It feels weird to me to see Vint Cerf propped up in this title by being associated with Google. It seems the guy's name holds enough merit on it's own; I didn't even realize he was working with Google these days!
It is Hacker News, not Computer Engineer/Scientist news. It was always supposed to tilt towards Silicon Valley. I don't think that's a good thing, but it's a free service from Y Combinator, so it's very understandable.
Well, since it was pg's idea, I'd guess "hacker" would slant more towards MIT-style hacks than anything to do with the valley.
I think the lack of Vint Cerf stories has more to do with the temporal nature of news than any biases that people might have. I'd chalk it up to him just not making news lately. Any text on the history of the Internet, TCP/IP, and networks in general liberally mentions him.
It's interesting to examine which historical figures have been remembered and which have been forgotten. Money is part of it but not the full story. Political leaders during times of chaos are remembered. Artists are remembered, many of which died penniless. Mass murderers are remembered. I think in general people just need a simple story to attach to a single person to remember them. Building a computer and making a billion dollars is a simple story. Walking on the moon is a simple story. Designing packet switching protocols is not. It's not a tragic situation it's just interesting to note.
Interesting... maybe an age thing? I'm 39 and Vint is right up there in my mind with the heavyweights of computer science.
That said, I wasn't aware he was with Google either. I imagine his place in my mind was solidified back in college during the .com days when history internet was a common topic of discussion.
He was supposed to give a talk here at Goddard Space Flight Center recently, but it was snowed out. I've heard him speak before, on his Interplanetary Internet work, and he's a good speaker.
I feel like we are already experiencing that. I have a drawer full of pictures given to me by my parents, some of them 100 years old,but I can't open pictures I took with my digital camera 10 years ago,because the CDs I burnt them to are unreadable. Obviously the answer to that is that I should have printed at least some of them,but I don't know anyone who prints pictures nowadays. Nowadays I backup all of them to Picassa(google plus albums),but I have no idea what happens if I can't pay for monthly storage anymore? Or if I die and no one knows my google password? All of that data will be gone permanently.
>> "I don't know anyone who prints pictures nowadays"
Really? Most of the big supermarkets here (UK) and quite a lot of pharmacy chains have machines where you connect your device/USB/SD, select the pictures you want, and press print. It's relatively inexpensive too and I know people who use them all the time. Personally I don't actually mind if all of my digital photos are unreadable so long as I have the important ones in a physical format so this is a nice solution and takes away the pain of owning a good printer, buying lots of ink and photo paper.
Oh maybe I didn't phrase it correctly. I meant that I don't know anyone who prints out their own pictures - most of my friends only keep their pictures on their phones/digital cameras, physical copies are only ever printed as presents and such, never to archive anything.
I know it doesn't answer the problem at large, but if you are worried about this (or even if you're not and you just want your google data to survive you), Google released an inactive account manager a couple of months ago:
In the event your google account becomes inactive for a long period of time you choose, you can trust specific contacts with access to specific data. It has other niceties. I highly recommend setting it up.
>but I can't open pictures I took with my digital camera 10 years ago,because the CDs I burnt them to are unreadable.
We still had jpegs 10 years ago. All of my digital pics from 2005/6/7 till now are still on multiple hard drives and multiple systems. Each time I get a new computer I transfer them over.
Why are they unreadable? It sounds like you put them on a proprietary format or some obsolete picture software you used to burn them on.
I think its absurd that you are paying for picture storage and think its your only option. Amazon prime customers can upload pics for free now [1]. Why not use one of the free ones like Dropbox, Google Drive, SkyDrive etc...
CD-R's, and especially older CD-R's, turned out to not have as long a shelf life as initially assumed. I recently found a bunch of old CD's I burnt ages ago, back when I got my first CD burner, and only one of them was still completely readable.
