5 ms·

Alas, I think there is no "best way" right now. All solutions fall short. And the problem is not just with library management (although that's bad, too), but ev

by jwr 12y ago

Alas, I think there is no "best way" right now. All solutions fall short. And the problem is not just with library management (although that's bad, too), but even with sharing photos with your family. For a couple of years I've been paying SmugMug to host my galleries, but recently finally got annoyed: they are great for pro photographers (who sell photos), but their goals are not aligned with mine. What I need is fairly simple: * online site, available at a specific (ONE) URL, that my entire family can bookmark (ONCE), * login/password required to enter, and I want multiple login/passwords, * HTTPS (duh), * no borg-minds harvesting my photos for advertising data, or doing face recognition, * good image quality, fast and responsive, both on mobile and desktop, * single gallery index, automatically ordered by date, * gallery date ranges extracted automatically from photos, * easy adding of new galleries, preferably just by uploading a directory with a bunch of images exported from Aperture/Lightroom/Photos/Whatever, * simple text-based search that finds galleries. Google Picasa, Flickr, SmugMug, whatever you pick — they all fail at this simple task (for various reasons). And some people even use Facebook to share images with family (sigh). I got so annoyed that I finally wrote my own solution. Which works. Costs me 6EUR/month for a Hetzner.de virtual server, but I can finally have all of the above "features" without any extra annoyance. Plus I get significantly better image quality than any other solution I've seen (mozjpeg kicks ass). And I don't depend on anyone, in particular on any company's whims or "strategic cutbacks", nor am I "the product being sold". I find it amazing that it's 2015 and I'm writing software that manages private online photo galleries. I think it doesn't speak well of the current business/tech landscape. (Incidentally, software to manage your photo collection is a whole another story. After Apple screwed me over with Aperture I am thinking hard about what to do about my photo archives.)

