Comment by roel_v | Hacker News Reader

Comment by roel_v | Hacker News Reader

"For another, the DOI is a persistent link. Broken links are a growing problem for academia, as link structures are changed and online content is edited. “If your persistent link is pointing to something on Figshare, which is ‘this GitHub repository at that version, at that release,’ then even if the GitHub repository changes or Figshare changes its link structure, that DOI will always point to that object,” Mark Hahnel, founder of Figshare said."

Yeah... well there is the achilles heel of this whole thing. "Always" in internet time means "2 or 3 years, or until we fail to close another round of funding" in real-world time. I'm probably being a cynical old fogey again, but let's see in 5 years time if this whole thing still exists (mozilla, for example, doesn't have a very solid track record of keeping projects up, to put it mildly) before I start putting URLs to this in papers that people will probably still reference every once in a while 5 years from now.

Then again, if everybody thinks like I do, it'll never get off the ground - classic catch-22.

DOI exists since 1997, which is pretty much since forever in internet time.

Figshare on the other hand, not so much.

Yes, I don't trust either figshare or github to be here in 10 years (there was a time when I thought sourceforge would be around for ever, now I keep expecting to see it disappear any day now).

Why don't they partner with the Internet Archive? Its specifically structured to be a long term archival/reference system.

Figshare is mirrored by several universities in the CLOCKSS scheme:

http://figshare.com/blog/Ensuring%20persistence%20on%20figsh...

Twelve research universities keep copies of all the data and will make it publicly available should Figshare implode.

Other data repositories, like Dryad, do the same thing, and grant agencies with data deposition requirements usually require you to deposit data somewhere that has a long-term plan to ensure access.

So as long as the twelve universities manage to update all the DOIs to point at the new locations, Figshare DOIs should be reliable.

I agree, it always make me cringe when privately held companies want to define an "open standard" for scientific citations that (surprise!) relies completely on their proprietary infrastructure. I still remember the case of Mendeley, which promised to build an open repository for research documents, and which is now a subsidiary of Elsevier, an organization that does not really embrace "open science", to put it mildly. I think what we really need (and the article mentions this at the end) is an open standard that can be implemented by anyone, most notably universities and research institutes.

Reply on news.ycombinator.com