News for hackers. Stuff that matters.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by danielecook on Wednesday January 31, 2018 @06:58AM (#16273186)
As the title suggests, the site is centered around discussions regarding scientific pre-prints and publications. Current sites supported are:

* arXiv * bioRxiv * Pubmed * Pubmed Central * DOI

Threads are submitted using the publication ID, DOI, or URL of a pre-print or publication. PDFs are fetched when available and used to generate a thumbnail.

Currently, the site only supports academic email addresses. I am trying to figure out the best way to open the site to all without having to manage spam. Let me know if you have any thoughts/ideas. If you are interested in accessing the site and do not have an academic email address please PM me with an email address to whitelist.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by o2348diuu on Thursday February 01, 2018 @10:16AM (#16282372)
I completely understand the issue of not wanting spam of various sorts broadly defined, but I do think restricting it to academic addresses is a huge mistake. As an established researcher on the verge of leaving an academic institution, but who is still very much interested in research, I'm becoming acutely aware of how cliquey academics is, and how problematic it is for the field and society as a whole.

The assumption that anyone who might have anything important to say would be at an academic institution is extremely dangerous (just as is the problem of access to journals outside of academic institutions, etc.).

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by afandian on Thursday February 01, 2018 @10:43AM (#16282618)
Seconded. I'm not an academic but I deal with academic publishing and I might want to participate.

My suggestion: use ORCiD.org for authentication with OAuth. This is a researcher profile. Most academics have one (or should!). They're free, the service is open. I think it strikes the right balance between barrier-to-entry for spam and accessibility.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by Vinnl on Thursday February 01, 2018 @11:38AM (#16283219)
I've implemented sign-in using ORCID, and although I support the idea, its UX is far from great, its documentation isn't either, and it also isn't terribly stable (it was down on my launch day...).
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by icebraining on Thursday February 01, 2018 @11:35AM (#16283186)
How does ORCiD.org itself deal with spam? Seems like they just use a captcha, no?
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by killjoywashere on Thursday February 01, 2018 @12:11PM (#16283561)
More about ORCiD for the skeptical: https://en.wikipedia.org/wiki/ORCID
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by afandian on Friday February 02, 2018 @05:06AM (#16289561)
If anyone has any questions I can ask someone from ORCiD to come over here and answer them. They're very open!
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by throwawayaway12 on Thursday February 01, 2018 @02:59PM (#16285307)
I somewhat disagree.

Academic journal papers are meant to be digested by experts in that topic, not the general population. It seems necessary to have a filter on registration to ensure that the platform is useful to these experts. You could imagine the restriction could be relaxed to research institution emails -or- a referral.

Even with a restrictive registration, the platform could still be useful to the a non-expert who is interested in the topic by reading the exchanges between experts.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by millzlane on Thursday February 01, 2018 @03:47PM (#16285722)
>Academic journal papers are meant to be digested by experts in that topic, not the general population.

TL;DR: Studies like these need scrutiny from everyone. Not just people that declare themselves experts.

I disagree.The latest E-Cig study says they found evidence of tumors in mice who were exposed to cigarette smoke....What they don't tell you in the abstract is that they used 10mg/ml nicotine E-juice and exposed the mice for 3h/day, for 12 weeks. The highest ratio most people vape at is 9mg/ml. No one smokes them for 3 hours straight. The study failed to include how exactly they produce the vapor, no mention of wicking material or coil metallurgy, temperature, or voltage of the coil. It could very well be that the tumors were cause by combustion of either the coil being heated above it's melting point, the wicking material burning, or the juice being combusted and not vaporized. Even though I'm not considered an expert in tumors or cancer I have learned that this test is flawed is missing information and doesn't really show anything that we as people don't already know "Heating things beyond the combustion point produces byproducts that could be harmful to our health." I can rest assured instead of listening to headlines that didn't bother to read or scrutinize the study. I think studies like these need scrutiny from everyone. Not just people that declare themselves experts.

Let's also be fair...just because I have an academic email address doesn't mean I'm an expert. I could be a janitor at JHU for intent and purposes.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by throwawayaway12 on Thursday February 01, 2018 @03:58PM (#16285821)
It would be helpful to link the study. Without seeing the paper, I would be cautious of separating the paper and conclusions it draws from a possible news article that has extrapolated or expanded the conclusion actually made in the paper.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by Arkanosis on Thursday February 01, 2018 @12:37PM (#16283815)
> I do think restricting it to academic addresses is a huge mistake

I have to agree. I'm no academic for almost 10 years, now, yet I've spent the last three years working with academics on my free time and it's really frustrating how hard it is to work without a “proper” affiliation. I've had to create my own “lab” (it's just a name, and it's only me) to have my name on posters, but I'm still unable to get access to some services because, well… because I'm not paid for the work I do, I guess.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by Kalium on Thursday February 01, 2018 @02:57PM (#16285293)
The academic world seems collectively addicted to credentialism and status. A friend of mine recently went from being a grad student in academic geology to private-sector geology. She's experiencing some pretty severe culture shock.

She's getting paid a reasonable amount. She gets paid overtime. Her opinions are given weight. Her physical needs are taken seriously. She gets credit for her work. She is, in general, treated as an actual human being capable of having expertise and bringing value.

The total result is that she's spent months going "WHY DID NOBODY TELL ME THE PRIVATE SECTOR WAS AWESOME?!"