They're unreadable because CDRs aren't an archive mechanism. They break down rather rapidly. Their shelf life is similar to a floppy disk. After 10 years, the odds of being able to read a CDR that wasn't stored in perfect conditions are pretty low.
Are you sure it would be as quick as 10 years? I have music CDs and mixes easily from the early 2000s that still work. A few years ago I found a Jurassic Park soundtrack that still worked.
My parents have CDs from the 90s that still work.
Would there be a difference in a music CD you bought from a store and one that burned yourself?
>>Amazon prime customers can upload pics for free now [1]. Why not use one of the free ones like Dropbox, Google Drive, SkyDrive etc...
All of them have space limitations - and I have over 400GB of pictures backed up to Picassa. So if I am not paying google, I would be paying Amazon or Microsoft or Dropbox or someone else. And my main point was - what happens to data on all of these paid-for services once I am dead? Amazon Prime backup is "free" as long as you keep paying for prime.
And yeah, CDs become unreadable after sitting in their envelopes for 10 years. Not all of them,but mine certainly did.
"The project was stored on adapted laserdiscs in the LaserVision Read Only Memory (LV-ROM) format, which contained not only analogue video and still pictures, but also digital data, with 300 MB of storage space on each side of the disc. Data and images were selected and collated by the BBC Domesday project based in Bilton House in West Ealing. Pre-mastering of data was carried out on a VAX-11/750 mini-computer, assisted by a network of BBC micros. The discs were mastered, produced, and tested by the Philips Laservision factory in Blackburn, England. Viewing the discs required an Acorn BBC Master expanded with a SCSI controller and an additional coprocessor controlled a Philips VP415 "Domesday Player", a specially produced laserdisc player. The user interface consisted of the BBC Master's keyboard and a trackball (known at the time as a trackerball). The software for the project was written in BCPL (a precursor to C), to make cross platform porting easier, although BCPL never attained the popularity that its early promise suggested it might."
And I believe the content on the original Laserdiscs has been reverse engineered more than once. Hopefully once content is on the web (unless it's behind robots.txt) it gets sucked up by the Internet Archive.
That of course doesn't take care of any rendering issues.
Last summer I discovered, by accident, a little computer museum[1] in the Kvarner Gulf. It's run by a single guy, who collected an incredible amount of old hardware over the years and exhibited them in couple of rooms.
It rained the day I visited the place, and since it was an old building, it actually rained a bit inside. The guy took it with humour, but it was actually a pity. It is a great place and on some days the visitors can even run old programs on those ancient machines or play games.
From their site:
Opened on 22nd of September 2007, Club PEEK&POKE is one of the few permanent displays of vintage computing technology in Europe. Located in the centre of the city of Rijeka and spread across 300m2 of space, it contains more than 1000 exhibits of the world and local computer history, ranging from very early calculators and game consoles to rare and obsolete computers from the nineties.
I've been to a couple of these museums as well. What makes them fun to go to are:
- you may touch everything, often including things inside the computers itself
- the people working there (or, as is often the case, the one guy working there) is knowledgeable, willing to talk about it in depth, cares, and you can have a real conversation with them.
- there's a lot of surprising hardware around I had no idea existed (from failed computer companies, to specialized I/O of decades past, to special-purpose hardware)
You cannot get the same full experience with most regular museums, as then you have to follow their narrative instead of following your curiosity.
On the other hand, these museums are often not much more than an erudite collection of historical artifacts they got their hands on. To preserve is to select with a plan, not just collect everything.
Technical/science museums in general often seem to be closer to this ideal than other museums, because while some of the artefacts may be valuable, more often they are trying to showcase how things worked rather than specific objects.
True. Although the more popular/populous museums tend to offer models you can interact with, whereas the smaller ones more often offer the real thing.