StavrosK 12y ago
I use Pixieset for that (mostly to share shoots with clients), and it does most of what you describe, but yes, even that isn't perfect.
walterbell 12y ago
For organization of a local collection of photos or other digital assets, http://www.cdwinder.de http://www.cdwinder.de looks promising. I wish there was an open-architecture alternative that could be scripted.
k-mcgrady 12y ago
Congrats on writing your own solution. I wonder though do you know how much your family actually uses it? In my experience if I share some photos via a link they view it and that's it. If I send a few select photos via MMS/WhatsApp they save them to their device and come back to them every now and then when browsing their own photo gallery. Does a lack of solution signal a lack of need in this case? P.S. Any chance you'll open source that software if there are other people in the same position (or create a closed source SAAS product)? P.P.S A product I've found that come close to what you need is iCloud shared photo streams. Did you look at that or does it fall short in some areas?
- jwr 12y ago
  This is a good point. Yes, they do use it — as opposed to all other solutions. People will use something which is simple and permanent. There needs to be just ONE URL (so forget about those picasa E-mail links) with a complete archive of everything. It needs to work quickly and reliably. That way everybody knows that whenever they want to access any photos from the archive, they can do so within seconds.
kerkeslager 12y ago
Have you considered releasing your code under the GPL? A lot of your concerns are shared by other people in the free software community. Ideally the solution to these kinds of problems would be decentralized, as a centralized server creates some issues: a) it requires you to acquire and maintain a server and b) it is susceptible to abuse[1]. But there are some technological issues remaining to be solved before decentralization is really available. I think federated systems like this--where multiple people connect to the same server but anyone can spin up their own server--are be a good stand-in for the present. [1] https://www.gnu.org/philosophy/who-does-that-server-really-serve.html https://www.gnu.org/philosophy/who-does-that-server-really-s...
onethumb 12y ago
Hey there, CEO & Chief Geek at SmugMug here. Seems like we do all of those things with the exception of always-on-HTTPS (which we're actively working on as I type this). You sound like exactly our target customer, so if you've got some time, I'd love to understand what these things are. I'm sure there are nuances here that I'm missing or not fully understanding. Thanks!
- jwr 12y ago
  I don't think I am your target customer. Or perhaps you think I am, but it certainly doesn't show in your product. I just checked and here's what I wrote to you in our support discussion from 2013 (when you introduced Sharegroups, which I thought would solve my problem): > "I'm not sure you thought this out — I can't be expected to manually add every new gallery to the sharegroup (which is not an easy process if you have >50 galleries), and then rearrange them manually (something I already did within my folder anyway) using the old interface." There are a number of problems with SmugMug. The most important one is the amount of work I have to do to add each new event (gallery). I had to fight the category system (I don't even want categories), remember presets (will that one work?), manually rearrange galleries, and perform a number of other chores. And there were minor annoyances, such as that at the end of it all, SmugMug still insisted on placing "Powered by SmugMug" branding on my pages. For comparison, so that you understand, these are the steps I have to take right now (with my software), starting from a folder full of images exported from Aperture/Lightroom: * run a script that converts/resizes/prepares images, * upload the whole resulting directory to my server. That's it. There is nothing more to do. Galleries get sorted automatically, date ranges are automatic, featured photo is chosen either randomly or based on a "featured" keyword. Galleries are named after directories by default. Galleries are reverse-sorted by date. There are no categories. Family and friends can log in, view pictures, search galleries. Oh, and it also supports retina images, and works very well on mobile devices. I think if you want to address use cases like mine, you have to realize that those photos are not the center of our lives, they are just a tiny part of life. I am not a pro photographer obsessing over how I want to categorize my galleries and how I will sell my pictures. I want to spend the absolute minimum of time dealing with placing them online.
swimfar 12y ago
I've been looking for something like this for a long time, not just for new photos, but for old family photos too. Another thing that I've really wanted was the possibility for anyone to add different tags to photos. So it could be people's names, places, objects in the photos, date (even generic "1980's", or just a year with no specific month or day if they are unknown), etc. Allowing anyone (from the group of people you allow to view or tag the photos) to add tags helps with old pictures where the person uploading it might not know who certain people are. And someone might see it and say, "oh that was at great grandpa's birthday party up at the cabin" and add his name and other tags to the picture. Then with the tags you can do cool things such as "search for all pictures of my dad and my grandpa together before 1975" or "search for all pictures of my aunt with a horse", etc. Also, I love the idea of being able to host the content myself. It might be a little more of a hassle, but then you don't need to worry about the service shutting down or removing features. Making your code available to others would be nice, but if you have a decent implementation it's certainly something people would pay for as well.
- jwr 12y ago
  Tags would actually require very little effort. I already read the IPTC "keywords" exported by Aperture/Lightroom/Whatever, to determine which image should be featured on the gallery cover (the first one with the "featured" keyword). As for hosting and business side of things, I thought about providing a service like that, but decided not to. It's a tough business: I don't want to deal with people posting dick pics or other kinds of porn, I don't want to go to jail because someone posts child porn to their account. And I definitely can't match the support level that companies like SmugMug provide. That business is mostly about support. I might release the software for everyone to use. It's not that complex anyway, you just have to make the right architectural decisions (this is what is really hard). I would probably have to change the thing so that it does everything on the server side, because right now image processing is done on the client (you run a perl script, then upload the directory that it produces). And to quickly answer those suggesting the GPL — if I do release it, it will not be the GPL. It will be one of the more permissive licenses. I don't want to block people's freedom to do anything they want with the software, and I think the threat of companies "taking stuff closed-source" is no longer a threat at all, because it makes no economic sense to do so. So if I do get around to it, it will be MIT or something similar. What could actually work well would be an Amazon AMI image, a Docker container, or something similar, that would be very easy to host.
ss1111 12y ago
Ever see OpenPhoto? It was on github and open source.. You have to self-host but it's pretty good. I was using it for this purpose a couple of years ago. I believe the dev team behind it also have a commercially-hosted offering called (maybe) Trove?

by jwr 12y ago

StavrosK 12y ago

I use Pixieset for that (mostly to share shoots with clients), and it does most of what you describe, but yes, even that isn't perfect.

walterbell 12y ago

For organization of a local collection of photos or other digital assets, http://www.cdwinder.de http://www.cdwinder.de looks promising. I wish there was an open-architecture alternative that could be scripted.

k-mcgrady 12y ago

Congrats on writing your own solution. I wonder though do you know how much your family actually uses it? In my experience if I share some photos via a link they view it and that's it. If I send a few select photos via MMS/WhatsApp they save them to their device and come back to them every now and then when browsing their own photo gallery. Does a lack of solution signal a lack of need in this case? P.S. Any chance you'll open source that software if there are other people in the same position (or create a closed source SAAS product)? P.P.S A product I've found that come close to what you need is iCloud shared photo streams. Did you look at that or does it fall short in some areas?

jwr 12y ago
This is a good point. Yes, they do use it — as opposed to all other solutions. People will use something which is simple and permanent. There needs to be just ONE URL (so forget about those picasa E-mail links) with a complete archive of everything. It needs to work quickly and reliably. That way everybody knows that whenever they want to access any photos from the archive, they can do so within seconds.

