A Thread about Internet Archive's “Silent Killer”
twitter.com
twitter.com
https://en.wikipedia.org/wiki/Wikipedia:Link_rot
The Wikimedia Foundation's budget is about 10 times that of the Internet Archive. If you see the fundraising banner on Wikipedia and want to help out the site, but don't think the Wikimedia Foundation needs the donation, consider donating to the Internet Archive instead.
It's been awesome to see archiving and Archive links become standard practice on Wikipedia. For a long time, links just rotted and were replaced. Then later, dead links would be switched to Archive content if it was available. But it's so much easier and more reliable to archive a page when it's referenced, and provide both links even while the page is still up.
It's one thing to automate dead link detection, but archiving defends against a lot more. Universities seem to redo their sitemaps and redirect all their old links about once a year. And reference pages get edited, which goes unnoticed for a long time and then leads to frustrating [Not in citation given] tags. It's incredible what a boon the Internet Archive has been to wiki sourcing.
Links submitted to Hacker News should be auto-archived, too: I often stumbled upon dead-links [0] which otherwise had generated insightful discussion on news.yc. Adding archive next to web would work nicely.
[0] Launch HNs, in particular.
https://gitlab.com/tildes/tildes/issues/586
I believe all link aggregators should do this, they're a great way of getting defacto quality/interesting content discovered by the archive.
It's been on our list to do that for some time, but this should make it much easier. Thanks!
Hook it up to RSS and away you go.
It is immensely important, especially as there is a strong trend to hide the content behind the paywalls and registration pages. Even if something is accessible today...
And there are always forces who have an interest that some inconvenient data just disappear.
Also, there is a 2:1 donor match thing happening now, making the value of each donation effectively triple. Moreover if your company matches 501(c) donations (mine does, at least up to something like $20k a year), you can have them match your donation too, effectively giving the archive a 4x your original donation.
Deceptive tactics like this are toxic to charitable giving culture.
The Internet Archive is a world wide treasure. Please donate!
And it's really easy to set up an automatic monthly donation as I have done. One time donations are great. Ongoing income helps insure that this precious resource can continue to fulfill their mission.
Wouldn't it be 6x, assuming your company could have their donation tripled as well?
If you happen to have power to do that at a large cloud provider, consider donating storage space and data movement services.
Having a backup of the Internet Archive on cloud providers would offer an additional safeguard against data loss.
And on that train of thought, I just had a little flashback about storage sizes. There was some time when 3.14 MB was a big unit of measurement, and some 50MB drive was huge. The classical example: Monkey Island, Indiana Jones or Day of the Tentacle on ~12 floppy disks. But you had to choose which 1 or 2 to install because you didn't have enough space on your hard drive, or you had to swap disks every few screens :)
Or, they are archiving a lot of video playthroughs in the lets-play style from video sites, in case those sites go through a meltdown like viddler does.
I guess I'm rambling. Point is: This isn't just a storage dump. There are also interesting projects around the internet archive to make the old things accessible on new systems. Very worthwhile donating to.
http://appletimewarp.libsyn.com/apple-time-warp-podcast-epis...
I don't know what storage medium that refers to. Plain old 3.5" floppies only went up to 1.44 MB.
(Still sore Apple did not stick to the 1kB = 1000 bytes of the original MacOS and gave in to the "powers of two" insanity ;-)
Actually, Apple switched back to 1kb = 1000 bytes ("back" assuming you're right about the original MacOS) with Snow Leopard back in 2009. [1]
[1] https://blog.macsales.com/1852-snow-leopard-changes-they-way...
I'm not entirely sure how well they were supported under Windows though.
That 'scene' is also full of drama and very insular. Donate to IA to help them build better infrastructure and better search. But donating because they host old abandoned DOS games I would not call a good reason, and misplaced. Because the people doing the majority of the work are doing it because they like it, not money. IA does have some work around that such as getting dosbox and mame running in a browser. Just be clear on what you are donating to.
I found out that the local games scene I used to love as a kid had been almost wiped off the face off the earth. These were small non-commercial and shareware games localized in just one language so it was already a niche. The free hosting services of the 90s are gone so those sites are down and no one wanted to keep paying for hosting for 20 years+ on sites that get very few visitors nowadays.
The only way to get these games again was to find a Discord group and a friendly stranger who agreed to seed a torrent (which had 0 seeds when I found it). I'm looking to upload them to a couple of different places and compile a basic website catalogue (static site on CDN) one of these days. For the layperson, these games are already gone from the internet.
The internet archive does a great service but it is breadth-first and quite surface level. The depth has to come from people who were familiar with the sites at their peak. And there's a big change that no one is doing that for your specific interest.
I know many people feel that everything should be available forever. But for me... it's pushing me away from doing much on the web. I liked it in the 90s when things were more ephemeral. When you could make mistakes and not have them easily found by anyone with a few clicks, forever.
I'd argue no, and there's a huge amount of waste in there - so many bot posts, or just spam.
But here we are, people want to archive every byte of information that traverses through the internet.
These days you have to approach cautiously, as every thing you do or post may be archived.
That was obvious in the 90s.
But sadly enough, it hasn't worked out that way. Before the Internet Archive anyway.
And obviously this is already happening— in the Canadian election a few months ago, there was a big scandal where some pictures turned up of the prime minister wearing blackface at a holiday party he attended in 2001. And on top of that there was a pretty steady stream of candidates (some of whom were indeed booted from their parties) who were challenged over social media comments/posts made 5+ years earlier, especially on hot social topics where national sentiment has rapidly evolved in recent times.