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by politician on Thursday February 01, 2018 @05:13PM (#16286425)
Perhaps they can add support for alternate emails - sign up with your primary academic email, add a secondary. When a user leaves an institution, they can promote the secondary.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by killjoywashere on Thursday February 01, 2018 @11:04AM (#16282861)
Can you please allow .mil and .gov addresses? Those "government scientists" add up to a lot of researchers.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by euid on Thursday February 01, 2018 @09:07PM (#16287810)
Totally agree. I did government research for a while and published papers. Don't exclude them - anyone at the intersection of government work and research is likely to have meaningful contributions to this website.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by kiliantics on Thursday February 01, 2018 @09:02AM (#16281865)
Great job! I had thought of building something like this before. Some ideas I'd toyed with were:

- A tool for annotation so that people could comment on specific portions of papers. This could help new researchers in a field to more quickly understand the different broader contexts in which a result can be interpreted, or the technical limitations of a given method and how that should be considered when say using the results to motivate other work.

- An option for some form of anonymised commentary, so that people could voice their views on a particular work without fear of reprisal. I think this would be amazing for non-tenured academics to exert some form of power in the community, if a platform like this were to take off. Of course, it would make for potentially very messy and difficult to moderate discussions.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by throwaway080383 on Thursday February 01, 2018 @09:34AM (#16282076)
The first thing you mentioned, annotations, or maybe even inline Q&A/discussions, was my dream in grad school (math). I imagined something like a separate MathOverflow for each paper.

The basic problem is that I was constantly finding myself struggling with a tricky part of some paper from the 70's ("Why is this sentence true??"), and I knew many students before me had struggled with the same section, but there was no way for one really great expositor to clarify it, get upvoted, and be done.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by JorgeGT on Thursday February 01, 2018 @04:54AM (#16280627)
I'm sorry, but what is an academic email address?
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by lou1306 on Thursday February 01, 2018 @05:02AM (#16280662)
I guess, any email address provably bestowed by a university (like .edu emails in the US).
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by bringtheaction on Thursday February 01, 2018 @05:08AM (#16280686)
Speaking of which; https://github.com/leereilly/swot

Swot is a community-driven or crowdsourced library for verifying that domain names and email addresses are tied to a legitimate university of college - more specifically, an academic institution providing higher education in tertiary, quaternary or any other kind of post-secondary education in any country in the world.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by danielecook on Thursday February 01, 2018 @06:15AM (#16281018)
I actually used the swot database for the site. If anyone is interested this gist will download the swot repo and generate a dictionary of the email --> school/university in python:

https://gist.github.com/danielecook/19793d2faad66d83cef2514c...

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by JD557 on Thursday February 01, 2018 @05:04AM (#16280671)
Usually when you study/work on an university, you get an email address with the university domain (e.g. [email protected]).

Those are used in many places as a simple way to validate that you are a student or a researcher.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by JorgeGT on Thursday February 01, 2018 @05:33AM (#16280814)
Ops, I meant to ask how they were validating academic addresses. I'm actually a postdoc, but my university domain is @upv.es, and I've found that sometimes it doesn't get recognized in American websites which expect a .edu for claiming academic discounts or things like that.

@bringtheaction comment mentioning SWOT is an interesting solution. Although the name of my institution is outdated! I'll send a PR.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by avian on Thursday February 01, 2018 @05:33AM (#16280809)
The important part here is "usually". It should be stressed that using lists (such as SWOT mentioned in another comment here) can add to the bias towards US and other well-known universities.

I have been involved in academia for 10 years now, and I never held an email address that would match against the SWOT list.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by semi-extrinsic on Thursday February 01, 2018 @02:12PM (#16284799)
How about doing a scraping of e.g. 10k Google Scholar profiles with more than ~10 publications to get the domains of their verified email addresses? Then you have a much broader whitelist than just *.edu.

Or you can crowdsource this, say I want to register as [email protected]; in the registration form simply require me to give a link to someone's Google Scholar profile with an email address from the same domain.

It should also be noted that e.g. on Figshare, anyone can upload any PDF and get a DOI for it.

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by lars on Thursday February 01, 2018 @12:49PM (#16283933)
Only 39% of the submissions to the latest ICLR (machine learning conference) were from .edu emails [1]. There's a lot of research going on outside of academia, and there's also a large portion of academia that doesn't use the .edu TLD. That said, I don't have a solution to your problem of fighting spam.

[1]: http://webia.lip6.fr/~pajot/dataviz.html

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by khr on Thursday February 01, 2018 @11:17AM (#16283007)
Do you plan to add PsyArXiv? (https://psyarxiv.com/)
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by throwawayjava on Thursday February 01, 2018 @12:20PM (#16283656)
> Let me know if you have any thoughts/ideas

Many papers list authors' email addresses.

Scrape PDFs on the services you support for email addresses. If a domain name occurs in enough author email addresses and isn't on a "known generic" list (gmail et al), consider it "safe".

Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by afandian on Thursday February 01, 2018 @08:45AM (#16281774)
This looks really cool. I've emailed [email protected] about tracking your use of DOIs.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by akavel on Wednesday January 31, 2018 @08:02AM (#16273483)
Did you advertise it in any academic circles, then?
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by danielecook on Wednesday January 31, 2018 @02:24PM (#16276726)
I have mentioned sent it out to people I know and will be posting it elsewhere as well.
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by liamcardenas on Friday February 02, 2018 @12:17PM (#16292215)
Have you actually had problems with spam? Or is this a purely hypothetical problem?
Re:Show HN: A HN/Reddit-style site for scientific pre-prints... by claudius on Thursday February 01, 2018 @09:34AM (#16282081)
swot seems to be restricted to places of higher education, leaving out research institutes, such as the MPG association in Germany or the national labs in the US.