If you're ever offered to visit the archives/warehouse of a museum, accept. Showing (large) stuff is expensive in space, so most exhibitions are only the tip of the ice berg of the enormous amount of stuff the museum has collected and stored. Often an exhibition is a mix of well-known artifacts that visitors expect to be there (and are therefore included, but are not very interesting as you already know them), some artifacts that are good specimens over-all to show (but not necessarily that interesting), and some truly interesting pieces (you probably never have seen before). I found that when you visit the archive/warehouse of a museum, curators tend to show artifacts from that last category in particular.
The Deutsches Museum in Munich is the world's largest and IMHO best museum of science and technology. Especially the mechanical engineering, cars and airplane departments are one of its kind.
Though, the computer history department is a bit too small and similar in size as the equivalent Science Museum in London. In both you find a Cray super computer, first Zuse computers and many older mainframes and terminals, etc. But everything is death, every historical computer sits just there. There is no interactivity, what a shame. They should at least re-work a Cray 1 or 2 super computer and let visitors play around on its terminal - that would be awesome.
Amazing to hear that there is a computer museum in Rijeka. My family spent nearly three months staying right outside of Opatija while working in Croatia. We'd travel to Rijeka regularly when we needed to do any larger shopping.
My wife and I originally planned to organize a workshop or attend some meetups while there, but the tech scene is so tiny it just wasn't feasible.
We'll be going again this spring, this will be on our list of places to visit.
The only long term answer is to detach the data stores from the applications which process them. It is the migration between apps that leads to this data orphaning.
This also leads to the conclusion that a standard mechanism for interfacing to the data from multiple apps on multiple data backends is needed. The Android storage framework is probably the best effort at this so far, but it's far from clear how used it is.
Google won't, that's the fact. That completely useless snapshot will live till the end of the digital age, whenever it will be. That's why I don't understand Cerf's concerns.
The whole point he's making that we do not know what might be valuable to future civilizations. We simply cannot foresee the way they would study our culture, in the same way the ancient civilizations would not be able to comprehend our current civilization.
We don't preserve muddy door matts today "for posterity." That we can't know exactly what will be important in the future doesn't mean we can't reasonably predict what won't be.
I feel like http://archive.org/ deserves a mention here for their efforts of preserving digital history. Personal photos or other data are not in their scope, but they do wonderful things with web pages and other stuff that's publicly available. I just read Brewster Kahle's interview from Founders at Work this week and it blew my mind how determined to build the Internet Archive he's been for practically his whole career.
Strange, I would call these preoccupations BS if not for the source. This was a common concern some 20 to 30 years ago, when the internet was practically non existent and people still used old floppy disks and cds. Today, a good part of our personal files are stored in the cloud and storage devices have become more standard through the usage of universal interfaces. Some of our personal data will be lost (whatever is left in the hard drives of old computers and not backed up) but the trail of information the world is leaving behind is so huge that what worries me is rather that time seems to have frozen. Everything we leave behind us, documents, pictures, looks as fresh today as it was the day it was produced. And as easily retrievable.
On another thread people were saddened by rapidshare-like website being put down because a lot of content was hosted there. Let's be sure the cloud will stay long. So far most 80s data I see on the web comes from .. magazines. Ha paper.
The trick with digital data (including that stored in cloud services) is to keep it in motion. Movement from one service to another, leaving behind a copy that may eventually get deleted but just might stick around a while, is the best way to keep data available. Whatever disappeared with rapidshare would have been just fine if people have moved it to the various "latest and greatest" systems when they popped up, the only way it gets lost is if it stops moving.
What happens when one dies, who keeps moving that data? Companies eventually close. Millennia old documents made their way to us without anybody carying about them. Digital data seem to be intrinsically different from analog ones in that continuous care is needed. What you propose is a solution but I can't see who's going to do it for free as walls and caves did for paintings and inscriptions.
Not necessarily. Analog data decays too. The issue is to be blind to the fact that neither `technology` is timeless and perfect. People are giving in too easy and too deep in digital data nowadays.