Perhaps we will all just become inured to this, and rightfully be able to accept our leaders as human, judging them for who they are today and not past words and deeds. But is it possible to still retain the ability to fairly judge someone in the present while forgiving the past? Or will we lose the ability to discern the difference? The GOP's attitude toward president Trump does not give me hope on this.
And, although I assume I'm in various Usenet archives, my participation there was always from a work address and was pretty tame. (BBSs mostly were too although all that content is gone anyway AFAIK.)
I could definitely picture political parties and businesses being willing to pay $$$ for an internet presence dossier as part of their larger candidate vetting process.
For example, about Sabu etc: https://sites.google.com/site/avalonlogsefnet/
Oct 29 12:28:53 <Cambion> we have logs that date back to 1993 and identify pretty much everyone from nickname change to nickname change
Oct 29 12:28:59 <Cambion> phone numbers, everythingI thought the rareness of such finds is what makes it a treasure, not the content itself.
https://www.businessinsider.com/the-internet-archives-100-ce...
This argument is so horrible as to be actively harmful to Archive's work. Jason Scott is a god, and if we didn't have him, we'd have to invent him.
WE DO NOT GET TO CHOOSE WHAT THE FUTURE FINDS INTERESTING.
We live in the only point in human history where we can actually save all of humanity's knowledge and culture, and we can do so without having to worry about physical space or staff to work the "library." It's a remarkable time we live in, and yet, 99% of our society either doesn't care, thinks this work is stupid, or actively works against it through horrific copyright laws.
We know more about how Rembrandt painted and lived than we do about how Atari 2600 programmers worked and lived. I can go to Rembrandt's house and see where he lived, where he painted, how he worked, where he slept and ate and mixed his paints and taught his classes.
Atari's old HQ is just another office building. The source code to those games is mostly gone (thankfully, it's assembly and easier to disassemble). We need to save our culture and digital heritage, else we forget where we come from.
Deleting some old tweets is one thing, but actively worrying about Archive's work is just harmful to us all. We need 10,000 more Archives, dammit. It's supremely important work that is helping stem the tide of lost culture due to stock market forces. Geocities is gone forever because Yahoo! didn't find it profitable. This cannot keep happening.
I’m also not sure your Rembrandt example shows what you suggest it does. The average Atari 2600 programmer would be more equivalent to the hundreds of now unknown artists in Rembrandt’s time. The John Carmack’s of today will be remembered in detail with or without blanket archive efforts.
Maybe, just maybe, Rembrandt’s Status in our minds is a result of generations of people each seeing the individual value in his work. That is, each generation does indeed get to decide what future generations remember. Or at least it used to be true until the digital age.
Maybe the change is an improvement. But maybe not.
And libraries are the epitome of what you’re fighting against. They are by definition works chosen by humans based on judgment calls of their perceived value.
Let’s at least acknowledge that blanket archive efforts are a fundamental change in themselves and a departure from the human status quo for thousands of years. Then let’s debate whether the change is an unabated good.
To my mind, this was the only known find of a vintage Atari 2600 game and its iterative development process. So, 30 years later, the only reason we had this snapshot is because someone found these chips and sold them at the flea.
The current state of digital preservation is abhorrent. Those roms would have taken up less than 1/4 of a 5.25" floppy, but the company behind them never thought to preserve that information or data.
Take2 Interactive republished BioShock in 2012. They couldn't find their source code. They didn't save it. They had to go machine to machine looking for it. The reissued game is not the same as the original.
As a society, we don't place any value on this stuff, but the potential value of it cannot be understood until the future has occurred. Letting it vanish is a disservice to the future. In the past, if a book was published, it wasn't going to vanish if the publisher went out of business, there would simply be no new copies.
In our digital online age, things vanish in seconds, days and hours. This is also a very different state of affairs. In the past we could not save everything, but everything didn't have a clock counting down from the end of the quarter over its head, counting the seconds until it is deleted.
The Library of Congress tries to save everything. Yes, libraries weed the stacks and choose items to host. This is due to space concerns: they can't host everything ever. Digitally, they can, and many host reams of microfilm and old newspapers because they can.
Libraries can, thanks to tech, now host every book ever, digitally, for very low costs. Copyright prevents that.
This is an unabated good. Leaving things behind and forgetting them is how you get Tulsa Oaklahoma, or the Armenian Genocide denials. We don't get to choose what the future finds interesting, and for the first time in history, we do not have to. Why in the every loving fuck would you worry about that?
Most likely, only for personal reasons. This is a humanity level problem. Your personal worries are irrelevant in 100 years when everyone who ever knew you is dead anyway. Geocities would be more interesting at that time, as a subject of study.
The bar - perhaps found accidentally - seems quite important in deciding what must be archived, and what probably shouldn't.
Releasing whilst the probably famous sender is alive is most often in the realms of to do damage, simply tasteless or paid for revelations in the gutter press.
There is plenty of evidence for the Armenian genocide, the Holocaust, and 9/11. That doesn’t really stop deniers or conspiracy theorists. When it becomes politically advantageous, spreading misinformation becomes weaponized and mainstream. A bunch of nerds saving some ROM dumps isn’t going to really change that.
Like the library of Alexandria it’s also quite idealist to think archive.org will be around in 100 years or more. Not that we shouldn’t do it... but the future can be unkind to even all modern technology.
(amateur digital archivist and data recovery hw/sw dev here, I find this stuff fascinating!)
I'm not convinced that we should feel a compulsion to save all of that. Just because it's more practical to be a pack rat about digital content doesn't mean that, taken to extremes, it doesn't still seem like being a pack rat.
