4 ms·

The fact that reputable scientific journals can just disappear from the historical record is one more reason why public non-profit archives like arXiv are gradu

by cs702 3y ago

The fact that reputable scientific journals can just disappear from the historical record is one more reason why public non-profit archives like arXiv are gradually, deservedly, becoming the dominant platforms for sharing and preserving new research.

greentea23 3y ago
I think libgen and scihub are even more future proof. Something as important as scientific (or any) literature must be distributed and uncensorable. Libraries are destroyed by authoritarians all the time throughout history.
- Vt71fcAqt7 3y ago
  Sci-hub hasn't been working for me for more than a year now for new articles.
  - senderista 3y ago
    "In accordance with an order by the Delhi High Court, Sci-Hub has suspended the addition of new papers since January 2021." https://www.editage.com/insights/a-look-at-sci-hubs-current-state-and-its-impact-on-scholarly-communication https://www.editage.com/insights/a-look-at-sci-hubs-current-...
xhkkffbf 3y ago
Why does making it "public" or "non-profit" make a difference? Things disappear because no one is paying to keep them up, either directly in time or indirectly with money. The government cuts off projects all of the time causing them to disappear. Witness, for instance, how Henry the VIII cut off Catholic monasteries effectively destroying their archives. Even things sent to well-funded organizations like the US National Archives disappear in confusion. So while I think you're right that places like arXiv make more sense, I don't think that they're automatically better. And it's an interesting question whether distributed, open source versions have an advantage. Sure, people can clone repos, but there's often no main person who promises to keep the repo alive making it possible for everyone to just assume someone else will cache a copy. This is a very complex problem at the deepest level.
- YetAnotherNick 3y ago
  There are many companies which have complete archive of arXiv, even arXiv put it in S3 and kaggle[1]. If someone just wants to serve static pdf it could be done very cheaply using R3 or something. I don't really think the old articles could die in the same way as journals which forbids archiving materials. [1]: https://info.arxiv.org/help/bulk_data/index.html https://info.arxiv.org/help/bulk_data/index.html
IanCal 3y ago
That's not the role of arxiv, and this problem is addressed much more easily with things like https://clockss.org/ https://clockss.org/ Edit - Clockss is a backup of academic content, mirrored at 12 academic institutes around the world. It's built on existing tech and means that things like dois should still work (if I understand right, https://doi.org/10.4103/2315-7992.204679 https://doi.org/10.4103/2315-7992.204679) . They currently backup over 50M articles.
- cs702 3y ago
  That may or may not be an intended "role" of archive services like arXiv, but it's one more reason, on top of many other reasons, why such services are gradually, and in my view deservedly, becoming the dominant platforms for sharing and preserving new research. Regardless of how we may think archival is most "easily" addressed, note that arXiv alone is already ~50x bigger than clockss.
  - IanCal 3y ago
    Sharing sure, it's a preprint platform. But why is it better for archiving content than clockss? How is it guaranteeing versions? What's the doing resolution like? > note that arXiv alone is already ~50x bigger than clockss. I'm sorry what's your measure of that? Arxiv is about 2-3M iirc and clockss is 50M.

by cs702 3y ago

greentea23 3y ago

I think libgen and scihub are even more future proof. Something as important as scientific (or any) literature must be distributed and uncensorable. Libraries are destroyed by authoritarians all the time throughout history.

Vt71fcAqt7 3y ago
Sci-hub hasn't been working for me for more than a year now for new articles.
- senderista 3y ago
  "In accordance with an order by the Delhi High Court, Sci-Hub has suspended the addition of new papers since January 2021." https://www.editage.com/insights/a-look-at-sci-hubs-current-state-and-its-impact-on-scholarly-communication https://www.editage.com/insights/a-look-at-sci-hubs-current-...

xhkkffbf 3y ago

Why does making it "public" or "non-profit" make a difference? Things disappear because no one is paying to keep them up, either directly in time or indirectly with money. The government cuts off projects all of the time causing them to disappear. Witness, for instance, how Henry the VIII cut off Catholic monasteries effectively destroying their archives. Even things sent to well-funded organizations like the US National Archives disappear in confusion. So while I think you're right that places like arXiv make more sense, I don't think that they're automatically better. And it's an interesting question whether distributed, open source versions have an advantage. Sure, people can clone repos, but there's often no main person who promises to keep the repo alive making it possible for everyone to just assume someone else will cache a copy. This is a very complex problem at the deepest level.

YetAnotherNick 3y ago
There are many companies which have complete archive of arXiv, even arXiv put it in S3 and kaggle[1]. If someone just wants to serve static pdf it could be done very cheaply using R3 or something. I don't really think the old articles could die in the same way as journals which forbids archiving materials. [1]: https://info.arxiv.org/help/bulk_data/index.html https://info.arxiv.org/help/bulk_data/index.html

IanCal 3y ago

That's not the role of arxiv, and this problem is addressed much more easily with things like https://clockss.org/ https://clockss.org/ Edit - Clockss is a backup of academic content, mirrored at 12 academic institutes around the world. It's built on existing tech and means that things like dois should still work (if I understand right, https://doi.org/10.4103/2315-7992.204679 https://doi.org/10.4103/2315-7992.204679) . They currently backup over 50M articles.

cs702 3y ago
That may or may not be an intended "role" of archive services like arXiv, but it's one more reason, on top of many other reasons, why such services are gradually, and in my view deservedly, becoming the dominant platforms for sharing and preserving new research. Regardless of how we may think archival is most "easily" addressed, note that arXiv alone is already ~50x bigger than clockss.
- IanCal 3y ago
  Sharing sure, it's a preprint platform. But why is it better for archiving content than clockss? How is it guaranteeing versions? What's the doing resolution like? > note that arXiv alone is already ~50x bigger than clockss. I'm sorry what's your measure of that? Arxiv is about 2-3M iirc and clockss is 50M.