2020-07-07 15:37:5 (GMT)
to continue the discussion from the mailing list: I think git is kind of the best shot we have, for the moment, as it's easily copied. The question is, how to point people at it instead of a website whose academic hosting service will disappear within a few years… :p
2020-07-07 15:41:42 (GMT)
By git you mean github?
2020-07-07 15:42:8 (GMT)
No, I mean git. The tool, the storage format, etc.
My point is that a git repository can (and will) outlive github :)
2020-07-07 15:46:25 (GMT)
(so is this the point where it becomes obvious I'm a crackpot?)
2020-07-07 15:48:55 (GMT)
Yeah, that might be the reason most people voted for GitLab.
2020-07-07 15:51:45 (GMT)
On the other hand, IIRC Jasmin went for Inria GForge because he thought that Inria would continue to keep that service up for a long time. It turns out they won't. So will Inria Gitlab really live longer than GitHub?
2020-07-07 15:58:21 (GMT)
I wish ipfs worked well, it'd be perfect for that kind of thing. Alas it's not good software.
2020-07-07 16:15:35 (GMT)
well normally, I think, we think that published stuff lives long because it's in some libraries and/or archives. it's absurd to try to archive things really long term as individuals. All of those solutions really can only serve as some medium term thing.
2020-07-07 16:16:24 (GMT)
Papers live long. Academic websites, where raw data, benchmark tarballs, software snapshots etc. live, tend to not live that long in my experience :(
2020-07-07 16:16:36 (GMT)
As soon as the URL contains a ~, all bets are off.
2020-07-08 0:24:9 (GMT)
Long live artifacts!
2020-07-08 0:24:35 (GMT)
Stored in meant-to-be permanent stuff
2020-07-08 0:24:40 (GMT)
(like zenodo?)
2020-07-08 0:24:55 (GMT)
I forget what CAV and TACAS use to store their artifacts