Looping in Phoebe Ayers, now at MIT Libraries, who has been getting a few questions from faculty about this email.
Cheers, Jake On Tue, Sep 5, 2017 at 6:54 AM Jason Priem <[email protected]> wrote: > 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 >
_______________________________________________ OpenAccess mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/openaccess
