, 465x465)
45 KB JPG
>>109089298
>i Tyrone all my HDDs. not my problem
Lulz, that's what I was thinking. Go and steal 256-TB NVMe SSDs from slop companies >>109081744
>>
>>109088952
Reading might be too strong of a word here, but I do skim the thread. Some of the info is useful even if you're not heavy into archiving.
>>
File: Bathtub_curve.svg.png (83 KB, 960x678)
83 KB PNG
>>109089446
>ALL HDDs will fail. It is occur a matter of time. This is why charging money for spinning rust, packed with helium that will fail is sinister and downright criminal. Especially when they don't pass the bathtub curve.
I was reading about this:
>https://en.wikipedia.org/wiki/Bathtub_curve
>Many but not all electronic consumer product life cycles follow the bathtub curve.[3] It is difficult to know where a product is along the bathtub curve, or even if the bathtub curve is applicable to a certain product without large numbers of products in use and associated failure rate data.
>The 'bathtub curve' hazard function (blue, upper solid line) is a combination of a decreasing hazard of early failure (red dotted line) and an increasing hazard of wear-out failure (yellow dotted line), plus some constant hazard of random failure (green, lower solid line).
>A product is said to follow the bathtub curve if in the early life of a product, the failure rate decreases as defective products are identified and discarded, and early sources of potential failure such as manufacturing defects or damage during transit are detected. In the mid-life of a product the failure rate is constant. In the later life of the product, the failure rate increases due to wearout.

So some HDDs don't follow their claimed bathtub curve because they fail way too often. I've had experiences with (and am continuing to deal with) HDD models which suck ass.

A related topic is the annualized failure rate (AFR) of hard drives. The likelihood of a hard drive failing is higher each year. You're testing your luck at year 3 more than you were at year 2.
>>
>>109080067
>I never heard of "Leto the pedo's proxy" before.
It's responsible for most if not all of the spam 4chan went through these past few years. Although now it seems to be under control, at least on /g/. Let's hope it lasts.

Also, regarding the flash drive thing, look into dead dropping, you might at least get a good laugh out of "archiving".

https://deaddrops.com/
>>
I have 136 MB of Windows Vista Sample Music (11 source files) which was deleted off of archive.org/details/ by someone other than the uploader.

Prior to me finding this data, I only ever listened to Windows 7 Sample Music (never really used Vista but did use Win10), so it was interesting to listen to this. One of the smaller audio files in that collection is "I Guess You're Right.wma", added that Rock music track to ar://:
https://grafananode.store/raw/Aj_xclh9LS8AnXVSzPSNaRT3pgEC0FqW_3gs8uOUEX4

Other genres or categories this sample music set in a certain OS features are: elevator music / Muzak, French or Spanish female singer, keyboard+violin ("Despertar.wma"!), ...

>>109090287
>deaddrops.com
Funny stuff. Such as the part where someone plugged a laptop into a USB drive embedded in a brick wall.
>>
I recently saw that this guy was vandalizing pages (or the site/account was compromised):
https://archive.is/2026.06.19-124347/https://web.archive.org/web/20260619124155/https://realbooru.com/index.php?page=history&type=tag_history&id=992973

The user added the tag "fuck_realbooru" and deleted other tags. No shit realbooru.com sucks, it's owned by untrustworthy deletionist collective booru.org.
>>
>>109090476
He pwned 6,695 webpages by changing or vandalizing tags like that:
https://web.archive.org/web/20260619124149/https://realbooru.com/index.php?page=post&s=list&tags=fuck_realbooru



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.