All the stuff Tumblr users intentionally wrote and published publicly, but none of their IP address logs and other incidentally collected information, is exactly what ought to be archived and preserved, in my opinion. This is in strong contrast to incidentally collected data including clear PII like IP addresses that many companies today are hoarding forever, when they ought to be ephemeral.
There's certainly value in archiving social media but I think it has to be balanced against the harms, instead of defending the practice with literal religious fervor and dismissing all criticism out of hand.
Don't want it archived forever? Don't put it on the Internet. Seems simple enough.
There obviously should be ways to ameliorate your mistake, which is why it is absolutely critical that Archive.org has a removal process.
I agree that there is a danger that people may not realize how public and permanent the things they published to Tumblr were, or how dangerous it can be to do so (and I downvoted a sibling comment dismissing this danger). However, I think you and I have different threat models.
In my mind, archiving PII that is intentionally published is not particularly harmful because most lay people do, in fact, understand that their avatar, username, and by default, posts are public on Tumblr. They have had the opportunity to remove that information this whole time, and they still do, Archive.org removes stuff if you ask them.
By contrast, lay people have no mental model for what kind of information is incidentally collected nor how dangerous or benign it is. Certainly, lay people also can and do misjudge how public and how dangerous the things they intentionally publish are, but the gap is far, far less than incidental information. "Would you tell a stranger this" or "would you write this on a bathroom wall" are decent heuristics: the only difference in danger between text written on a bathroom wall and written on Tumblr is due solely to the potentially wider reach and possibility of even going viral on Tumblr. (Photos, of course, can also subtly compromise privacy in ways surprising to a lay person, but the gap is still much smaller than incidental information.)
In my threat model, that gap in understanding is much, much more dangerous than the intrinsic danger of PII. That's why I think that as long as Archive.org has a usable removal process, I think pretty much all the danger is in surveillance capitalism's collection of incidental information, not Archive.org's permanent record of intentionally publicized information.
Yet such data is invaluable to historians and can give us a window in time through the eyes of people who lived that time. Having that publicly available data lost for all time would be an immense loss to future generations.
I'm sure in a few generations, some historians will study those archived porn blogs and get an insight on the evolution of humans' relations to sexuality that today's historians can only dream of.
Sure, but this leaves us a distorted view of history, where we have lots of details on the lives of "great men" and next to none on how ordinary people lived. Which means the vast majority of people who lived and died in that period end up written out of their own history.
Archaeologists spend a lot of time rooting around in ancient rubbish piles and cesspools, because these are some of the very few places where physical evidence of how ordinary people lived has survived. Nobody in ancient times would have nominated those sites as culturally important or worthy of preservation. But what we know of how ordinary people in those times worked, played, ate and drank comes largely from things dug up from them.
Because the the welter of proprietary, undocumented formats, media bitrot and the like we are actually moving away from such a point.
Turns out historians may not be so upset. You can be a historian of early medieval France have a chance of reading 100% of the surviving documentation. Too much data can obscure the story.
Of course historically you could get a PhD for compiling a concordance to Shakespeare, something that can now be done mechanically in seconds. Future historians could (and will) apply the same tools to today's surviving documentation. But I don't believe there'll be as much of it as you seem to think.
The best we can probably say is that it's different. We're capable of saving far more but, in practice, a lot of digital media is locked up in walled gardens and accounts that have to be paid for and require logins.
It's presumably easier to save a bunch of photographs or videos in a way that they'll be accessible so long as key Internet sites or their successors are. A fire or flood probably won't destroy them. OTOH, unless you've taken affirmative steps to upload that media to the right place, it won't be serendipitously discovered in a shoebox some day in the future.
I don't understand this reasoning. Yes, more data = more work, but less data = more likely you're wrong.
The incentives in academia are messed up.
Or are you advocating for the idea that history is arbitrary and it’s better to just have a simple story than to have to worry about what really happened?
I’m having a hard time understanding what idea you’re trying to position in this debate.
1 - it's more likely we will be an information-sparse region in the historical record rather than an information-dense region.
2 - professional historians have their own set of incentives which can be counterintuitive to the layperson.
3 - but indeed if there turns out to be a huge amount of stuff (there will likely be mountains of some forms of ephemera) to go through some people may be able to find value using new tools not available in the past to historians.
As someone trained as (but never worked as) a historian I do indeed have a bit of cynicism on point 2. I suspect most if not all actually working in that domain have the same cynicism.
And that's France. What about Medieval Africa? The Americas?
Oh, we mostly burned that documentation because "it wasn't important". Right?
It is more like, either you burn the library down or every thing you have written in your private journal is now available to be checked out by anyone.
It really shouldn't be that way and I think we should fix the problem of holding people responsible for bad behavior in the past. But how do we draw lines (for example, what about holding people responsible for past crimes).
>We need to save our culture and digital heritage, else we forget where we come from.
I agree, but we also need to ensure this is done without costing individuals. Technology has advanced, but society has not. Out technology outpacing our culture has and will continue to hurt many people and we should try to find a way to fix it.
Missed opportunity in that the book was only readable if you took it as a deliberately over the top "if this goes on" fable. And the film was mostly notable for how it squandered a top-notch cast.
There is value in forgetting. What we forget, then is a very relevant question. "NEVER FORGET ANYTHING RAWR!" is not a useful point of view because it denies the very right to have a conversation on the subject.
If you can't agree that I should have some say in what is remembered (or at least archived) about me or generated by me, there's not much we can talk about.
And to make it even clearer, one just needs to think about leaked images. Should we not allow a person to delete such images leaked without their consent?
1,000 years from now archaeologist may have some academic interest and those involved and even generations of their descendants are long dead.
But what about 1 year from now? Benefiting those 1000 years from now as the cost of those alive today is a hard position to justify, especially with such a blanket justification.
Playing devils advocate here for a moment. . .
Considering we as humans learn little from our past, keeping all of this knowledge is a benefit to whom then? Some people who feel nostalgic about Sony's first walkman? Or maybe people using it for nefarious reasons? If humans continue to make the same historical mistakes over and over, what benefit does the human race gain from cataloging all this information? I would venture to guess, its more plausible it will be used against us instead of furthering our own culture.
>> We know more about how Rembrandt painted and lived than we do about how Atari 2600 programmers worked and lived
There is a huge difference between saving all of Rembrandt's stuff than it is some 22 year old college drop out programmer who created a video game in the hey days of long forgotten startup company. And yeah, there have been numerous documentaries, and articles written about Atari in those early days. Who would want to save a dilapidated roller rink under the auspices that a great and noble video game company used it as their HQ for a few years??
https://www.polygon.com/2018/7/6/17542154/atari-book-valley-...
But then this roller rink down the block became available: 10,000 square feet! I mean, we were just jam-packed, and we had people on roller skates actually running around on the roller-skate rink building Pongs.
While I do think leaving certain things to the sands of time is a good thing, vacuuming up everything is just as worrisome. Are we going to be hoarders of a bygone technological past where a large majority of the "stuff" we save will have little, if any use to anybody anymore??
Having a background in anthropology, I find it fascinating there will be many generations of kids who leave no physical trace of their existence since a large majority will be in electronic form. Just imagine how people's lives are in a sort of suspended animation after passing away and having their Facebook pages live on forever.
It's a bit hard to wrap your head around tbh.
- kindling sources, like a LiveJournal post that inspired Lin Manuel to write Hamilton (for a fake example)
- early work of a future star, like imagine Lorde posted early songs to MySpace. This is already a clear issue as many posted songs have been deleted or lost for various reasons.
- valuable things on shaky ground. Yahoo Groups, for the latest example. But I just saw on Reddit someone was looking for a deleted scene from Blair Witch Project that was supposedly the first video ever published on Amazon Prime Video... and now it has nearly vanished. That seems crazy to me from so many directions.
- the value of the ephemeral. Gold and jewelry from old civilizations is nice but we know so much of how people actually lived by examining their garbage, scrap notes, broken bowls, etc.
- the myth of permanence. We feel like 10 million people see a video, it is probably preserved. But there are no master tapes of any of this and so much of everything is interlinked and hard to piece together after the fact. What were people's tech stack when they were making MySpace? How big were people's hard drives? Did rhey share sonngs theough Kazaa or play them on MySpace directly? What was the state of Javascript then, what were the security issues or underground trends? How did songs propogate, where were they shared? Were people sending links in email or AIM, were people sharing links on Digg? This is stuff from like a decade or two and already you need to think like an archaeologist to have any sense of how the culture really existed because there were so many moving parts from year to year.
- the value of datasets. Imagine putting some thought against the Geocities archive to see how HTML blink tags grew then fell in popularity over time. Or how a meme propagated, or analyze the link structure between groups of people or by topic or any make any number of interesting inquiries about how humans operate culturally in digital space and how interact socially through certain set of tools and limitations. There are very interesting possibilities here for understanding ourselves better as a species.
Jeez... You may want to re-read my comment. I have written no such thing, and it is not my opinion at all.
I feel you got the comment backwards: a better analogy would be "if a used-books store full of Dan Browns were to burn down, would we regret the loss of maybe one chapter that has some value?"
Your position seems to be "yes", but I wouldn't dismiss so easily the opposite view: that 90% of everything is crap, and that keeping everything forever "just in case" sounds surprisingly similar to hoarding.
I do not oppose "purposeful archiving" - as someone mentioned, saving outgoing Wikipedia links seems smart. But my old twitter account, where I kept track of missed trains? There are better sources for that, and no one missed it when it was gone.
Imagine an author writes a paperback. It isn't very good, but a few people read it. Later, one of those people goes on to rework some of those ideas into their own script for a film. The film is a success. Years later, the scriptwriter mentions the paperback as an inspiration while giving an interview, but it's long out of print.
To a biographer or a devoted fan of the film, this forgotten book, while of little value in and of itself has become a valuable part of a larger story. If it were culled when the contents of that used book store burned down, we would have lost something without realizing it. And that's how we lose most things. The only way to minimize this is to store as much as we can, in the hopes that we may find a use someday, and thankfully digital storage has made this very, very cheap. The opportunity cost is tiny, and the potential reward, given enough time, is unbounded.
And I do recognize that the thousands of petabytes will likely be chump change to store in a decade... but necessarily the economies of storage will keep pace with the rate of content production. It will always be expensive to store everything.
The question is, do we gain back this investment from future uses of these archives? I dunno. I’d be interested to hear what value archivists have gotten out of the archives, given it is decently old already.
Even if you look at classical art with an honest eye, you can find plenty of works that in themselves are, well, crap - but they're being preserved and reproduced and talked about, because they acquired meaning over time. They've become relevant in context.
Take your old Twitter account. It's probably not interesting. It probably won't ever be. But it might. Imagine several decades from now, your great-granddaughter becomes a well-known, influential politician. This might retroactively and posthumously make you relevant, and in the process your Twitter account. Biographists might find it useful. Or independently, people who're into historical train schedules. Etc.
It's near-impossible to predict what the future will find relevant, so if storing some memories is nearly free on the margin - as it is today, with digital technologies - then just storing it is a no-brainer.
Very good point!!!
