Some small features for consideration

[Noon Silk]

Hi,

  So as mentioned on the github page, I wanted to add one or two things to Scirate (that aren't listed on the dev roadmap page - https://github.com/draftable/scirate3/wiki/Development-Roadmap)

 These are:

  1) Allow sorting list of personal scites by date scited, or date published,

  2) Export to bibtex on article page, (and a link to the direct pdf also),

  3) Create a watchlist of authors where I could see their recent publications. Could potentially be dealt with via custom searches to arXiv, but if the list is sufficiently large, might be better to deal with in a more robust manner.

  I'm also not sure of the best way to export a large bunch of papers to bibtex (along with downloading the pdfs). Probably best is, as the wiki says, create an API that lets me write my own program to do this; i.e. query my list of scites, recognise ones which have already been "processed" (perhaps just by date) and then get the bibtex and pdf link for processing (i.e. downloading and adding to my own local bibtex files).

--
Noon Silk

[Jaiden Mispy]

Those all sound like useful things to me! Potentially for the watchlist you can just expand the existing Twitter-like subscription thingy I currently have for categories such that it applies to authors as well (and maybe make a tab on the home page that shows only authors instead of authors+feeds?).

A proper API is a bit of work but an ad hoc solution for bibtex shouldn't be too hard. Maybe a url like http://scirate.com/api/scites/list?user_id=:id which serializes a sequence of scited papers to JSON, and another like http://scirate.com/1305.3615/bibtex which formats the information we already have in the Paper models as per http://www.crcg.de/arXivToBibTeX/

[Noon Silk]

Okay great, I'll look at adding some of this after exams.

I agree that /bibtex is an easy way of dealing with the situation, but think we could easily include a bib more than they do there. At the moment I use (a variant of) this - https://github.com/adamcw/pyarxiv - to get arxiv bibtex information.

On 24 May 2013 17:27, Mispy <^_^@mispy.me> wrote:

Those all sound like useful things to me! Potentially for the watchlist you can just expand the existing Twitter-like subscription thingy I currently have for categories such that it applies to authors as well (and maybe make a tab on the home page that shows only authors instead of authors+feeds?).

A proper API is a bit of work but an ad hoc solution for bibtex shouldn't be too hard. Maybe a url like http://scirate.com/api/scites/list?user_id=:id which serializes a sequence of scited papers to JSON, and another like http://scirate.com/1305.3615/bibtex which formats the information we already have in the Paper models as per http://www.crcg.de/arXivToBibTeX/ 

--
Noon Silk

[Ben Toner]

On Fri, May 24, 2013 at 12:10 PM, Noon Silk <noon...@gmail.com> wrote:

  I'm also not sure of the best way to export a large bunch of papers to bibtex (along with downloading the pdfs). 

Now that Mendeley is part of Elsevier, maybe we should be more ambitious in this direction... 

[Noon Silk]

In what sense, exactly? A new local reference managing client?

JabRef is what I use currently, and it's pretty much complete as I see it, aside from perhaps having the ability to search within PDFs, and one or two other trivial features ...

For a short period I was thinking that a Android version of it would be great, but for me to use it it'd need to support more "weird" workflows like mine (where I keep all my bibtexs and pdfs and whatnot in source control).

On the other hand, if there was a "Sync to <foo>" style system, where I could somehow link up JabRef and Scirate (and perhaps other such websites, say if Journals started getting involved...), that'd be pretty cool. Which is basically what could be accomplished by a suitable Scirate API + local python tools...

--

Noon Silk

https://sites.google.com/site/noonsilk/

"Every morning when I wake up, I experience an exquisite joy — the joy
of being this signature."

[Ben Toner]

I was thinking more about the discussion and group collaboration features of Mendeley, rather than the local reference management. Anyway, that's for later.

Ben Toner

Founder & CEO, Draftable Pty Ltd

Phone: +61 412 996 288

http://bentoner.com  http://draftable.com 

[Noon Silk]

Oh okay, that makes a lot more sense.

Yeah, you'd of course be aware of say https://www.writelatex.com/ and similar sites (I don't use them, but probably people find them useful); indeed I suppose this is part of Draftables area; so it would make sense to consider building something like that in.

[Noon Silk]

Regarding the watchlist, I agree that probably a more general thing about saving custom searches is a good idea, so will not worry about adding this in general.

As for a proper API, I admit to cheating and writing a hacky scaper to do what I want:
  - https://github.com/silky/scirate3_scraper

It only works in combination with pyarxiv, but suits me fine. For a system for use by other people, would probably be a good idea to have some use cases about how everyone else getes the data into their own reference management systems and then build something that will work for all, if necessary?

On 24 May 2013 17:27, Mispy <^_^@mispy.me> wrote:

Those all sound like useful things to me! Potentially for the watchlist you can just expand the existing Twitter-like subscription thingy I currently have for categories such that it applies to authors as well (and maybe make a tab on the home page that shows only authors instead of authors+feeds?).

A proper API is a bit of work but an ad hoc solution for bibtex shouldn't be too hard. Maybe a url like http://scirate.com/api/scites/list?user_id=:id which serializes a sequence of scited papers to JSON, and another like http://scirate.com/1305.3615/bibtex which formats the information we already have in the Paper models as per http://www.crcg.de/arXivToBibTeX/