It's not just about third-party storage - I have a picture of my great-grandmother and her parents from 1898 when she was 3 in my photo album. It survived several wars, regimes, and revolutions. No current digital tech in general can provide this kind of longevity, you have to actively watch the media yourself and transfer it when it's about to expire. (which I also do for the rest, of course)
Your house burns down, now what?
> I have a picture of my great-grandmother and her parents from 1898 when she was 3 in my photo album. It survived several wars, regimes, and revolutions. No current digital tech in general can provide this kind of longevity
That's because digital storage didn't exist then... Now it's easier and cheaper to store media than ever before
Maybe I'm missing something in your argument, but to me it sounds naive?
Or maybe you're just in a nostalgic mood!
It may seem that way, but it's not as simple. A digitized history project in the UK made this assumption, and later worked really hard to get back at the data, [0] is a long description, the Wikipedia entry on the BBC Domesday book doesn't go into all the details. You might believe we've learned a lesson since then, but I don't.
DRM formats are creeping into everything. Right now it seems JPEGs will be usable forever, and that may be; But cultural artificats - games and videos - are already becoming hard to preserve; People only care about convenience, and I suspect a popular 3D-or-lightfield-or-something-format in the future will be similarly hard to preserve 20 years later.
Print media has two main advantages: It requires no equipment/software/license/electricity to use; and treated properly, it has been shown to last hundreds of years.
I have a couple of 20-year-old hard drives that are still mostly readable (12 bad sectors so far), and that's considered a miracle. Usenet history is disappearing, even though it doesn't take a lot of space - because it isn't in anyone's economic interest to keep it - and the same will happen to free cloud storage sooner or later.
> Your house burns down, now what?
He still has the digital copy. And perhaps physical copies that were sent to relatives.
If one has both digital and physical, then either can disappear and be recreated from the other.
Personally, if Lloyd's of London offered an insured storage service, I might believe them, but I can't think of any other organization I would trust with a 100 year promise.
Nothing stops you from having redundancy in printed media. In fact, I have that photo reproduced in another album of my relatives.
>That's because digital storage didn't exist then...
What can you realistically do to keep the digital photo through 120 years or more? Imagine you stored it on the most durable media you have, M-Disc perhaps, or whatever they advertise today. Let's assume it even survived 120 years with its data intact. Are you sure your descendants will have the means to read that M-Disc in the year 2140? How about the file system on it? JPEG or DNG format? With a printed photo, you don't have to be sure.
Online storage is much more complex and thus inherently volatile. You depend on other people and on the assumption that business and legal frameworks will remain the same throughout the archival period.
Digital tech heavily depends on an unbroken chain of content transfer, which is a pretty big assumption for archival-grade storage, considering that companies, encodings and formats come and go, and there are also things like censorship and AI "enhancement" to worry about. (and who knows what else it will be many years into future) Digital media is not static by nature. Admittedly, any media depends on content refreshing, but the traditional media has much longer life between transfers - for an extreme example we still can read Sumerian clay tablets.
>Now it's easier and cheaper to store media than ever before
Storage cost was never a problem. The problem is the separation of the archival-worthy content from the info noise. Now, when the noise is multiplied, selecting is much harder than before. For my case, I only have a few photos worthy of putting into a 120-year photo album for my descendants, because I'm just an ordinary person.
It’s like the post about digitizing VHS the other day. It gets put off until the project is overwhelming.
I am in the process of scanning paper photos and photos on glass plates, dating from 120 years ago or so. Some of them survived two world wars, and mind you surviving WWII in Poland is no small feat for a piece of paper.
So while I do agree with the general sentiment of having multiple copies and backups, our digital world is woefully short-lived when compared to physical media.
On a related note, I am working on gallery/archive software, where the main assumption is longevity: whatever the UI du jour is, actual data must be kept in the simplest and most long-term file formats possible. Which means well-established and documented "traditional" image formats, text files with metadata, etc.
The odds of physical photos destroyed in a fire is far lower than a photo hosting service going out of business. There's no money in it because monitoring it for legal compliance is expensive and time consuming. Apple and Google and Facebook can do it because the monitoring has a business upside in terms of training ML models.
Physical reality is a useful abstraction.
I also remember when I left Facebook and asked for an archive of my photos, I got 800x600 images when I had uploaded much higher resolution ones. Service degradation is actually more common in FAANG because they have a massive moat and rely on the fact that 90% of their customers don't know their 1's from their 0's.