In my view the Internet Archive should be the Digital equivalent of the the role of the National Register of Historic Places (NRHP). Shepherds of documentation, to give it a cool-ish sounding name.
>1. Archiving isn't just capturing data or downloading it, it's making it available into the future. Without an intense amount of planning around that, the act of capturing is pointless.
>2. We don't need all of Yahoo Groups. We need a subset of Yahoo Groups. Choosing what, exactly, is worth keeping is called appraisal in the archives world - not like monetary appraisal but cultural appraisal.
>3. We also don't need all of Yahoo Groups because hoarding data long-term is terrible for the environment. Digital preservation is also terrible for the environment. So we should be extremely judicious about what digital content we choose to attempt to retain permanently.
>4. It's an incredible violation of privacy and doesn't align with the ethics of the archives profession to collect all that data without permission from the people involved, especially in the case of private groups. People should also have the right to be forgotten.
>It’s also part of a long-term pattern of IA (and this dude in particular) deciding to “archive” things that people have not given consent for—and, in some cases, have explicitly asked not to be preserved. There’s another gendered element here: IA tends to get tons of accolades and funding, and is largely seen as a group of do-gooder dudes just trying to preserve the internet. Meanwhile they routinely ignore or denigrate the work of librarians and archivists trained in digital preservation—professions that are overwhelmingly gendered as female.
>Plus Jason Scott is generally a dick to anyone who brings up any ethical qualms about their work (and he has a stupid hat in his avatar.)
Not to mention that Archives is historically a female dominated industry. This is a real world example of a loud, boisterous man "disrupting" an industry.
If someone thinks they can do better, they are free to try. No one is gatekeeping their attempt. Help yourself to some storage & VMs and write some code. If you do better (regardless of gender), everyone benefits.
[1] It's not bad to be able to wild west it and cowboy shit in non-regulated industries, where someone's life safety, finances, etc aren't at stake. Two cents.
On the other hand, clearly the Internet Archive is a competent digital archiver, and they've earned the capital-A "Archive". They publish a larger digital commons than anybody else, I think, especially at the low low price of gratis.
It sounds like your entire complaint is in two points. First, that IA doesn't ask (much) permission, which is unsurprising. The history of libraries is not one of asking permission, but of simply doing it. The public has been convinced repeatedly, over the decades, that libraries are good for them, and this public support helps insulate librarians from corporate interests.
Second, that IA doesn't employ enough women. I can't help you with that, but you are free to improve yourself.
Obviously there's value in archiving some information, but a save all or even same most approach starts sounding a little hoarder-ish. Sure you might one day make use of that 1997 November TV guide, but chances are you won't and in the meantime you're paying the opportunity cost of storing it.
Maybe we need to take a page from Marie Kondo and only keep that which sparks joy and learn to let go of the rest. There's a chance someone will need a bit of info that no longer exists, but we'll probably be ok.
Presumably some sort of search engine, not a person.
In that sense, curating all of it doesn't really matter as long as you archived it. Someone trying to find the data later (or curate it!) can find their way to the right URLs using other sources, and then begin the process of curating this archived data after-the-fact.
There is probably a lot of junk data on IA though especially video site archives but its worth keeping stuff that isn't needed if it means keeping stuff that was useful.
But you see, even if the Internet Archive didn't exist, someone would probably still be saving a copy of the things you do. It'd just be a megacorp or surveillance agency instead of a more egalitarian organization.
So the choice isn't "things on the internet are ephemeral" or "things on the internet are available forever to everyone", it's that or "things on the internet are available forever to some subset of the rich and powerful".
In the offline world the National Libraries get a copy of every book, magazine and newspaper published, by law. At least that's the way the UK and US do it. They archive a lot of other stuff as well, including music, audio, adverts, but that's more informal, and there is no requirement to preserve.
Personally I'd like things politicians and personalities (by dint of having chosen to live large) say online archived, all business (to later hold them to account) along with the sites of anyone in the business of influence - think tanks, parties, lobbyists, activists, "grass roots" organisations etc. Individuals, anon forums, HN and reddit subs and other places of shooting the breeze should be allowed to stay ephemeral. In fact I think conversation is freer that way - some will choose to say less, say different, or say nothing if all everyone says is forever...
I consider HN pretty much permanent and tread carefully with controversial opinions or things that might one day be considered not-PC.
In a sense this is also a good demarcation point. If something is serious enough to be worthy of copyright protection, it's probably worth archiving.
I've published several books the LoC does not have.
Also the national libraries aren't the sole archives of culture. Univ and private libs preserve all the important stuff government has not the interest or budget for.
[1] https://www.copyright.gov/help/faq/mandatory_deposit.html
Moving from "somebody knows this happened" or "this is in a file drawer somewhere" to "there's a searchable record of this" expand everyone's access to the info, and can do a lot to stave off forgetfulness and bit rot. But the people whose gain the most access are the ones who weren't involved in the first place, and the intersection of "uninvolved" and "cares enough to check" tends to be people who are actively hostile. Hence doxxing, stolen photos, and callouts over years-old tweets.
But that's a broad result of digitization. If a reporter or opposition researcher wants to embarrass someone, they can already look through digitized student newspaper essays, find interview subjects off class rolls, or simply comb through Twitter for long-forgotten offenses. (This holds for both good and ill - it applies to both serious skeletons and misleading or trivial issues.)
