5 ms·

join sci-hub rescue mission https://www.reddit.com/r/DataHoarder/comments/nc27fv/rescue_mission_for_scihub_and_open_science_we_are/ https://www.reddit.com/r/Dat

join sci-hub rescue mission https://www.reddit.com/r/DataHoarder/comments/nc27fv/rescue_mission_for_scihub_and_open_science_we_are/ https://www.reddit.com/r/DataHoarder/comments/nc27fv/rescue_... there is a new p2p client being developed (in Go) that makes sci-hub papers accessible via DOI https://www.reddit.com/r/libgen/comments/ompcvk/need_help_from_libgen_scimag_archive_data_holder/ https://www.reddit.com/r/libgen/comments/ompcvk/need_help_fr... it could use some help with indexing & IPFS integration (or just code review), see e.g. https://github.com/Trim21/sci-hub-p2p/pull/23 https://github.com/Trim21/sci-hub-p2p/pull/23

agilob 5y ago
With commits like: * fix (lines changed: +1,645 −95) * style * update * config * fix * fix style I have no trust for developer of this project Also these are torrents for libgen and sorted by torrents that needs seeders the most: https://phillm.net/libgen-seeds-needed.php https://phillm.net/libgen-seeds-needed.php
- rzzzt 5y ago
  They squash commits when PRs are merged and use "Conventional Commits"-styled messages in the main branch: https://www.conventionalcommits.org/ https://www.conventionalcommits.org/
the_biot 5y ago
I don't mean to sound cynical about such projects with good intentions, but I'm not convinced this effort will last without the cooperation and approval from the sci-hub admin herself. Having said that, any alternative way to access sci-hub would have to use a p2p architecture just like this.
- Anunayj 5y ago
  Looking at the project, They don't really connect to SciHub at all, SciHub on downloading a paper (using the credentials) saves it on libgen so it could be accessed later by others quickly. These just look like attempts to p2ping easy access to libgen (Which has a huge database (public), and would be sad to lose)
de6u99er 5y ago
SciHub would be a perfect example for a decentralized searchable repository.
betwixthewires 5y ago
I would seriously consider focusing on IPFS integration and indexing using a distributed table. Most of these documents are already pinned on IPFS, focusing on bittorrent requires users to change how the files are shared, and we have all seen the rot that happens with bittorrent and trackers, try to find something from 30 years ago using bittorrent, or even 10 years ago in many cases. If you want to make these documents available well into the future and make it easier for scihub users to transition, focus on IPFS.
- 6510 5y ago
  > try to find something from 30 years ago using bittorrent The clients are bad nowadays. Old utorrent could easily seed many thousands of torrents. With utorrent web I don't even know how to get the torrent files or magnets out. I hear some ppl with seedboxes run multiple clients and restart each periodically. Thats not how things are suppose to work.
  - betwixthewires 5y ago
    That might be a problem, but the big problem is that users of the system focus on consumer content such as movies and TV shows, and so people wind up seeding newer stuff or very popular stuff and as time goes by the number of seeds for older content goes down and the content gets lost, and there's little interest in seeding stuff from the 70s or 80s. Bittorrent is a bad way to archive anything long term because it is unreliable due to user dynamics. Can some other system fix those dynamics and become a more robust way to archive data? Maybe. Or maybe the problem is due to how people think. I like to think though that a system where downloads are not tracked by seeders would be more conducive to archiving.
    - 6510 5y ago
      Yes, things are overly focused on pirating new releases and the clients are feeding into that problem (or that line of thinking) I use to seed say 100 old things, sort by seeds, if there are leaches and no seeds give it 5-10 kb until a copy exists again (everything else I don't seed) I've also had things complete after years. Disk space is cheap. It seems doable for clients to periodically run over archived torrents and seed the dead ones. In stead the whole list is actively parsed for every action. Say someone has a youtube channel with 500 videos. Also distributing the videos over bittorrent shouldn't have to cost much. But then they have to jump a lot of hoops, run a client with 500 active seeds in it. Its so user-unfriendly a different [more dedicated] medium is probably a better idea at this point.

agilob 5y ago

With commits like: * fix (lines changed: +1,645 −95) * style * update * config * fix * fix style I have no trust for developer of this project Also these are torrents for libgen and sorted by torrents that needs seeders the most: https://phillm.net/libgen-seeds-needed.php https://phillm.net/libgen-seeds-needed.php

rzzzt 5y ago
They squash commits when PRs are merged and use "Conventional Commits"-styled messages in the main branch: https://www.conventionalcommits.org/ https://www.conventionalcommits.org/

the_biot 5y ago

I don't mean to sound cynical about such projects with good intentions, but I'm not convinced this effort will last without the cooperation and approval from the sci-hub admin herself. Having said that, any alternative way to access sci-hub would have to use a p2p architecture just like this.

Anunayj 5y ago
Looking at the project, They don't really connect to SciHub at all, SciHub on downloading a paper (using the credentials) saves it on libgen so it could be accessed later by others quickly. These just look like attempts to p2ping easy access to libgen (Which has a huge database (public), and would be sad to lose)

de6u99er 5y ago

SciHub would be a perfect example for a decentralized searchable repository.

betwixthewires 5y ago

I would seriously consider focusing on IPFS integration and indexing using a distributed table. Most of these documents are already pinned on IPFS, focusing on bittorrent requires users to change how the files are shared, and we have all seen the rot that happens with bittorrent and trackers, try to find something from 30 years ago using bittorrent, or even 10 years ago in many cases. If you want to make these documents available well into the future and make it easier for scihub users to transition, focus on IPFS.

6510 5y ago
> try to find something from 30 years ago using bittorrent The clients are bad nowadays. Old utorrent could easily seed many thousands of torrents. With utorrent web I don't even know how to get the torrent files or magnets out. I hear some ppl with seedboxes run multiple clients and restart each periodically. Thats not how things are suppose to work.
- betwixthewires 5y ago
  That might be a problem, but the big problem is that users of the system focus on consumer content such as movies and TV shows, and so people wind up seeding newer stuff or very popular stuff and as time goes by the number of seeds for older content goes down and the content gets lost, and there's little interest in seeding stuff from the 70s or 80s. Bittorrent is a bad way to archive anything long term because it is unreliable due to user dynamics. Can some other system fix those dynamics and become a more robust way to archive data? Maybe. Or maybe the problem is due to how people think. I like to think though that a system where downloads are not tracked by seeders would be more conducive to archiving.
  - 6510 5y ago
    Yes, things are overly focused on pirating new releases and the clients are feeding into that problem (or that line of thinking) I use to seed say 100 old things, sort by seeds, if there are leaches and no seeds give it 5-10 kb until a copy exists again (everything else I don't seed) I've also had things complete after years. Disk space is cheap. It seems doable for clients to periodically run over archived torrents and seed the dead ones. In stead the whole list is actively parsed for every action. Say someone has a youtube channel with 500 videos. Also distributing the videos over bittorrent shouldn't have to cost much. But then they have to jump a lot of hoops, run a client with 500 active seeds in it. Its so user-unfriendly a different [more dedicated] medium is probably a better idea at this point.