Sure but at what rate? Most 20 year old CDs are still fine, to say nothing of paper stored with a sliver of care, what percentage of web companies from 1995 are still around?
Consider this fact, for perhaps a hundred years prior to digital technology consuming just about everything we collectively produced gigatons of paper output. How much of that is available today? A thousandth of a percent perhaps? Almost all of it has been lost and most of what does remain is available because it was related to someone famous or some famous event, or because it was widely replicated. Digital data is easy to replicate and spread widely. Given how cheap it is to store and how it gets cheaper every day it is not inconceivable that every bit that is generated and "donated to the public" will survive forever.
Not so fast. If you abandon your house anything in it will be trashed in 20 years. Humidity, fungus, animals, whatever. Data you mention has been wiped or the medium has been destroyed. I have some old Seagate 6GB IDE hard drives that are still working properly. So maybe web companies data could have been around if their hdd had been stored in a nice location.
It's not BS and you've laid out why. There are games and software on 5.25 and 3.5 disks that we can't access now. Even if we could, the software was written for a world that doesn't exist anymore and we've have to write other software to adapt them to our machines now. This isn't a problem that is just going to go away, formats and OSes and hardware will come and go as our technology needs and abilities change.
First, you're talking about a very short timespan right at the beginning of the digital age. Even if we were to lose everything, a missing 20 years timespan (1975-1995) of videogames wouldn't amount to a digital middle age.
I don't know about digital dark ages, but before GOG this was a serious problem. It's not that games were not accessible at all, it's that it was so hard to access them that most people would not bother. Are you really "preserving" history if it's just some CD in a box no one ever opens?
You are. Most of our history has been preserved in books that were inaccessible to almost everybody and that nobody opened for hundreds of years. The important thing is that things are preserved somewhere, even if to read them you need a fully equipped laboratory and boffins in white gowns. That's how we read our most ancient documents anyway. Once the information has been retrieved and copied, it's safe.
I think you are confusing history with archeology. They are obviously related, but they are not the same thing. Preserving history requires keeping it alive. Those books you speak about were endlessly copied and read. I am willing to bet that most of you know about history you have learned from articles or books of people who read other books, and so on. Archeology steps in when history ends up as a CD-in-a-box or tablet-in-a-tomb. And even then we often loose context and meaning.
No one ever considers the volume of photos and documents we are now creating. My parents probably took dozens of photographs when I was younger, and kept maybe one or two albums. Nowadays, most people have hundreds if not thousands of photographs and videos. So it seems to me that the converse will actually be true. If you've got a 100 fold increase in the number of documents and photos being created, and only 1% of those make it, you're still preserving the same number of documents. And frankly, I think that's a conservative number.
As some of the other posters here mentioned, with the advent of cloud based services and easy backup systems, retrieval is getting easier. So if storage and retrieval is solved, that leaves format evolution problems. But somehow I doubt in 20 years or even more that JPEG is somehow going to be harder to read than it is now.
A couple of years ago, Kodak's online photo storage service (in the UK at least) shut down. We just barely managed to get copies of the hundreds of pictures we had stored there in time.
Other recent (non-photo specific) examples of vast amount of data disappearing:
Geocities. Yes, large chunks of it was archived last minute. MegaUpload.
And we have Rapidshare on it's way to disappearing.
Cloud services can and will shut down, and it is not at all a given that we manage to preserve the data.
This creates a relentless churn where some proportion of older data disappears every day. All we really can do is to fight to keep the churn rate low enough, because we have no realistic prospect of saving everything all the time.
But that's not really any different than any other medium. If there exists two copies of a book in different libraries and then one of the libraries burns down it's not any different. I mean we still talk about the library of Alexandria 2000 years later.
I don't think this invalidates the point though. In those alone we may have already lost more pictures than were ever taken before 1990 but at the same time the amount we have left is staggering. If you go back 100 years, there are famous people that we have maybe 1 picture of. No matter how sloppy we are with the majority of internet content today, I can't imagine that 100 years from now, they'll only be able to find a couple pictures of Obama and Putin.
