Can't access your app. Looks like it has crashed. Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)
Looks good now
Show HN: Paperkast
51–60 of 62 posts
Re: Show HN: Paperkast
#52TypeError in Home#index Showing /opt/lobsters/app/views/layouts/application.html.erb where line #85 raised: no implicit conversion of Symbol into Integer
Should be OK now.
It looks like you are running the app in RAILS_ENV != 'production'. This is dangerous and not recommended
You are exposing internal error messages to users.
Re: Show HN: Paperkast
#53Re: Show HN: Paperkast
#54There is also https://paperhive.org/ which works really well with comments directly within the document.
Re: Show HN: Paperkast
#55Really great - I signed up. One useful feature might be for the index to show one-sentence blurbs about each paper (in addition to the title of the paper itself). Unless I'm reading papers very specific to my subfield, I need some context to know what I'm looking at & why it's important. Otherwise I fear that only articles with the most accessible titles (or disciplines whose titles tend to be most accessible) will b…
Thank you very much for the feedback. Do you think it has to be in the front page -under each link, perhaps?
That's what I was thinking - maybe something under the link. Could be optional as well.
I'd been remembering how Marginal Revolution, which despite having a fairly educated/technical audience, still often writes its own link text when linking to papers: e.g. item #1 here https://marginalrevolution.com/marginalrevolution/2018/08/sa...
But obviously that would have a big impact on UX for submitters and readers which I haven't thought through in detail!
Re: Show HN: Paperkast
#56Any particular reason for the 100 character title limit?
Re: Show HN: Paperkast
#57Any particular reason for the 100 character title limit?
Fixed to 150 characters. Thank you man!
Re: Show HN: Paperkast
#58Re: Show HN: Paperkast
#59This site is a good idea, and the choice of tag-based navigation is great. As of now, your tags are mostly ad hoc and that will very quickly lead to a jumbled disaster. Already we are seeing people saying that their field's tag is not there.
I strongly recommend that you implement and design a semantic, controlled vocabulary for the subjects in order to keep it from being and hoc. It would be pseudo-hierarchical, but not strictly so. For instance, Physics would include Classical and Modern and both of those would break down into things such as biophysics and quantum chemistry. The question is begged, "But if we put biophysics in the Physics category, it would not be in Biology unless biology were also to be included within Physics." To this, there are two follow-up statements. The first of which is that you could include Biology in Physics. The second of which is the more relevant point in that the beauty of semantic organization of information is that things can belong to multiple groups. Biophysics could be under biology and physics. One doesn't have to choose.
This brings us to modifiers? For each of these, you could have, for instance, a "Quantitative", "Analytical", "Numerical", etc. variant of the discipline. Should you have tags for "Quantitative Physics", "Quantitative Biophysics", and "Quantitative Biology"? I argue, No. Instead, take advantage of semantic structuring of information and simply attach the "Quantitative" descriptor to articles having to use quantitative methods or categorizing themselves as the "Quantitative" variant of the discipline. The end result of this is that you can "build up" to disciplines that we know by using tags, but you aren't limiting yourself nor are you generating duplicate or excessive tags in the long run.
A final question, Is the above overkill? I argue, No. If you are trying to be a hub to discuss any and all academic publications, then your site will eventually expand beyond its initial STEM focus to include the humanities, the arts, and so forth. You will be confronted with the task of trying to essentially categorize all knowledge. I argue that you should just plan for that from day one and take advantage of semantic structuring of information to not only allow discoverability of disciplines we know exist, but to eventually generate insights into combinations of disciplines of which we as a society have not yet conceived.
Re: Show HN: Paperkast
#60One piece of feedback I have is to be careful with tagging. What is physics to a biologist [1] is probably not physics to a physicist. If you want communities to build around this hub, then you need to be careful about flooding a small community with unrelated topics because a larger community is more active on the internet. This looks neat otherwise and I joined. [1] https://paperkast.com/s/mbiq74/maximum_entropy_mo…
Good point. I would say that as the community grows, there will be more and more tags and therefore specialization. Do you think it’s better to stick with broad tags like Physics or Genetics, rather than Synthetic Biology or Quantum Mechanics? I would say the latter would be much better.
Specific tags are good to carve out communities as they really are in academia.
I work in materials science, which there is a tag for, but I work in a super narrow slice of materials (it’s a broad field) and I imagine many others are the same. Even on a list of about 10 journals I keep RSS feeds for, I have a hit rate (of a paper I will read) if no better than 10%. And these are journals that I have already selected as “my flavor” of materials.
Specific tags are bad in that they could harm discoverability. It would be neat if the tags could be hierarchical. E.g. a tag for quantum mechanics shows up in materials science and physics so that the discoverability is there from the general tags, but the narrow slices are available as a filter.