The Internet Archive, then, seems like sousveillance offsetting surveillance. For those who can point time, money, and connections at a target, it's enough that evidence exists, and more than enough that it's available online. But for the general public, it's much harder to keep track of countless sources or publicize news. If you can't dedicate interns and an archive to tracking every news story you read, you can't find or prove edits. (And while most newspapers noted corrections or morning/evening revisions, silently changing online stories has become common practice even for the likes of the BBC.) If you can't point out a webpage or tweet to thousands of people at once, the evidence is likely to be taken down before it's recognized. There are a lot of dedicated sites like NewsDiffs working on this problem, but Internet Archive provides a general-purpose answer to "let an average person see the history of a page or create a trusted record of it".
I worry that this just amounts to an eye for an eye, and still increases the total amount of scrutiny we're all under. But as long as more content is becoming permanent, it still seems better to have symmetrical access to it.
Not to mention that quite a lot what is archived today has been made by companies, there's no "right to be forgotten" that companies could ever deserve. For example I've uncovered quite a few mistakes in currently public datasets/websites based on archived sites, who knows how many mistakes are made now and never fixed because we lose the original sources. Point being that the lack of original source doesn't mean the information gets lost, it just becomes a big version of the kids game "telephone" where everyone recites what they heard and it gets distorted in the end.
The truth is, only the things someone has an interest in archiving will be archived, and only so long as someone has an interest in maintaining those archives. Just look at the recent announcement about Yahoo Groups... no one was, and likely no one is, going to permanently archive most of that. Sites, content and history get lost all the time.
I'm sure glad I went to efforts to scrub my personal sites I made when I was a teenager!
I don't think all blogs and personal content(for the lack of a better word) should just be archived. You should need consent. Most people have no idea it's going on. Or it should be very easy to delete something from the archive.
The real problem here is the runaway cancel culture, where we attack people for things they said or did years or decades ago which were (at the time) perfectly acceptable and reasonable.
The most egregious example I have seen so far is cancel culture advocates who think we should disregard the late Richard Feynman’s legacy because he said some rude things to a lady back in 1946, even though the lady herself was not offended, since she did sleep with him later that same evening.
There’s a point where we just have to say “That was a long time ago, no one at the time was offended, get over it.”
A comment made years ago is only a reflection of a person's opinions at that point in time; opinions which may have changed since.
A website can be archived, vanish from the web, and then vanish from the archive for technical copyright reasons (new owner's robot.txt file on the root). So "archiving the archive" might be useful. Or something.
Ezboard was an old discussion site that contained much of interest - archived and now the archive is not accessible.
https://archive.org/post/389127/ezboard-content-suddenly-not...
I've always wondered why it was called the "The Oakland Archive Policy".
http://groups.ischool.berkeley.edu/archive/aps/removal-polic...
I work for a company that owns a website containing information that in my opinion is valuable to the public. The website may go offline forever in the coming months. How can I get in touch with the Archive, to ensure that the content is saved? Parts of the content are not easy to index (e.g. there are "hidden" pages that you will only find if you have the exact URL), I can assist with that.
[1] https://www.archiveteam.org/index.php?title=Archiveteam:IRC
Please do post your efforts on HN if you succeed in getting the material archived. I'd love to hear more about it.
I would be perfectly ok if I was trying to see a copy of a web page from five years ago, and it said that I had to make a request and it would be available in five or ten minutes.
I think I could wait five or ten minutes for a web page to get pulled from the archives.
A problem with most forms of warm storage is that they involve moving media around, which itself can lead to degradation.
I would like to see some write-one and really last forever form of storage to come into widespread use. That seems ... like it's still a few years off.
And you still have the issue of how such storage is read, and how that affects total lifetime.
In that case if the low-demand data front-end machines get too much load, you'd be entering a queue to be served.
Not sure that's realistic here though, it sounds like the main problem is just the volume of data so whether the storage for the data is hot, warm or cold isn't necessarily going to make a big difference and having enough frontends to serve the content isn't a big problem (in comparison) either
Leaving a lot of dead links on the FDA site. Sometimes they tell you to look in the archives for the old information, without giving you a link to it, and sometimes they don't, they just expect you to know.
Now why can't the FDA afford the space to keep their pages forever on their own site? Fill in your favorite conspiracy theory...
Some of the information that has been removed, such as the 2015 hearings on Fluoroquinolone antibiotics, are important health research as just one example.
This seems like a prime example of Hanlon's razor. Tight government budgets and lowest-bidder contractors not bothering with page permanence strike me as the most likely explanation.
Both are highly worthwhile projects -- some of the best of the Web. But a straight apples-pomellos comparision is difficult.
I think it's a fair enough comparison.
Or are you saying that they should have backed up 10x less data in order to have the 10x more redundant copies? That's a value decision on what to prioritize (more data saved vs 'good enough' resiliency).
Donations would be the best way to solve this problem so that they don't have to make this trade off in the first place.
Basically yes. It's not quite same cost due to the CPU/memory issues I mentioned earlier, but it's the same cost for disk drives.
> Donations would be the best way to solve this problem
Donations don't solve this problem. The vulnerability to faults is part of the system design, independent of size.
The decision here was: For the given budget we have, where do we make the reliability vs data backed-up trade off? Increasing one necessarily decreases the other.
The options are:
- Increase reliability but decrease how much data capacity they have
- Increase data capacity at the cost of lower reliability
- Remove the "given the budget we have" constraint via fundraising. With more funding they can afford to increase both
You can see the decision they made for the fixed budget option. Now we can help them remove their constraint with extra donations (Every $5 helps. I've already donated)
Simply not true. Encoding schemes represent a whole different set of tradeoff possibilities which you don't seem to have considered, and simple replication is worse than other options in terms of both reliability and efficiency.
It just doesn’t seem like that use case (mirroring) would be relevant to the Internet Archive.
https://docs.ceph.com/docs/master/radosgw/
https://docs.ceph.com/docs/mimic/rados/operations/erasure-co...
