search

openjournals / brief-ideas

The Journal of Brief Ideas
http://beta.briefideas.org
MIT License
93 stars 10 forks source link

Handling citations within Brief Ideas #32

Closed physicsdavid closed 9 years ago

physicsdavid commented 9 years ago

What would you think of having citations within an idea exempt from the 200 word limit and referenced like this [1]. Then you could have a list of references as individual boxes to fill in either with a citation or DOI [2]. It would also be good to make it really easy to cite other brief ideas (and multiple brief ideas in one follow up idea). This would probably require that we have a search capability on the new idea entry page. (I'll talk about that in a separate issue though.)

[1] Reference 1 [2] Reference 2

arfon commented 9 years ago

Quoting from your email @physicsdavid:

Referencing and citations: let's get that as solid as we can. I think one that is perhaps missing in the format is a way to cite other work without needing to use up most of your 200 words just as citations. JOBI ideas won't work in a ghetto--they need to interact with mainstream literature. That means we should have some way to cite references in the work that won't take up with word count. I'm not sure whether we use reference-style markdown syntax or have an extra set of input boxes on the form for references and instruct people to refer to them with a [1], [2], [3], etc. These references should either be JOBI entries or entries in other journals. In particular, we need to be able to reference more than one JOBI idea in another idea.

Do you think we should encourage people to cite DOIs? I'm wondering if we can then grab the formatted entry from the CrossRef API: http://search.crossref.org/help/api

@lnielsen - what do think about trying something along these lines?

physicsdavid commented 9 years ago

@arfon - I'm happy to encourage citation by DOI but can people find DOIs easily for papers in general?

arfon commented 9 years ago

@arfon - I'm happy to encourage citation by DOI but can people find DOIs easily for papers in general?

Yeah, that's a fair point, I don't think they're that easy to discover. I guess I'm looking for 'strong' links to the literature rather than someone just saying 'Smith et al, MNRAS, 2012'.

We can likely make explicit connections between JOBI entries but for the rest of the literature perhaps string references will have to suffice. Do you have a favoured citation style @physicsdavid?

physicsdavid commented 9 years ago

I don't have a favorite citation style but I agree we need to encourage strong references. I'm not sure we even need to recommend a particular citation style - seeing as it changes from field to field, people should use what they are familiar with, but just be consistent within each idea.

lnielsen commented 9 years ago

+1 for strong references, which I would count anything with a resolvable identifier (DOIs, Handles, arXiv ids, ORCIDs, PubMed/PubMed Central ids, ADS bibcodes, urn:nbn's and lastly just plain URLs)

The type of most identifiers can be detected (see https://github.com/inveniosoftware/invenio/blob/pu/invenio/utils/persistentid.py for how I'm doing it in Zenodo). Nearly only plain integers causes trouble (e.g. PubMed).

Advantage of DOIs is that we could use http://www.crosscite.org/citeproc/ to format citations in +700 different styles, but e.g. arXiv papers doesn't have any DOIs, so I hardly think we can enforce it. Likely the best is just adding features like the citation formatting, that will make life easier for the end-user and hence encourage the right behavior.

Btw., the references can be passed to Zenodo in a separate field in the API. Also, if each one has a persistent id, I can properly link it in the metadata I submit when registering a DOI and thus make it available to anyone harvesting DataCite.

arfon commented 9 years ago

The type of most identifiers can be detected (see https://github.com/inveniosoftware/invenio/blob/pu/invenio/utils/persistentid.py for how I'm doing it in Zenodo). Nearly only plain integers causes trouble (e.g. PubMed).

This looks decent. Can you point me to where the parser is in use so I can have a play?

arfon commented 9 years ago

@stuartlynn this is somewhere I'd love your help :pray: . Ideally I'd really like some kind of client-side auto-complete for Journal of Brief Ideas content (see GitHub screengrab below). I would imagine we'd want to try and auto-complete based upon username, idea sha or DOI?

grab

I wonder if we could also do something to help people general references formatted in a particular way?

stuartlynn commented 9 years ago

Cool will take a look today.

— Sent from Mailbox

On Wed, Dec 17, 2014 at 6:36 AM, Arfon Smith notifications@github.com wrote:

@stuartlynn this is somewhere I'd love your help :pray: . Ideally I'd really like some kind of client-side auto-complete for Journal of Brief Ideas content (see GitHub screengrab below). I would imagine we'd want to try and auto-complete based upon username, idea sha or DOI? grab

I wonder if we could also do something to help people general references formatted in a particular way?

Reply to this email directly or view it on GitHub: https://github.com/openjournals/brief-ideas/issues/32#issuecomment-67316277

stuartlynn commented 9 years ago

Ok I have a version of this for users using a standard @ notation. My question is about sha's and Doi's Do we really want that to be the way we reference other ideas? It would make more sense to me to use something like #idea_name. That feels more intuitive, I am not sure if autocomplete makes sense for arbitrary codes that no one is going to remember. Happy to implement Doi's and SHA's but not sure what the value there is.

arfon commented 9 years ago

It would make more sense to me to use something like #idea_name.

Agreed. I think if someone does copy and paste a URL of an idea (or the DOI) we should recognise that (and make the link) but I agree it makes much more sense for people to start typing the name of the idea.

arfon commented 9 years ago

OK https://github.com/openjournals/brief-ideas/pull/43 breaks the 'create an idea based upon this one' functionality. While there is now a data-model to support more flexible citations there's no UI to go with it...

@stuartlynn - does https://github.com/openjournals/brief-ideas/pull/47 include citation helpers?

stuartlynn commented 9 years ago

Yeah if you start typing # it will autocomplete based on other ideas titles. We could change it to use your new system though

— Sent from Mailbox

On Fri, Dec 19, 2014 at 9:27 PM, Arfon Smith notifications@github.com wrote:

OK https://github.com/openjournals/brief-ideas/pull/43 breaks the 'create an idea based upon this one' functionality. While there is now a data-model to support more flexible citations there's no UI to go with it...

@stuartlynn - does https://github.com/openjournals/brief-ideas/pull/47 include citation helpers?

Reply to this email directly or view it on GitHub: https://github.com/openjournals/brief-ideas/issues/32#issuecomment-67723266

arfon commented 9 years ago

OK, we've made a bunch of progress here and I'd like to review where we're at. I've made a quick screencast (below) that shows the current functionality.

citations

To review, we currently support the following:

Questions about the current approach:

We could encourage some kind of formatting such as cite[Paper name](url to paper) but after reading Martin Fenner's post about this I'm inclined to keep them as 'standard' links and do something clever to parse out the links in a post processing step.

Anyway, I think we're in pretty good shape right now.

Thoughts @physicsdavid @stuartlynn @lnielsen ?

physicsdavid commented 9 years ago

I like the idea of keeping things as standard links and parsing out the links as a reference list at the end. However, there is one problem: what do you do for references that aren't online? How do we accommodate them - this could be important as not enough stuff, particularly old archived stuff, which could be very important for brief ideas, is on the open web.

arfon commented 9 years ago

What do you do for references that aren't online?

Good question... and to be honest I'm not sure I have a good answer right now. I feel like there's a whole bunch more we could do here but it feels like a substantial amount of work to do this well.

My vote at this point would be to recognise this as a limitation of this first release and then look to implementing this if/when we get any traction.

What do you think?

physicsdavid commented 9 years ago

Yep, let's just have this as a limitation for now.

arfon commented 9 years ago

Closing this for now. We should revisit later.