That reminds me of the Encyclopedia Galactica http://en.wikipedia.org/wiki/Encyclopedia_Galactica. I wonder if this is a scheme to get comp scis to a remote Location and eventually start a new galactic empire.
I, for one, feel that humanity would benefit from this generation's digital presence being wiped from existence. Everyone on this planet is now dumber for having experienced it. I fully support and will tell my congressman to vote for a digital dark age.
Camlistore seems relevant - it's whole intention is to last (at least) 100 years, primarily by making the data format simple and making data migration (ie, moving between providers, between hard disks, etc) a common thing. https://camlistore.org/
How is this different than any other point in history? Throughout time, the vast majority of records and artifacts were lost, and only some given percentage survive. Let's say it's ~2% for argument's sake...actually, the amount of material that survives would decay over time - 40% after 10 years, 20% after 50 years, etc.
I have no idea what the rate of decay for digital records, information, and archives would be, but I would think it would be higher than information stored as hard copies of paper, books, etc. Of course, we are also producing orders of magnitudes more information than we were in the past.
You can virtualize a PC (or replicate the filesystem like Dropbox), but it is harder to virtualize a distributed system like Facebook's or Google's. And most mobile platforms are DoA without their vendor signing and cloud services.
It's always been hard to migrate out of social systems--convincing a well-connected user of a photo service to move away from the place where they've accumulated comments and tags is really hard. That metadata is not portable because identity is not yet portable, and it's what we're spending time on (lots more time than we spend making spreadsheets).
I think we might manage to keep "JPG as file" alive for 30-50 years, but there is lots more to manage.
We are getting better at archiving files. And thanks to a combination of people dedicated to emulation and the rise of virtual machines, there are few popular pieces of hardware we can't emulate in excruciating detail.
But many of the services I used a decade ago are already gone. And pretty much all the services I used two decades ago have completely disappeared, with some very few notable examples. And with them, vast amounts of data.
Some of it the Internet Archive have at least captured static snapshots of (and they really should have magnitudes more funding), but ten times that - or more - was data in walled gardens, behind logins or otherwise restricted in ways that means it is lost forever unless we're lucky and it turns out some admin held onto backup tapes they weren't really meant to keep.
And the problem with there is not to create a snapshot of a single server, but as you say that distributed systems are far harder. Even recent. Twice I've been contracted to help companies take over infrastructure that involved systems I'd worked on, and try to "package it up", and it was incredibly hard, because no matter how much you try to tear down and bring up individual servers or groups of servers and automate deployment, very few places running complex services ever try - or could afford to try - to tear down and bring up a full copy of their entire infrastructure.
Suddenly all kinds of nasty interdependencies and bootstrap problems nobody had needed to think about shows up.
DRM also plays a role in this. Obsolete DRM techniques whose algorithms have long been forgotten will make it difficult to archive data protected by DRM onto new storage mediums.
It sort of is a big issue if you consider that preserving the history of something relies on lawbreaking motivated by piracy to enable it. Outside of the against-the-law bit it requires sufficiently motivated attackers and insufficiently good DRM.
The status quo is probably stable for games and big name software packages for the time being, but it doesn't inspire confidence for the long term.
"Vint Cerf is promoting an idea to preserve every piece of software and hardware so that it never becomes obsolete - just like what happens in a museum - but in digital form, in servers in the cloud."
The point is that the description/emulation would need to be a virtual machine, so it can be transported and duplicated, as long as you have the basic system to run it on. Otherwise the problem is just as bad when those "regular servers" die.
I see a lot of comments amounting to saying that contemporary digital content isn't worth much, and its loss wouldn't be a big hardship. This isn't the only kind of digital content though. What I am worried about is older, pre-digital matter which is thrown out because digital copies now exist (often crappy scans of marked-up, faded pages, but that's another rant.) When the digital copies disappear, we will no longer have the durable pre-digital copies to revert to.