The trick is you have to create the ROM yourself apparently. Sharing it is bad. I would assume copies made by anybody would be identical, so it would be interesting to see such a case in court against The Internet Archive, not that I want them to get sued, but would love for them to win if they did.
Nintendo is the most aggressive about ROMs than anybody.
They are, but they’re also the most proactive in making their old games available on new hardware. It wouldn’t be fair to call Super Mario Bros “abandonware” because Nintendo has done a lot to keep the game alive and accessible to new generations.
Here's an article that asks this very question, not sure if it was posted on HN prior or not.
"Computer programs and video games distributed in formats that have become obsolete and which require the original media or hardware as a condition of access." can be argued to cover many of the video games they currently archive. You could probably argue this applies to classic nintendo ROM dumps and such as well but I doubt anyone wants to have a 10 year long, multi-million USD court battle with a notoriously abusive legal department like Nintendo's.
How valuable is it?
I don't remember the exact formula, but it was similar Shannon's equation. Basically, the more valuable information is, the larger the change it affects in the probably of what an organisms next behavioral state is. So, if information signaling didn't change the behavior of another organism, it wasn't considered communication.
There's so many retro patches that just don't exist publicly. For example, a number of files on SciTech's IA's WBM have zero captures. Most FTP sites weren't captured in WBM adequately either. There are spots of FTP archives hosted here and there on IA and elsewhere, but they're not like WBM for static content sites, and a single snapshot archive lacks the history and the changes, before and after. It is what it is, unless folks donate their vintage personal/work local mirrors to add to the collective.
Anything that was not securely scanned by WBM but was donated should have a different and clearly denoted status, if admitted.
Interestingly, the British Library uses web crawlers based on the Internet Archive's Heritrix web crawler[2], which demonstrates how important IA's work is for many other archival organisations' work.
[1] https://www.webarchive.org.uk/ [2] https://github.com/internetarchive/heritrix3
I wish Semantic Web had taken off. "Pages with styling" was suboptimal. Web apps are a such a weird evolutionary branch we've descended into that don't relate to documents.
Content instead should have fallen under a type of ontology: news item, blog post, technical reference, comment, status update, ... If we'd adopted such a markup grammar and styled around it, we could parse out meaning, have stronger links in the graph, and compress.
Semantic Web would have happened if commercial web didn't outpace it.
And coming soon, you’ll be able to get paid to make content available using IPFS and FileCoin[2].
[1]: https://ipfs.io/
[2]: https://filecoin.io/
I was thinking more about the 90s era dream of uncensurable "data havens". That led to Freenet, for example. Which is slow, and forgets stuff that doesn't get accessed. And Tor onion services, which are more readily taken down.
But the problem is that there's no way to know in advance whether something is about to disappear from the Internet Archive. So you'd need someone inside who'd discreetly alert the backup service.
But:
Let's also support the public service players in its space that often get forgotten or marginalised by it's well funded marketing.
I'm thinking particularly of the various national libraries that preserve content under incredibly tight budgetary, PR and legal constraints that the IA is relatively free of.
On IA-aware media like HN, there is a tendency to present it as the only preservation initiative out there, which is absolutely not the case.
https://en.m.wikipedia.org/wiki/List_of_Web_archiving_initia...
There are probably a lot of reasons, but I don't know where to chat about it to learn more.
edit: now that I think about it, if the original is pdf, then jpeg makes sense for loading one-page-at-a-time in the browser, but it seems like mp3 transcoding to ogg is reasonable to leave up to the user?
And also, I guess, transcoding the least-popular format on-the-fly is still too CPU intensive for large files (zips of hundreds of jp2 images, etc)
Guess I'm looking for a magic bullet where there might not be one, I just want to see Archive.org keep doing what it's doing far into the future.
Who knows, maybe we will stumble on some magic bullet, new compression algos (Zstandard? AVIF? AV1?), user clouds for compression (like boinc; archive already lets users assist via bittorrent for bandwidth costs). Thinking out loud.
Anyway, keep doin' what you're doing. And podcasting. That's good too :)
Surely this would get enough support that they could host torrents of the content stored in chunks, and have many peers download and seed many chunks making the backups entirely distributed? I’d gladly seed a few hundred gigs of data to ensure they maintain good backup procedures.
Let me remind that there already was a fire that destroyed some equipment for internet archive but luckily not digitalized data: https://www.theverge.com/2013/11/7/5076166/the-internet-arch...
Luckily that article from 2013 mentions: > None of the Internet Archive's digitized data was lost in the fire as backups are held in multiple locations.
Seven versions of the Primos operating system, from rev 18 to rev 24, have been recovered from disks, 9-track tapes, and 8mm backup tapes. The company died around 1992.
There has been a huge amount of Prime software that as far as I know, is lost to time. Oracle ran on it. SPSS. A native DBMS. Every one of the major OS revs had 5-10 minor revs. Some software products are only available for certain revs. I actually used rev 12, so at least half of the OS versions are completely missing. And Prime released source for their products. Most of that is missing.
For me personally, the emulator was / is a very rewarding project, and maybe for a handful of others who are still alive and used Primes in high school, college, or work. It's been really fascinating to "relive" my Prime days now and then, and others have made similar comments.
But how valuable is it really? Is it more than just a curiosity now? Sure, for a current or future computer historian, it might be viewed as a gold mine, but as time goes on, it has less and less value IMO, especially as the people who actual used one die off. If in a few years, only 10 people in the world care about such a thing, does it make sense to save it for all time? I have my doubts. And if I did have all of the versions of Primos, and all of the software ever written for Primes, would it make a difference even now, when there are people alive who actually used this computer system? Also seems doubtful.
