My mom digitized many many old family videos, and wanted them online for sharing with family (including elderly & not-super-tech-savvy relatives). She asked me “should I just upload them all to a YouTube channel?”
Thankfully it was a phone call so my mom didn’t see my aghast expression. I prefer that big tech not index this stuff! Better to keep “in the family”
Seriously why does big tech deserve this free & super-private window into me & my ancestors lives?
So I wrote something[1] where:
* it’s fully free & open source
* cloud native
* plays on any device, any bandwidth, even if shitty
* yes my 90+yo Aunt Loretta (w00t to you Aunt Lo!) can use it on her phone & computer
* all data can be always encrypted, both source videos and derived/optimized assets
* and there’s more. please have fun
Basically point it at a source bucket on S3 or B2, and get your own private YouTube.
What I’ve built is very limited in functionality atm, but I believe the foundation is solid and plan to extend media support to photos and audio.
This can be a nice alternative to Plex/Google Photos/YT/etc.
It’s for when you don’t care about “building an audience” and in fact prefer that big tech can only see encrypted bytes from you.
Not saying you shouldn't do this, but by publishing under AGPL plus
If you are an individual person or a not-for-profit organization, and your usage of this software is entirely non-commercial, you may use this software under the terms of the GNU Affero General Public License, version 3, summarized below and reprinted in full thereafter.
you have effectively created a new license and it's not completely clear to me what that new license even means exactly, except that obviously a company should stay far away from it.
With regular AGPL, there is not a problem for a company to use the AGPL licensed software, it "just" can't offer Tivo-ised experiences or a website running modified AGPL code.
> All other non-permissive additional terms are considered "further restrictions" within the meaning of section 10. If the Program as you received it, or any part of it, contains a notice stating that it is governed by this License along with a term that is a further restriction, you may remove that term.
So it seems at best there is a need for a middle man who gets the AGPL licensed version that can then propagate it further under pure AGPL.
I think that language is meant for the case where someone takes an AGPL programs, slaps another restriction on it, and sends it along.
The last person in the chain can disregard the extra "conditions".
But this only works if someone distributed it under (only) the AGPL in the first. In the specific case with the software we are talking about now, that is not the case. It was originally distributed under this almost-AGPL.
But yes, the wording inside the AGPL makes it extra confusing exactly. It reads like those test where the instruction is "before you do anything, read all the questions".
This doesn’t make sense to me. If one clause says another clause can be removed, then doesn’t that create a contradiction where legally it’s unclear which of the two clauses “wins” the fight - the one adding additional restrictions or the one removing that clause?
The danger here is that your additional terms might be ineffective. The problem is, if you distribute it to an individual or a non-profit organization under the terms of the AGPL, then that individual or non-profit organization has the right to redistribute the software to anyone, including for-profit organizations, under the terms of the AGPL. It says so in the very first sentence after the copyright notice!
So if you do not want to allow this, then you must use something other than the AGPL. (Or maybe you could do some funky patching of the AGPL's terms, but you'd have to be careful to only do so by reference, since its text is copyrighted by the FSF, who only permit distribution of verbatim copies.)
Also, you should probably avoid calling it "fully free & open source" as you did in your original comment, since you intend for it to be neither Free nor Open Source in the sense ordinarily meant by FOSS.
In all dual-licensed FOSS projects which require licensees to comply with the terms of both licenses, both of them include an unlimited right to redistribution. Also, roughly no projects do that with the AGPL, since the AGPL includes the right to cast off any additional terms when redistributing.
But your intent, if you want to avoid for-profit organizations using your software, is to impose a restriction on users not to redistribute the software in certain ways. It is your right to add such a restriction to your software, but it is no longer FOSS, since that term necessarily implies an unlimited right to redistribution (within the law).
Your first claim is provably wrong; look at the license for MySQL as an example of a dual license where the licensee chooses which terms to follow based on their usage.
Redistribution is unlimited for all non-commercial use. This has been through legal review.
If you’re still confused, please contact me directly and I’ll be happy to answer any other questions.
> With regular AGPL, there is not a problem for a company to use the AGPL licensed software, it "just" can't offer Tivo-ised experiences or a website running modified AGPL code.