Indeed. Every time we scan a document and throw the original away, we're betting for the continued existence of current technological civilization. If it collapses, the thing we digitized and thrown away is the thing future generations will have lost from the cultural heritage. The Dark Ages may end up extending far beyond the beginning of information age.
Meta: I know this is the Internet, and we're all supposed to complain all of the time. And I am aware of HN's submission guidelines. But do we really need to refer to him as Google's Vint Cerf? The guy is famous in our community. It's not like he needs another adjective, and it's not like Google owns him. How about just his name? I find this phraseology a bit disconcerting.
The irony is that Google have also participated in acquishutdowns and data deletion. They still have the Deja archive online, but I think it's reasonable to worry whether they might just bin it at short notice like Google Reader or Wave.
We observed this problem first hand when the inventor of Powerpoint emailed us to ask for help in opening old presentations which he could no longer view !
When the guy who wrote the software which produced the file can't view it 20 years later you know you've got a problem.
This is why I'm a firm believer in open, free standards. Of course they don't necessarily guarantee you'll be able to open those files, but it's a great step in the right direction.
We shouldn't have to rely on reverse engineering to preserve what's ours.
Unfortunately public administrations don't realize (or don't want to realize) this, at least where I live. I even discussed this with information professionals and they wouldn't understand! We're right in the middle of some big change and it will be too late when they finally realize.
"A company would have to provide the service, and I suggested to Mr Cerf that few companies have lasted for hundreds of years. So how could we guarantee that both our personal memories and all human history would be safeguarded in the long run?"
153 comments
[ 3.2 ms ] story [ 246 ms ] threadI think the lack of Vint Cerf stories has more to do with the temporal nature of news than any biases that people might have. I'd chalk it up to him just not making news lately. Any text on the history of the Internet, TCP/IP, and networks in general liberally mentions him.
That said, I wasn't aware he was with Google either. I imagine his place in my mind was solidified back in college during the .com days when history internet was a common topic of discussion.
Really? Most of the big supermarkets here (UK) and quite a lot of pharmacy chains have machines where you connect your device/USB/SD, select the pictures you want, and press print. It's relatively inexpensive too and I know people who use them all the time. Personally I don't actually mind if all of my digital photos are unreadable so long as I have the important ones in a physical format so this is a nice solution and takes away the pain of owning a good printer, buying lots of ink and photo paper.
https://www.google.com/settings/u/0/account/inactive
In the event your google account becomes inactive for a long period of time you choose, you can trust specific contacts with access to specific data. It has other niceties. I highly recommend setting it up.
The question is what will happen to Picasa? Will it be free in 5, 10, 50 years? Will it be free?
https://support.google.com/picasa/answer/6558?hl=en
We still had jpegs 10 years ago. All of my digital pics from 2005/6/7 till now are still on multiple hard drives and multiple systems. Each time I get a new computer I transfer them over.
Why are they unreadable? It sounds like you put them on a proprietary format or some obsolete picture software you used to burn them on.
I think its absurd that you are paying for picture storage and think its your only option. Amazon prime customers can upload pics for free now [1]. Why not use one of the free ones like Dropbox, Google Drive, SkyDrive etc...
[1]https://www.amazon.com/clouddrive/primephotos
CD-R's, and especially older CD-R's, turned out to not have as long a shelf life as initially assumed. I recently found a bunch of old CD's I burnt ages ago, back when I got my first CD burner, and only one of them was still completely readable.
My parents have CDs from the 90s that still work.
Would there be a difference in a music CD you bought from a store and one that burned yourself?
There are also special archival disks and burners, like M-Disc, which literally "engraves in stone": http://www.mdisc.com/
All of them have space limitations - and I have over 400GB of pictures backed up to Picassa. So if I am not paying google, I would be paying Amazon or Microsoft or Dropbox or someone else. And my main point was - what happens to data on all of these paid-for services once I am dead? Amazon Prime backup is "free" as long as you keep paying for prime.