To me, an "archive everything and let someone else figure out if it's useful" strategy seems impractical. If I had my choice, sure, I'd like to have every major version of the OS and all of the products for each version. But all of the dot-revs too? Nope - not important. Would I like all of the manuals? Yep!
Some of the products like Oracle, DBMS, etc. were a bitch to configure back in the day, and certainly would be even harder now with very few people around to help. So even if I did have them, I doubt I could get them running. And even if they were running, they were a specialty product at the time, with few experts, so finding someone today that would be interested in them would be a needle / haystack thing.
It all reminds me of family pictures. Before my grandma died, she went through a large box of B/W photos, explaining to us who all the people were, and we wrote on the back of the photos. But just 3 generations after her, the generation after me, no one knows the people in the photos (except my mom is in some - their grandma). In a way it seems that this family history is somehow important, but no one younger than me is interested at all. And truthfully, I never look at the pictures myself. I ended up scanning the photos with my mom or grandma, putting them on a digital picture frame for my nephews, and leaving the rest out. I asked them first - they didn't care about any of the others.
I'm not an archivist, but it seems to me that curation is a much more important aspect than grabbing everything and keeping it alive forever. Not curating seems like a "scale to infinity" problem. Those tend to not end well.
I think the IA is great project and am not intending to be critical, but giving my own perspective about preserving some very small piece of computer history.
Edit: If anyone would care to explain why they feel that chopping an article up into pieces of arbitrary size makes for a good user experience, please do so. I know Twitter gets a pass from HN in general but it is simply not suited to this type of content by design. This was a conscious choice by its creators.
Edit: Twitter is obviously the correct platform for this because of reach. Hacker News is reading it, his 33.3k twitter followers are reading it, reddit is probably reading it, everyone's reading it! So if the goal was to draw more donations, he's winning.
For example, here's the thread author quote-tweeting the beginning of his thread, which allows him to prepend both a tl;dr and a link to "archive.org/donate"
https://twitter.com/textfiles/status/1204431184069058560
If he wrote a blog post that includes the fundraising URL (e.g. at the top and the bottom) and tweeted a link to that blog post, you believe that that call to action would be more visible and obvious to the average reader?
On a deeper level, content follows form. The limits of Twitter inherently force me to cut my wordiness and elaboration to a concise point. Of course whether this makes the overall thread better than a blog post is up to debate, but it most certainly makes composition easier on the creator.
Once in awhile I'll read the ThreaderApp's rollup of a thread, and almost always find it less of an optimal experience than just scrolling through the thread on mobile, though that may be attributable to differences in visual design. I can think of a number of Twitter threads that I simply cannot imagine working well in a blog or article format.
For example, here's a Twitter thread that was posted by a newspaper reporter, as she followed a mailman on his last day after 35 years:
https://twitter.com/Jennifer__Brett/status/11312980440104837...
(HN Thread is here: [0])
AFAIK, she never posted this as an article on her newspaper's site – which to me, as a former newspaper reporter, is very surprising given publishers' demand to feed the content/pageview beast. I don't know if she ever explained this decision, or if the thread eventually did make its way into native article format. But I think the thread format worked beautifully here. By necessity, it minimized whatever tendency the reporter had to insert superfluous prose. And the Twitter visual design places heavy emphasis on the photos and videos, which each exemplified the aphorism of "a picture is worth a 1,000 words"
edit: the WaPo did do a recap of her thread [1]. Being a recap, of course, means it's inherently different than an original article about the mailman, but you still get a taste of the extra verbiage needed to connect scenes in traditional prose, which is not needed (maybe because people don't necessarily expect it) in a typical Twitter thread.
And again, the visual design aspect is critical imho. Twitter's format integrates images and video much more seamlessly than a standard website design, partly because tweets have a strict container max-width of ~500px, which is far narrower than most modern sites. And I suspect knowing that you're recording media for a Twitter thread fundamentally affects how you do it – e.g. you may feel less restrained about context and quality.
For example, one of the best tweets in the mailman thread imho is this one:
https://twitter.com/Jennifer__Brett/status/11319201271575920...
The text is very informal: It took forever to #MrFloyd to even get to his own party - people kept stopping him for photos and hugs!
But the attached video tells the story better and more immediately than any prose could. I can't imagine that 49-second raw footage clip being inserted into a standard story without being a massive distraction from the reading.
[0] https://news.ycombinator.com/item?id=20001559
[1] https://www.washingtonpost.com/lifestyle/2019/05/24/mail-car...
I tried to read it. I have no idea what the "silent killer" is. And I have no idea what a lot of the images had to do with anything.
Is there any useful purpose to all this other than nostalgia? And at what point does the nostalgia overwhelm the useful purpose?
Even if most data is uninteresting doesn't mean that there are data there that is very useful to someone.
Information destruction is not necessarily a bad thing.
There's a problem with the Wayback Machine in specific which can kill your ability to access it quite silently, unless you know how to use the browser's development tools and interpret headers.
It has to do with cookies: Somehow, the Wayback Machine sets cookies... and sets cookies... and keeps setting cookies, until it overflows its own ability to accept cookies. At that point, your browser tries to access a Wayback Machine page, handing the server all of the cookies it currently has, and the server refuses to deal. It absolutely denies everything, sending an error header and a blank page. You have to clear all web.archive.org cookies to get anything at all, at which point it works perfectly.
I've completely solved this problem by blacklisting web.archive.org in browser cookie blacklists. I haven't had it happen since then. As far as I'm concerned, the problem is diagnosed and just needs to be solved. At their end.