Nitpick: the company can use AGPL code wherever, it only needs to make the updated source code available to users. In many shops, devs would be perfectly ok with that, where legal depts are still mostly living in the world of "opensource it with MIT/Apache to make sure we can take it down tomorrow, just in case".
The goal was to have a platform that ingests the commonly enormous video files from old tapes, automatically cuts them, tags them based on content to make them easily searchable. My focus was on discoverability of scenes hidden in those long video files. The search bar would also randomly suggest tags to search for.
At some point I tried to work with a large video digitization provider and the video splitting ended up being too expensive to be viable for the proposed business model. Now it just auto generates thumbnails and lets you tag videos manually.
The project includes a business dahsboard that allows digitization businesses to send videos directly to customer accounts (deliveries need to be accepted).
Currently, I only use it for my own videos as well as for my MIL.
I like S3/B2 because the vendor only ever sees encrypted bytes. All decryption/plaintext is on the device. YouTube does not get you there. In fact their entire model is predicated on watching your every move.
As far as Peertube, I don't know enough about it. If I put a massive-bitrate video in some weird format on it, and then try playback on a crappy phone, will it work? If I go through a short tunnel, will it buffer or degrade quality gracefully? I don't want to worry about it.
> S3/B2 because the vendor only ever sees encrypted bytes
Got it, thanks!
> questions about Peertube
I don't know either. So far my experience with existing instances has been rather good but I didn't consciously test the use cases you mentioned. I've wanted to publish educational videos for a while but the idea of feeding the big nasty beast just breaks my heart.
I actually did try hosting the raw videos first. Playback is kind of terrible when the source bitrate is >20MB/s. You really need adaptive-bitrate streaming.
Thankfully ffmpeg support for DASH and HLS is very good. It’s not hard to transcode, this does all the right incantations. It can take a lot of CPU/time, but it’s a one-time thing.
nit: The language in the readme is called 'Marathi' instead of 'Maranthi'. It's my native language, and not a ton of people speak it online. Nice to see it here :)
Reverend Lovejoy: No, but He was working in the hearts of your friends and neighbors when they came to your aid, be they [points to Ned] Christian, [Krusty] Jew, or [Apu]... miscellaneous.
PS-- I also wanted my docs in any language, but also that anyone else could translate their app/docs easily, so I made this other thing[1], which is why these docs are available in so many languages
I wrote something like this, but for photos. I was tired of not being able to share photo albums with my family and friends.
And then I realized that there is a certain group of people who really need a private photo-sharing service without any oversight, and that not only do I want nothing to do with that despicable group, I don't ever want to realize that I helped them in any way.
So I decided to never release the code and never make it available.
Counterpoint: the bad people are already doing bad things. There is already plenty of secure communication technology, it’s just not super convenient. The highly motivated weirdos don’t mind that though.
There is a much larger group of people (parents) who need a secure and convenient way to share photos. They are not as driven as the perverts, so they need all the support and user-friendliness that we can provide.
We should build the tools that help the parents protect their kids, and let law enforcement deal with the other group.
If you discovered a new way to make a secure locking cabinet, and then found out that bad people would lock bad things in the cabinet, you’d stop working on the cabinet?
OK then, that’s entirely your choice. Just don’t forget you started working on this cabinet for your own legitimate use, and there are many more people who share your same legitimate problem, but who can’t write software, and they will be SOL.
Thankfully it was a phone call so my mom didn’t see my aghast expression. I prefer that big tech not index this stuff! Better to keep “in the family”
Seriously why does big tech deserve this free & super-private window into me & my ancestors lives?
So I wrote something[1] where:
* it’s fully free & open source
* cloud native
* plays on any device, any bandwidth, even if shitty
* yes my 90+yo Aunt Loretta (w00t to you Aunt Lo!) can use it on her phone & computer
* all data can be always encrypted, both source videos and derived/optimized assets
* and there’s more. please have fun
Basically point it at a source bucket on S3 or B2, and get your own private YouTube.
What I’ve built is very limited in functionality atm, but I believe the foundation is solid and plan to extend media support to photos and audio.
This can be a nice alternative to Plex/Google Photos/YT/etc.
It’s for when you don’t care about “building an audience” and in fact prefer that big tech can only see encrypted bytes from you.
Try it out and lmk!
[1] https://github.com/cobbzilla/yuebing