ResearchHub product update — 4/7/22
v2 of RSC rewards, updating the paper upload flow, adding DOIs to ELN posts
New RSC reward algorithm
Right now on ResearchHub, users earn 1 RSC for every upvote their content receives. This v1 RSC reward algorithm is simple, easy to understand, and has not changed since ResearchHub was first launched.
As we now have a solid core initial community, we think it’s time to design a new RSC reward algorithm that can better incentivize ResearchHub’s users to create and share scientific content on the platform.
Our target is to distribute 50m RSC (5% of the total RSC supply) to our community annually in large part via this new v2 reward algorithm. Here’s how it works :
rsc_by_minute = calculate_rsc(votes_in_past_minute * MINUTES_IN_YEAR, .25)
rsc_by_hour = calculate_rsc(votes_in_past_hour * HOURS_IN_YEAR, .3)
rsc_by_day = calculate_rsc(votes_in_past_day * DAYS_IN_YEAR, .25)
rsc_by_month = calculate_rsc(votes_in_past_month * MONTHS_IN_YEAR, .1)
rsc_by_year = calculate_rsc(votes_in_past_year, .1)
rsc_distribute = rsc_by_minute + rsc_by_hour + rsc_by_day + rsc_by_month + rsc_by_year
rsc_distribute *= (1 — GROWTH)
Translation — we take the number of votes in the past minute, the past hour, the past day, the past month, the past year and extrapolate them all out to 1 year. We then give each bucket a weighted total and add them up. Finally, this number is multiplied by a growth multiplier which leaves room for the platform to grow in the future.
ELI5 — we use a weighted average around the number of upvotes received in 5-time buckets. A variable portion of this upvote will go to a pot held aside for the authors.
Be on the lookout for the first test of this new v2 reward algorithm this week!
Improving the Paper Upload Flow
Most of the bugs being reported these days come from our paper upload flow. We hear you! It’s annoying to want to add a paper to ResearchHub, only for the upload flow to break.
Right now our upload flow relies on a third-party open-source software tool called Manubot. It’s a great software that’s helped us get to where we are today. It helps us find all the information about a paper with just a link.
The only problem is that sometimes Manubot doesn’t respond properly, and thus we’re unable to fetch the metadata. Because of this, we fail the paper upload. We’re fixing this issue by doing a few things:
- We are building out our own service that fetches metadata about papers, given just a link.
- Rethinking our UX so that even if fetching metadata fails, users will be able to move forward and continue the paper upload.
- Adding metadata retrieval to the end of the process after the paper is uploaded
- Making the upload flow asynchronous. This will allow users to click out of the upload module while their paper is uploading. We will also send a notification whenever the upload is complete
Here’s a sneak peek at that the new flow:
Caching on our Home Feed
A couple of weeks ago we released a new homepage “trending” algorithm. It weights quality discussions much more heavily. In this process, we discovered some wonky behavior about our caching on the home feed and in other places in our app, which were causing some performance issues.
We improved our caching here quite a bit and these slowdowns should go away.
DOIs for ELN Posts
Last week we released the ability to add DOI’s for posts shared from ResearchHub’s ELN. All you have to do is click a checkbox when you publish your post and a DOI will be assigned to it. This will allow others to cite the content that you post to ResearchHub via the ELN!
Right now, it costs 5 RSC to have a DOI assigned to a post. Here’s an example of what it looks like on our staging instance: https://www.doi.org/10.55277/ResearchHub.mjwfhn9n
If you are interested in contributing to ResearchHub’s mission of accelerating the pace of scientific research — join us by signing up at https://www.researchhub.com/