kerkeslager 12y ago

Have you considered releasing your code under the GPL? A lot of your concerns are shared by other people in the free software community. Ideally the solution to these kinds of problems would be decentralized, as a centralized server creates some issues: a) it requires you to acquire and maintain a server and b) it is susceptible to abuse[1]. But there are some technological issues remaining to be solved before decentralization is really available. I think federated systems like this--where multiple people connect to the same server but anyone can spin up their own server--are be a good stand-in for the present. [1] https://www.gnu.org/philosophy/who-does-that-server-really-serve.html https://www.gnu.org/philosophy/who-does-that-server-really-s...

onethumb 12y ago

Hey there, CEO & Chief Geek at SmugMug here. Seems like we do all of those things with the exception of always-on-HTTPS (which we're actively working on as I type this). You sound like exactly our target customer, so if you've got some time, I'd love to understand what these things are. I'm sure there are nuances here that I'm missing or not fully understanding. Thanks!

jwr 12y ago
I don't think I am your target customer. Or perhaps you think I am, but it certainly doesn't show in your product. I just checked and here's what I wrote to you in our support discussion from 2013 (when you introduced Sharegroups, which I thought would solve my problem): > "I'm not sure you thought this out — I can't be expected to manually add every new gallery to the sharegroup (which is not an easy process if you have >50 galleries), and then rearrange them manually (something I already did within my folder anyway) using the old interface." There are a number of problems with SmugMug. The most important one is the amount of work I have to do to add each new event (gallery). I had to fight the category system (I don't even want categories), remember presets (will that one work?), manually rearrange galleries, and perform a number of other chores. And there were minor annoyances, such as that at the end of it all, SmugMug still insisted on placing "Powered by SmugMug" branding on my pages. For comparison, so that you understand, these are the steps I have to take right now (with my software), starting from a folder full of images exported from Aperture/Lightroom: * run a script that converts/resizes/prepares images, * upload the whole resulting directory to my server. That's it. There is nothing more to do. Galleries get sorted automatically, date ranges are automatic, featured photo is chosen either randomly or based on a "featured" keyword. Galleries are named after directories by default. Galleries are reverse-sorted by date. There are no categories. Family and friends can log in, view pictures, search galleries. Oh, and it also supports retina images, and works very well on mobile devices. I think if you want to address use cases like mine, you have to realize that those photos are not the center of our lives, they are just a tiny part of life. I am not a pro photographer obsessing over how I want to categorize my galleries and how I will sell my pictures. I want to spend the absolute minimum of time dealing with placing them online.

swimfar 12y ago

I've been looking for something like this for a long time, not just for new photos, but for old family photos too. Another thing that I've really wanted was the possibility for anyone to add different tags to photos. So it could be people's names, places, objects in the photos, date (even generic "1980's", or just a year with no specific month or day if they are unknown), etc. Allowing anyone (from the group of people you allow to view or tag the photos) to add tags helps with old pictures where the person uploading it might not know who certain people are. And someone might see it and say, "oh that was at great grandpa's birthday party up at the cabin" and add his name and other tags to the picture. Then with the tags you can do cool things such as "search for all pictures of my dad and my grandpa together before 1975" or "search for all pictures of my aunt with a horse", etc. Also, I love the idea of being able to host the content myself. It might be a little more of a hassle, but then you don't need to worry about the service shutting down or removing features. Making your code available to others would be nice, but if you have a decent implementation it's certainly something people would pay for as well.

jwr 12y ago
Tags would actually require very little effort. I already read the IPTC "keywords" exported by Aperture/Lightroom/Whatever, to determine which image should be featured on the gallery cover (the first one with the "featured" keyword). As for hosting and business side of things, I thought about providing a service like that, but decided not to. It's a tough business: I don't want to deal with people posting dick pics or other kinds of porn, I don't want to go to jail because someone posts child porn to their account. And I definitely can't match the support level that companies like SmugMug provide. That business is mostly about support. I might release the software for everyone to use. It's not that complex anyway, you just have to make the right architectural decisions (this is what is really hard). I would probably have to change the thing so that it does everything on the server side, because right now image processing is done on the client (you run a perl script, then upload the directory that it produces). And to quickly answer those suggesting the GPL — if I do release it, it will not be the GPL. It will be one of the more permissive licenses. I don't want to block people's freedom to do anything they want with the software, and I think the threat of companies "taking stuff closed-source" is no longer a threat at all, because it makes no economic sense to do so. So if I do get around to it, it will be MIT or something similar. What could actually work well would be an Amazon AMI image, a Docker container, or something similar, that would be very easy to host.

ss1111 12y ago

Ever see OpenPhoto? It was on github and open source.. You have to self-host but it's pretty good. I was using it for this purpose a couple of years ago. I believe the dev team behind it also have a commercially-hosted offering called (maybe) Trove?