And yeah, CDs become unreadable after sitting in their envelopes for 10 years. Not all of them,but mine certainly did.
"The project was stored on adapted laserdiscs in the LaserVision Read Only Memory (LV-ROM) format, which contained not only analogue video and still pictures, but also digital data, with 300 MB of storage space on each side of the disc. Data and images were selected and collated by the BBC Domesday project based in Bilton House in West Ealing. Pre-mastering of data was carried out on a VAX-11/750 mini-computer, assisted by a network of BBC micros. The discs were mastered, produced, and tested by the Philips Laservision factory in Blackburn, England. Viewing the discs required an Acorn BBC Master expanded with a SCSI controller and an additional coprocessor controlled a Philips VP415 "Domesday Player", a specially produced laserdisc player. The user interface consisted of the BBC Master's keyboard and a trackball (known at the time as a trackerball). The software for the project was written in BCPL (a precursor to C), to make cross platform porting easier, although BCPL never attained the popularity that its early promise suggested it might."
http://www.bbc.co.uk/history/domesday
And I believe the content on the original Laserdiscs has been reverse engineered more than once. Hopefully once content is on the web (unless it's behind robots.txt) it gets sucked up by the Internet Archive.
That of course doesn't take care of any rendering issues.
It rained the day I visited the place, and since it was an old building, it actually rained a bit inside. The guy took it with humour, but it was actually a pity. It is a great place and on some days the visitors can even run old programs on those ancient machines or play games.
[1] http://www.peekpoke.hr
From their site: Opened on 22nd of September 2007, Club PEEK&POKE is one of the few permanent displays of vintage computing technology in Europe. Located in the centre of the city of Rijeka and spread across 300m2 of space, it contains more than 1000 exhibits of the world and local computer history, ranging from very early calculators and game consoles to rare and obsolete computers from the nineties.
- you may touch everything, often including things inside the computers itself
- the people working there (or, as is often the case, the one guy working there) is knowledgeable, willing to talk about it in depth, cares, and you can have a real conversation with them.
- there's a lot of surprising hardware around I had no idea existed (from failed computer companies, to specialized I/O of decades past, to special-purpose hardware)
You cannot get the same full experience with most regular museums, as then you have to follow their narrative instead of following your curiosity.
On the other hand, these museums are often not much more than an erudite collection of historical artifacts they got their hands on. To preserve is to select with a plan, not just collect everything.
If you're ever offered to visit the archives/warehouse of a museum, accept. Showing (large) stuff is expensive in space, so most exhibitions are only the tip of the ice berg of the enormous amount of stuff the museum has collected and stored. Often an exhibition is a mix of well-known artifacts that visitors expect to be there (and are therefore included, but are not very interesting as you already know them), some artifacts that are good specimens over-all to show (but not necessarily that interesting), and some truly interesting pieces (you probably never have seen before). I found that when you visit the archive/warehouse of a museum, curators tend to show artifacts from that last category in particular.
http://en.wikipedia.org/wiki/Deutsches_Museum
Though, the computer history department is a bit too small and similar in size as the equivalent Science Museum in London. In both you find a Cray super computer, first Zuse computers and many older mainframes and terminals, etc. But everything is death, every historical computer sits just there. There is no interactivity, what a shame. They should at least re-work a Cray 1 or 2 super computer and let visitors play around on its terminal - that would be awesome.
My wife and I originally planned to organize a workshop or attend some meetups while there, but the tech scene is so tiny it just wasn't feasible.
We'll be going again this spring, this will be on our list of places to visit.
