Sorry I know I'm late to this thread, but just wanted to mention the oaDOI API, which could be helpful depending on your goals.
We can easily handle 225k DOIs if you spread them out over a few hours...we currently serve 500k DOIs daily, and have handled 2M daily in the past with no trouble. Response time is about 200ms at that load. We also have a data dump so you could store it all locally and make it as fast as you like, if you want to put the resources into that. We currently have records for all 90M DOIs, and OA locations for ~10M of them. We're in the middle of a large rolling update (v2) that will add links to another 7M or so hybrid DOIs. ResearchGate and Academia.edu resources are not included in the index, which is either a feature or a bug depending on your goals :) FWIW, we now have an evaluation set <https://docs.google.com/spreadsheets/d/1kC9A9WscsPr-N3r8ULZYjTZdJb3NdhrbMtGlJSHJn-M/edit#gid=263352758> comparing oaDOI v1, oaDOI v2, the Dissemin API, and the Open Access Button API. oaDOI v2 is currently putting up the best numbers, particularly in precision (low false positives) There are some qualifiers, though. The evaluation set is a random sample of all DOIs, across all years. We count RG as closed. The gold standard is manual coding based on Google and Google Scholar searches, which was done independently (by Lisa Matthias and Juan Alperin, coauthors on our recent paper <https://peerj.com/preprints/3119v1/>). So feel free to interpret accordingly :). But we think it might be helpful, particularly since it's quite time-consuming to create the manual gold standard of availability (and without this there tend to be a lot of false positives). Feel free to ignore if it's not what you're looking for. Happy to help (or not) any way we can! Best, J On Thu, Aug 31, 2017 at 7:42 AM, Rudy Patard <[email protected]> wrote: > Hi, > > I'm glad to see this initiative and the momentum it is getting. > I know automation is needed and will provide scale. However, I have a > feeling we are missing an ingredient, something (not else) but more. > > Freeing academically produced literature is not just about pushing > formerly published work toward open archives. > A lot of work, resources and money is spent to sent publicly funded work > in the private lucrative repositories of a very limited number of > publishers. In France, Elsevier receive from Couperin-Abes (the organism > 'negotiating' and buying access to journals for most of french research > institutions) about as much money as the world wide Wikimedian budget. As a > whole, France could pay a double wikimedia if giving-up all what is know > from these subscriptions. Furthermore from the research I maid and lived, > academia in its vast majority simply do not care. It is part of the > economics of academia (some notoriety-academical-kapitalism). One aiming > at a carrier in research can hardly escape this system. And from first hand > experience I'd rather discourage writing about it without a secured > academical position. > So civil society must push a bit to get back what she made possible. In my > opinion Wikimedia is just the right intermed body. > > What is done to make contact with researchers showing them, repeatedly, > who is financially supporting their work and that *public *research can > directly be produced under free (libre) licenses ? Going toward opening > knowledge communities more than opening past research would benefit more > the interaction between researchers and the rest of society. > > As discussed with Dario and previously on some wikimedian discussion on > this matter (can't find the URL), I was proposing to enable emailing > corresponding authors via the mailing tool of a * wikimedian account *(in > use for notifications). And I mean a person-account (not *just *bots*)*. > Pre- established mails would be sent with links such as the ' Email this > user <https://en.wikiversity.org/wiki/Special:EmailUser/RP87> ', but > addressed to corresponding authors found in the meta-data or head of the > article. > > For instance: > ~To {{U|the user mailing the request}} Please, consider personalizing this > e-mail template ~ > "Dear 'Auto- Author-Name', > Searching the references on 'Auto- Matter-Topic', indicated on the article > 'Auto- Wikimedian-article', I found your paper entitled 'Auto- > Article-Title'. I was impeded in this work by paywalls. > Would you please publish this work in an open archive [Auto- major > archives links]. Their are by the way alternative way to publish scientific > work, direct open access, without author publishing charges : 'Link toward > wikiversities scientific journals'. > > Nice formulations about knowledge creation and dissemination etc. > Sincerely yours, > {{U|Auto- the user}} and the wikimedian community." > > The contact links