I compared checksums of (recently uploaded)
1) original source images
2) result of clicking "Download" in the Web UI
3) what was contained in my most recent Takeout
And they all match. There might be issues with the API (didn't check), but that API is VERY limited anyway.
Someone recommended the github project called Timeliner to help me get the photos though. I just haven’t tried it yet.
This is because of a change they made.
I was absolutely distraught tbh. The feeling of being at the mercy of google is a horrible feeling. Especially when there is zero support or assistance.
I received links to over 100 tgz files (4GB each), and checked a couple. Each had JPG files and JSON files with extra metadata ("geoDataExif", "geoData" (inferred from my location history?), "title", "description", etc).
I can't tell if the actual images are the originals I uploaded, but I can confirm the images were archived as actual images, and in a serviceable format/structure.
https://github.com/mholt/timeliner
Timeliner is a personal data aggregation utility.
It collects all your digital things from pretty much
anywhere and stores them on your own computer, indexes
them, and projects them onto a single, unified timeline.
Supported data sources
* Google Location History
* Google PhotosHowever, it has tons of metadata files, empty folders, etc. I highly suggest unticking 'preserve directory structure' when unpacking.
Now, when I was just browsing the zips, it was folder after folder of junk I didn't even recognize, I guess metadata files of some sort. Rather than do all that, at least in Gnome, when I unzip it has an option of 'preserve directory structure' that I unchecked, which basically gets rid of all the folders. From there I was easily able to mass remove all the json and other non-photo files, leaving only the photos behind. Good luck.
It’s only good for backups though. There’s no way to browse/share memories so we reluctantly still also use Amazon photos...
1. It’s still a kind of vendor lock-in that I’m not crazy about
2. I’m generally underwhelmed by the synology apps (eg music) and prefer docker based foss alternatives when available
3. I was reading somewhere about DS moments (after another mention on this post) and I think they require a special folder structure?
2. Yeah, I only use DS Photo for backup / browsing on my phone. And Drive to replace Dropbox.
3. Not entirely sure about DS moments, as i use Photostation. In there it's basically each folder is an album (or sub/album).
This is just about choosing a cloud storage provider that I can feel reasonably confident will still have my photos backed up on any given day for the foreseeable future.
I think you can make much better choices than FAANG or a random company, and I advise anyone to go with that better choice.
Far better IMO to not rely on a single provider, especially in the case of backups.
(Source: my personal experience with Google Pixel "original quality" backups, and providing customer support to several hundred PhotoStructure beta testers).
It comes with an adapter for Google Photos, so you can use it to download / backup photos and videos in your library.
Note: there are some limitations to this approach – namely that rclone won't be able to download your original image files, even if you uploaded with 'original' quality. These limitations are documented on the rclone website: https://rclone.org/googlephotos/
I have it run via a cron job and backup files to a local folder.
I also use this to backup my gmail account: https://github.com/jay0lee/got-your-back
I store these on a local OpenSuse box I use as a NAS. It has a local drive, with a 2nd drive that is mounted and synced to weekly. Then the important stuff is rsynced offsite to a cloud server, rsync for pictures and other various "family" stuff. Borg for home drives with more private information.
That internal drive I backup to, runs btrfs. After backup is complete. I take a snapshot and keep a monthly one going back at least a year.
Does not really seem fit for purpose. I'll be looking into it if I can update it.
From there RClone encrypts and copies them to AWS Glacier Deep Archive (11 9s, etc). I'll also start coping them to Google Workspace (was GSuite).
This removes a lot of product change risks such as Google APIs not returning the original files, etc. AWS Glacier is secure and cheap unless I have to restore. If I'm doing a full restore it's likely that's the cheapest option so then I don't care.
Glacier Deep Archive works out to about $1/TB/month and because it's paid per GB it removes the risks of a company changing plans or some "unlimited but not really" problem.
One piece of advice, think about your local disaster risks. I'm in an earthquake risk area. I ruled out most off-site backup options that involved moving drives around due to those all being in my same disaster area and costs of bank boxes etc.
Is iCloud better for this?
As Google Photos is my main usecase for being stuck in the Google Ecosystem, is this situation better with the iCloud?
If you delete something from the Adobe cloud, is it deleted from your backups too?
What do you use for sharing albums with friends?
What do you use for displaying photos on a TV?
I use iCloud photos for sharing with friends and family.
A takeout request can take "a few weeks", during which you just can't download them.
Google has also had too many cases of banning accounts with no recourse. I'd be scared for it to be my only backup of something so treasured.