Wishlist for the PENSOFT API?

[Scott Chamberlain]

Hi all, 

I recently had a chance meeting with the PENSOFT publisher (here).  All their journals are open access, and he says they are very interested in being innovative in the API space, and wants to know what API users would like to see. 

So, I am coming to all of you asking for your feedback.  What would you like to see?

• What would you want out of an API that connects with scientific journals?
  
• Some of the PENSOFT journals publish species descriptions and keys to species groups.  Maybe there's something interesting there that could be provided. 
• I think right now they use the OAI-PMH framework for their API.  Is this ideal, or something else?
• I think they just provide XML now.  I feel like JSON is better, but thoughts?
• Maybe we could get real crazy and ask if one could submit a paper through the API.
• Maybe more detailed altmetrics data?

Please forward on to anyone that has feelings on this topic.

Thanks! Scott

[Ross Mounce]

Well, if we're talking wishlists here...

They publish a fair few phylogenetic studies. e.g.
http://dx.doi.org/10.3897/zookeys.36.306 ,
http://dx.doi.org/10.3897/zookeys.91.709 ,
http://dx.doi.org/10.3897/zookeys.102.1240 and many more...

I can't say I'm particularly familiar with API's but I do know I'd
like ways of automatically identifying, refining selections and
extracting phylogenetic data content (trees AND character matrices)
from papers.

Can an API help with this?

Also, simple(?) things like markup of papers containing fossil taxa,
and what geological / stratigraphic age these taxa are from would also
be nice. e.g. Find me all taxon names published in Pensoft papers
which can be found in the Cretaceous.

Perhaps they could help to develop a distributed open architecture for
sharing phylogenetic data content - different to the centralised
repositories currently on offer (TreeBASE, MorphoBank)? Given they use
XML, NeXML or Phylo-XML might(?) fit well to help achieve this?

That's my ideas anyhow. Apologies if they're a bit over-optimistic or
unachievable.

Best,

Ross

[Scott Chamberlain]

Awesome, thanks Ross.  

The API can help to identify bits of text/data that a user wants.  But it all depends on how well planned and executed the API is.  The PLoS API is pretty well done in my opinion, but for example, the Springer and EOL APIs are not that great.

S