261–270 of 317 posts
I recommend Yandex.Disk (disclaimer - I used to work for Yandex) for backing up photos. It offers free unlimited photo auto-uploads with the original quality. Same for the videos for a very small price. The app itself is like Google Photos and Google Drive combined, but with fewer features.
OneDrive. If you can get 5 other people to go in on a family plan with you, it's $17/year for 1TB storage. You can get OneDrive app for your phones and tablets to auto sync camera roll, etc. to OneDrive. As an added bonus, OneDrive is really well integrated with Windows 10 (if you are a windows user) so your pictures will just magically show up on your system as files you can drag and drop.
Do your own tests, but Web and Windows filessystems get out of sync easily.
Pausing/unpausing the client does not work, resetting seems dangerous and I'd prefer not to do it.
You're certainly welcome to try PhotoStructure! It's got very robust de-duplication, supports portable libraries (so you can automatically sweep your photos scattered across many drives and computers into a single timestamped folder structure), and a novel UI showing a random sampled view of your library. There's also an active forum for both support and to discuss what and how new features are built out. https://pho…
WARNING! Be aware that this is an Electron app. I tried it, and using the interface was like swimming in glue. I presumed this meant "Electron app" and a quick check confirmed it.
https://photostructure.com/server/photostructure-for-servers...
Earlier quoted context omitted.
I had boxes and boxes of old drives that I wanted to get organized which drove that feature: https://photostructure.com/about/introducing-photostructure/... It's also why I implemented bitrot detection and metadata inference. A beta tester just called the date parsing heuristics "freaking awesome and so close to black magic…" :) https://forum.photostructure.com/t/combining-images/524/7 (And if you prefer the command…
Hey, PhotoStructure looks amazing. I'm intrigued. How and when does bitrot detection work? I couldn't find it on your site. I spend a decent amount of time making PAR2 volumes so I can detect failures as I've not come up with a better idea.
Earlier quoted context omitted.
> Haven't you all learned by now that you are the product if it is free. So when I'm running the Linux kernel, or compiling with gcc, or editing with vim, I'm the product? Could you explain to me how exactly? Sorry but that sentence is completely stupid.
A service is completely different... You are storing gigabytes of files on a remote server, which costs a lot of money, and get services to modify or whatever your pictures remotely... And, compared to Linux, it works out of the box... (ok, this was a troll).
Earlier quoted context omitted.
It's valid when speaking about a service. None of those are. Services have operating costs, if you are not paying it means someone else is and your data and attention is very likely part of the package.
> Services have operating costs So does ordinary, non-service software (like mentioned, Linux, gcc or vim). Hosting of their websites, mailing lists, CI, time of developers... they all cost very real money. And yet, somehow these programs don't track you and sell your data. I don't mind paid services, but I don't think they should get a free pass for being unethical just because they run on someone else computer, und…
Services are different, in that each individual user will tend to place an additional load on the systems. It is much more obvious if there is 10 GB of data that can be attribute to virgoerns which must be managed and much more obvious that virgoerns should perhaps contribute to that.
> I don't mind paid services, but I don't think they should get a free pass for being unethical just because they run on someone else computer, under a different label.
This is an independent question from the above. Services, paid or free, should be obliged to act ethically. If they won't do that voluntarily, then they must be regulated. Part of this should be placing an explicit value on the data that they extract and including it in invoices; during some months, Google should perhaps pay you for your use of their services. Certainly if they're going to go around and cap your usage of their storage at 15 GB, then they should accept the reasonableness to cap their usage of your data.
[1]: "Can" doesn't necessarily "invariably must be able to"; individual circumstances and tradeoffs will have an effect on the extent to which the costs can approach constant values as users grow.
Earlier quoted context omitted.
The self-hosted audience is also very small. I think you made the right choice.
For photos I’d agree as most customers are end users. Don’t underestimate this for other products though. Self hosting is a huge feature for selling into large corporations with data security policies.
Earlier quoted context omitted.
Can the search feature of ente become comparable to Google Photos? AFAIK, ente does E2E encryption, which means that unless they index the EXIF data on each single device, and also run the image recognition algorithms on device, the search is never going to be comparable to Google Photos.
I can't live without the face / object detection. It's been so useful
How are you housekeeping your photo library? I'm hosting mine in iCloud, and its size is increasing year after year. I'm close to the 150-200 GB of photos/videos. Deleting photos or videos manually would take a lot of time. How do you do it?
Earlier quoted context omitted.
I really didn't have much luck with the de-duplication. I tried photostructure for about 3 weeks and since stopped. I was trying to pull photos from multiple different drives/folders into one central organised place. A few issues: - The de-duplication didn't seem to work, it at least was defeated by my messy files. I spent many hours manually going through the results and finding many dupes. Then I lost interest in d…
> if I browse using the PhotoStructure app it will guide the dupes Yes: out of an abundance of caution, the unique (by SHA) variations of each asset are copied into your library. There are heuristics that pick which is the "best," and although those heuristics seem to be robust for most beta users, I didn't want to be the source of want data loss. > datestamp info in the filename? The metadata in the file is trusted…