You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Gajen has brought up some interesting concepts regarding custom validation layers for paper submission. So this issue intends to expand on that topic.
He said: "Validation happens in multiple layers. For higher validation layers, we should give flexibility to users. Each validation layer is like a vetting process to create a list of curated ideas with desired attributes. The best case scenarios is the researchers make groups by their research interests and are involved in creating validation rules and (...) we only maintain the base layer validation and tech".
I like this a lot! But I am concerned about the possibility of censorship and/or manipulating custom validation rules to satisfy a group's bias. How can we avoid this?
The text was updated successfully, but these errors were encountered:
gil-air-may
changed the title
Custom Validation Layers - Possible Shortcommings
Custom Validation Layers - Possible Shortcommings?
Feb 21, 2019
The word 'custom' and 'bias' come hand in hand. Whenever we customize something, we do it to bring it closer to a bias (or a goal).
Positive or Normal Scenarios
Bias by research interest: Researchers will make groups within themselves to make lists for their subjects of interest.
Bias by novelty: A list can be created to publish only the ideas that are deemed novel
Bias by size: A list can be created to publish only 1-page papers
Bias by country: A list to publish the ideas from Brazil
Negative Scenarios
A group can create a list that is beneficial to their ideas. This is inherently not a bad thing. But it is highly likely that nobody else will use this. For example, anyone can fork Bitcoin but if there are not enough nodes supporting the fork, it will be worthless.
A group can take over an already established good list with a decent size of users and change the validation rules to favor them. If the researchers are not satisfied, they can always fork the list and implement their own validation rules on it.
In this context, for the sake of analogy,
a list of ideas = a blockchain
number of users/validators in a list (e.g. number of peer reviewers) = hash rate of a blockchain
Gajen has brought up some interesting concepts regarding custom validation layers for paper submission. So this issue intends to expand on that topic.
He said: "Validation happens in multiple layers. For higher validation layers, we should give flexibility to users. Each validation layer is like a vetting process to create a list of curated ideas with desired attributes. The best case scenarios is the researchers make groups by their research interests and are involved in creating validation rules and (...) we only maintain the base layer validation and tech".
I like this a lot! But I am concerned about the possibility of censorship and/or manipulating custom validation rules to satisfy a group's bias. How can we avoid this?
The text was updated successfully, but these errors were encountered: