I kinda agree but there are still some bigs ''but''. Indie first of all means indipendent. Right now development of the procol relies very much on bluesky startup money. Semble, which also i too love, doesn't let me login with an account on eurosky pds if first i don't login into bluesky/similar. Also, can i backup my whole pds stuff on a single zip that i can restore on another pds? If not, we are still lacking fundamentals; if yes, please let me know how because i'd like to do it!
Also, the fact that we need to have RSS at appview level and not at pds level also makes it harder to follow directly a person...
Lastly, until the atmosphere main thing remains microblog, the rest is basically mmh ruined(?) by that. I should think more about this point but what I mean is that basically right now everything is like ''it also connects to bluesky'' while the basis should be ''it also connects to blento'' or ''it also backtracks to a community on the relevant topic'' etc instead of making impossible to follow people on bsky because they are used to posting infinitely like on twitter. I don't want social media...
sharpmars.space·4d
There are multiple points so I'll try answer each one best I know/can:
1) Protocol development (at least some parts) are getting moved to IETF, so that will make it less dependent on Bsky PBC
2) I'm not on Eurosky PDS (I have one account on another independent PDS) so can't really check, but from what I've seen you don't have to login to Bluesky, if you just put in your handle it redirects you to a log in screen on your PDS (aka the OAuth login screen)
3) Yea (somewhat). Backing up your data would be getting all your record data in a CAR file and then getting all your blobs (so images, videos or other files like that). You could be interested in https://pdsmoover.com/ as it handles moving data and also backups. I'm not really aware of any other user friendly software that could help with that and for example store backups locally or how to use them later to restore (didn't have to do a restore personally). I can mention that Bsky app has a "download CAR" button under account settings, but that doesn't include blobs. Also if useful I did write a Windows script for myself to make backups in an TAR archive file, would only need to change the handle inside the script for it to work for you, assuming you use Windows. https://tangled.org/strings/did:plc:irx36xprktslecsbopbwnh5w/3m44okqgu6d22
4) I mean PDS is mostly for storing user data, it's up to services to transform it to whatever format is needed, but it doesn't stop the person from deploying their own RSS service for their own data.
5) Yea, I get what you mean. Microblogging is still probably the biggest thing in the ecosystem, but hopefully with time this will change.
six.eurosky.social·2d
Thank you for the exhaustive answer! Regarding 3... What's up with the division between blobs and non blobs? Isn't everything in the PDS? I mean, obviously some services make you post stuff in storage buckets rather than directly in the PDS but is it the norm or...? Because if it is it means that using AT as a social Solid (protocol) doesn't make sense.
P.S. for the 2, I just had to force while it kept saying no user found on the overlay
sharpmars.space·5h
Well, records and blobs are really different from each other. One is structured data, has whole structure to prove that you wrote it and is expected to be exported in full for synchronization, which funnily enough is used for the same thing as the "Download profile data" button, while the other is just flat file storage and is addressed individually and fetched as needed. Size of these is also a big difference, record data is much smaller, while blobs can be big. Also blobs can be offloaded to S3, if you configure your PDS to do that, but all blobs are still managed by the PDS. Generally most blobs do go to your PDS, except ones that are far above the size limit of most of the PDSes (i think stream.place is doing something like that here, but not entirely sure).
PDS is kinda like a half database and half S3(in function). You wouldn't really store files in a database, you would just store a reference to the file (which is what atproto does with the blob reference, which is a hash of the file which is stored under your account).
(Probably should have just said this as it kinda explains why it is separate, also probably misinterpreted the comment, but hopefully I at least explained something.)
six.eurosky.social · 4d
I kinda agree but there are still some bigs ''but''. Indie first of all means indipendent. Right now development of the procol relies very much on bluesky startup money. Semble, which also i too love, doesn't let me login with an account on eurosky pds if first i don't login into bluesky/similar. Also, can i backup my whole pds stuff on a single zip that i can restore on another pds? If not, we are still lacking fundamentals; if yes, please let me know how because i'd like to do it! Also, the fact that we need to have RSS at appview level and not at pds level also makes it harder to follow directly a person... Lastly, until the atmosphere main thing remains microblog, the rest is basically mmh ruined(?) by that. I should think more about this point but what I mean is that basically right now everything is like ''it also connects to bluesky'' while the basis should be ''it also connects to blento'' or ''it also backtracks to a community on the relevant topic'' etc instead of making impossible to follow people on bsky because they are used to posting infinitely like on twitter. I don't want social media...
sharpmars.space · 4d
There are multiple points so I'll try answer each one best I know/can: 1) Protocol development (at least some parts) are getting moved to IETF, so that will make it less dependent on Bsky PBC 2) I'm not on Eurosky PDS (I have one account on another independent PDS) so can't really check, but from what I've seen you don't have to login to Bluesky, if you just put in your handle it redirects you to a log in screen on your PDS (aka the OAuth login screen) 3) Yea (somewhat). Backing up your data would be getting all your record data in a CAR file and then getting all your blobs (so images, videos or other files like that). You could be interested in https://pdsmoover.com/ as it handles moving data and also backups. I'm not really aware of any other user friendly software that could help with that and for example store backups locally or how to use them later to restore (didn't have to do a restore personally). I can mention that Bsky app has a "download CAR" button under account settings, but that doesn't include blobs. Also if useful I did write a Windows script for myself to make backups in an TAR archive file, would only need to change the handle inside the script for it to work for you, assuming you use Windows. https://tangled.org/strings/did:plc:irx36xprktslecsbopbwnh5w/3m44okqgu6d22 4) I mean PDS is mostly for storing user data, it's up to services to transform it to whatever format is needed, but it doesn't stop the person from deploying their own RSS service for their own data. 5) Yea, I get what you mean. Microblogging is still probably the biggest thing in the ecosystem, but hopefully with time this will change.
six.eurosky.social · 2d
Thank you for the exhaustive answer! Regarding 3... What's up with the division between blobs and non blobs? Isn't everything in the PDS? I mean, obviously some services make you post stuff in storage buckets rather than directly in the PDS but is it the norm or...? Because if it is it means that using AT as a social Solid (protocol) doesn't make sense. P.S. for the 2, I just had to force while it kept saying no user found on the overlay
sharpmars.space · 5h
Well, records and blobs are really different from each other. One is structured data, has whole structure to prove that you wrote it and is expected to be exported in full for synchronization, which funnily enough is used for the same thing as the "Download profile data" button, while the other is just flat file storage and is addressed individually and fetched as needed. Size of these is also a big difference, record data is much smaller, while blobs can be big. Also blobs can be offloaded to S3, if you configure your PDS to do that, but all blobs are still managed by the PDS. Generally most blobs do go to your PDS, except ones that are far above the size limit of most of the PDSes (i think stream.place is doing something like that here, but not entirely sure). PDS is kinda like a half database and half S3(in function). You wouldn't really store files in a database, you would just store a reference to the file (which is what atproto does with the blob reference, which is a hash of the file which is stored under your account). (Probably should have just said this as it kinda explains why it is separate, also probably misinterpreted the comment, but hopefully I at least explained something.)