Ask HN: How do you manage your photo library? In the age of excellent smartphone cameras, I would expect the average 30 year old person to have several thousand photographs they would rather not lose. How do you keep them straight? What software do you use? What naming strategy? Do you transfer to your long term storage on a regular interval? Exclusively cloud based? Please include the approximate size of your library as well. |
I've written a lightweight JS app authenticated by AWS Cognito that uploads my photos to S3. It even has a thumbnail making Lambda function, and a comments sections for family and friends. I wrote about it here: http://www.pesfandiar.com/blog/2017/03/10/serverless-photo-s...
One approach could be: buy a USB stick every few months and copy a few thousand random photos from your collection to it. If your hard disk dies, grab as many of those USB sticks as you feel like making time for, and restore the photos from them.
If you are really adventurous, put a quotum on the growth of your collection (e.g. 50 photos per month or x MB per month) and force yourself to throw away older photos if you want to keep more than 50 for a month. If you set the limit at x MB per month and really want to keep 100 photos, you can give up photo quality by recompressing some photos.
I feel the opposite, I keep everything (even 1000s of screenshots) because managing it would be a higher workload than otherwise. Furthermore, even the worst photo of something insignificant can bring back vivid memories of the time and location. If I "pick" only the "big" moments all the other ones are lost.
I have almost 30gb stored in Dropbox sorted by phone (I change phones a lot), every week or so I just click "import all" from my phone.
Yes, you can recover deleted files in Dropbox but to do that, you must know you deleted something, and then, what you deleted, or you must periodically check for deletions in the Dropbox UI.
Also, Dropbox probably will happily sync files that were written with bit errors on your hard disk.
Acceptable risk? I think so, but it means accepting some decent risk of losing photos.
My point is that many people likely would be happier if they spent, say, two evenings a year watching past photos together with their loved ones than when they spend that time making (more) sure their collection stays perfect (and at thousands of photos a year, two evenings is _nothing_ if you want to make sure your archive is bit-for-bit safely stored and properly indexed so that it is reasonably easy to find photos on demand)
Of those I delete the obviously-terrible images, where the exposure was off, the person blinked, or had their mouth open in an awkward position as we were talking during the shooting process.
The result of a good shoot is 20 images that are worth publishing, and a poor shoot might result in only 10.
All my images are RAW (.CR2), so they're pretty large. Despite that keeping them all is the only approach I feel comfortable with.
Backups are handled by archiving to a local NAS, and uploading offsite once per evening. I'm careful to ensure that when I import the images to my desktop system I don't erase the CF cards before I have that offsite copy uploaded (which sometimes takes a few days).
Whenever I have a new image, I upload it and tag it.
I periodically back up the database and keep it stored on a separate drive. All of my drives are also backed up to Backblaze in the event my house burns down or all my drives get electrically fried, etc.
After toying with Dropbox, Amazon Photos, Google Photos and a host of others, what I ended up doing was connecting a 1TB drive to a Raspberry Pi, then setting up a Samba mount on the local LAN. There's a read/write "dropbox" folder where we can upload photos manually, and I have a Python script running out of cron that sorts everything into date-based folders (https://github.com/andrewning/sortphotos). The date-folders are read-only on the Samba share.
I also have Dropbox configured on our mobile phones. Photos can be auto-uploaded to Dropbox, and I have CLI scripts that download them from there, and move them into the date-folders. The Python script de-dupes everything, and I move "duplicates" to an S3 folder where they're either purged after 30 days, or I'll review them to double check something wasn't ignored by mistake.
Everything on the 1TB drive is synched to S3 weekly, and moved quickly to Glacier. I also have a secondary 1TB drive that I'll sync to the master on a monthly basis.
So I don't get on-the-go access to my full photo archives, but I can move what I want to Amazon Photos if I really want something, and I've got enough redundancies to make me feel like I won't lose anything.
From there I'll curate a gallery of particular trip or event and upload to Google Photos where I can more easily share with friends and family.
The general workflow is that Dropbox automatically uploads photos from my phone and from my husband's phone to a "Camera Uploads" folder. My husband's photos are brought over to my Dropbox via a shared folder. Hazel watches the Camera Uploads folder and renames and then sorts photos into a directory structure of "YYYY > YYYY-MM." Each photo is names YYYYMMDDHHMMSS.EXT. It is incredibly easy for me to find any photo I am looking for. I am not locked into any software or file structure. I use Lyn.app or GraphicsConverter.app or even Finder.app to browse the photos. There are a couple of iOS apps that look into your dropbox directory and make albums out of folders.
So far, the biggest weakness of this setup is pictures that are not taken by me and the co-mingling of those images with personal photos. I need to be better about reviewing the monthly folder at the end of the month and placing those where they belong. Examples of the things that are in this category are photos of sports players/games/events, screenshots, memes, wallpapers, etc. I do have a separate but similar workflow for screenshots of video games, so that is not an issue.
I also had really poor backup habits when I was younger (yay for multiple copies of everything and just adding a random number to it) and have a huge chunk of photos that are all improperly dated for December 2013. This means that I have to spend a lot of time de-duping. I use DuplicateDetective.app to help with this, but photos are one area where I will not do that automatically. Sadly, I am missing a couple month chunk right around when my dog was adopted, but overall, my pictures from late high-school forward have survived pretty well. Once I have the old stuff mostly organized, this will self-maintain in the future.
I do not know the size of my photo library.
My photos are generally divided into 2 categories: high quality artwork photos and everyday pictures, usually from a cell phone.
The art pictures are divided into "collections", named according to whim. Most files have names as unique as the picture. I also have art reference photos here, as well as a few cell phone photo collections: Things that look glitched, manhole covers, and sunsets behind a utility pole.
The others are badly organized. I usually have a "dump" folder with unorganized photos. I try to name folders so I remember, but I forget.
Images from my iPhone are synced up to iCloud, and I never come around to put them in the structure above.
Costs are about $2/month for about 100 gigs so far.
Managed poorly. 1/2 of it Just slaved to Mac's Photo app. The other half in loose dated folders.
No tags, only a hand full of albums. No face detection. Scattered backups on DVDRS. Also backed up with the rest of the computer every week.
It's all coming from my phone camera, which I sync and cleanup every week or so.
I do not have a great strategy to access them all in one place. I was thinking of possibly utilizing Amazon photo backup that comes with prime, but it does not do video.
Maybe I should try setting up a cron job to rsync my DCIM folder. I haven't really heard of anything better that doesn't rely on some third party service.
I'd rather have access to a file system that has them and I can sort through that way, do copies, backups, etc. without having to touch a 20+GB file.
In addition to that, I find that photos are actually harder to manage when everything is a singular blob on all of your devices. When trying to free up space, I am not sure what can stay and go and what is safely stored elsewhere. I actually moved my parents from Photos to Dropbox for this reason (and that they are on Windows). It is now trivially easy to just erase all photos on anyone's phone and be confident it is living in an easy to find directory structure. I no longer have to spend hours making sure that the photos I think I have backed up are indeed backed up and nothing will be lost when Photos are taken off of an iOS device.
I have accidentally deleted files before and if it is a significant amount (e.g. if I try to do bulk processing on some) Dropbox notifies you by email.
I could easily duplicate my folder, or even triplicate it to minimise hard disk errors and not be near my Dropbox limit. However I poke around at pictures quite a lot and I don't think I've seen issues like this so I don't mind it.
(I think they are also synced to Google Photos now that I think about it).