This also leads to the conclusion that a standard mechanism for interfacing to the data from multiple apps on multiple data backends is needed. The Android storage framework is probably the best effort at this so far, but it's far from clear how used it is.
https://remotestorage.io
Sure but at what rate? Most 20 year old CDs are still fine, to say nothing of paper stored with a sliver of care, what percentage of web companies from 1995 are still around?
Second, https://archive.org/details/messmame
As some of the other posters here mentioned, with the advent of cloud based services and easy backup systems, retrieval is getting easier. So if storage and retrieval is solved, that leaves format evolution problems. But somehow I doubt in 20 years or even more that JPEG is somehow going to be harder to read than it is now.
Other recent (non-photo specific) examples of vast amount of data disappearing:
Geocities. Yes, large chunks of it was archived last minute. MegaUpload.
And we have Rapidshare on it's way to disappearing.
Cloud services can and will shut down, and it is not at all a given that we manage to preserve the data.
This creates a relentless churn where some proportion of older data disappears every day. All we really can do is to fight to keep the churn rate low enough, because we have no realistic prospect of saving everything all the time.
Now get off my lawn!
I have no idea what the rate of decay for digital records, information, and archives would be, but I would think it would be higher than information stored as hard copies of paper, books, etc. Of course, we are also producing orders of magnitudes more information than we were in the past.
I would argue that for the first time in history, we're not limited by storage space.
It's always been hard to migrate out of social systems--convincing a well-connected user of a photo service to move away from the place where they've accumulated comments and tags is really hard. That metadata is not portable because identity is not yet portable, and it's what we're spending time on (lots more time than we spend making spreadsheets).
I think we might manage to keep "JPG as file" alive for 30-50 years, but there is lots more to manage.
We are getting better at archiving files. And thanks to a combination of people dedicated to emulation and the rise of virtual machines, there are few popular pieces of hardware we can't emulate in excruciating detail.
But many of the services I used a decade ago are already gone. And pretty much all the services I used two decades ago have completely disappeared, with some very few notable examples. And with them, vast amounts of data.
Some of it the Internet Archive have at least captured static snapshots of (and they really should have magnitudes more funding), but ten times that - or more - was data in walled gardens, behind logins or otherwise restricted in ways that means it is lost forever unless we're lucky and it turns out some admin held onto backup tapes they weren't really meant to keep.
And the problem with there is not to create a snapshot of a single server, but as you say that distributed systems are far harder. Even recent. Twice I've been contracted to help companies take over infrastructure that involved systems I'd worked on, and try to "package it up", and it was incredibly hard, because no matter how much you try to tear down and bring up individual servers or groups of servers and automate deployment, very few places running complex services ever try - or could afford to try - to tear down and bring up a full copy of their entire infrastructure.
Suddenly all kinds of nasty interdependencies and bootstrap problems nobody had needed to think about shows up.
The status quo is probably stable for games and big name software packages for the time being, but it doesn't inspire confidence for the long term.
Why not regular servers?
By essentially saying that, if you don't use cloud technology your precious photos and documents might be unaccessible in years to come.
http://www.everpix.com and the like. Hopefully less often than hard drive failures, but more frequently than a format like JPEG becoming unreadable.
I really hope someone comes up with a good long-term data archive/retrieval system soon, but I'm not holding my breath.
"Never trust a corporation to do a library's job" https://medium.com/message/never-trust-a-corporation-to-do-a...
An example of acquishutdown from HN: https://news.ycombinator.com/item?id=8472047
When the guy who wrote the software which produced the file can't view it 20 years later you know you've got a problem.
You can read more about that here:
http://blog.zamzar.com/2012/04/17/open-old-powerpoint-presen...
We shouldn't have to rely on reverse engineering to preserve what's ours.
Unfortunately public administrations don't realize (or don't want to realize) this, at least where I live. I even discussed this with information professionals and they wouldn't understand! We're right in the middle of some big change and it will be too late when they finally realize.
This puts another perspective on the story about Japan's oldest companies, does it not? (https://news.ycombinator.com/item?id=9041040)