would be placed in the reference part or in a template > at the head of the article : > "This article *does not cite > <https://en.wikipedia.org/wiki/Wikipedia:Citing_sources> enough Open Access > sources <https://en.wikipedia.org/wiki/Wikipedia:Verifiability>* > Contact corresponding authors, in order to free this knowledge and related > references. > Click and send, from your wikimedian account, requests to theses authors > to free their work 'Auto-e-mail-LINK'. > * article 1 > * article 2 > ... > If you are working on this topic, please join an open-peer-reviewing group > [[open-peer-reviewing group portal]]' > If you are from any part of civil society interested in the topic and > searching for advanced knowledge on the topic, please join a > [[vulgarization group]] for not-understood content, or a [[bibliographic > intelligence and problematization group]] for 'non-researched yet' material. > " > > I'll watch for the visio-chat opportunity (some mumble equivalent at least > if we are many). But the current period is quite loaded, with social, > political and work context. So I may just contribute on any meta / wiki > page you point toward. I push this mail toward some friends, former > colleagues and contacts, as I'm sure some will see opportunities. > > Best Regards > > @+ > Rudy {{U|RP87}} > > On 29 August 2017 at 19:57, Joseph McArthur <[email protected]> wrote: > >> Thanks for the quick reply, tried the same below. >> >> On Tue, 29 Aug 2017 at 17:31 Federico Leva (Nemo) <[email protected]> >> wrote: >> >>> Joseph McArthur, 29/08/2017 18:52: >>> > Federico, this sounds amazing! >>> >>> Thanks. >>> >>> > >>> > I want to offer the help of the Open Access Button request system >>> here: [...] >>> > >>> > I'd love to work with you to figure out how we can help here, as it >>> > would be a shame for us to duplicate this! >>> >>> Sure. In fact I reused some of your learnings, so you already >>> contributed to my campaign. >>> https://github.com/OAButton/discussion/issues?utf8=%E2%9C%93 >>> &q=commenter%3Anemobis%20 >> >> >> Oh grand, I'll take a peak at those issues in the coming days and see if >> there is anything else useful I could add. >> >> >>> >>> >>> From https://github.com/OAButton/discussion/issues/587 I got the >>> impression you wouldn't be able to sustain my requests for 224k DOIs at >>> once, so I figured I'd try to add some additional capacity with WMIT >>> resources at least for this initial test of mine. In the short term >>> (September), if some of you have spare cycles, I could use your help in >>> drafting the next emails and replying to support requests. >>> >> >> From issues it sounded like you were doing a few thousand emails to >> authors, which we could manage with a bit of co-ordination (technically, >> you could do it via the API without consulting but it would be quite >> innefficient for us) e.g let us know what you want requested in a >> spreadsheet, and then we'll get it up the same day and have messages sent. >> If you simply wanted to run 224k DOI's to see if they were available, >> that's a pretty light lift, so I wouldn't worry about that (just hit the >> API when you fancy). >> >> We might have some spare time, shall we try and have a 30 minute chat by >> phone (or similar) and try to hammer something out? If that makes sense see >> if there a time that works for you here: doodle.com/joemcarthur. >> >> >>> >>> In the long term, if we want something stable as suggested by Dario (or >>> indeed John >>> <https://www.linkedin.com/pulse/accelerating-open-access- >>> adoption-john-dove>), >>> I agree we should go through some permanent help desk like the OAbutton, >>> ideally with some resources/support from SPARC, JISC, OpenAIRE or >>> whoever is interested in directly supporting green OA beyond specific >>> borders. >>> >> >> All the more reason to chat :) >> >> >>> >>> Nemo >>> >>> P.s: I've removed from Cc some addresses which I believe are already in >>> the list, to avoid making these messages always go into moderation. >>> >> >> Thanks! >> -- >> Joseph McArthur >> Assistant Director: Right to Research Coalition >> <http://righttoresearch.org/> >> Co-Lead/Founder: Open Access Button <http://openaccessbutton.org/> >> Twitter: @Mcarthur_Joe <https://twitter.com/Mcarthur_Joe> >> Skype: joseph_mcarthur >> >> _______________________________________________ >> OpenAccess mailing list >> [email protected] >> https://lists.wikimedia.org/mailman/listinfo/openaccess >> >> > -- Jason Priem, co-founder Impactstory <http://impactstory.org/>: Share the full story of your research impact follow at @jasonpriem <http://twitter.com/jasonpriem> and @impactstory <http://twitter.com/impactstory>
_______________________________________________ OpenAccess mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/openaccess
