Wikipedia talk:WikiProject National Institutes of Health
This is the talk page for discussing improvements to the WikiProject National Institutes of Health page. |
|
Archives: 1 |
![]() | This project page does not require a rating on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||||||||
|
Suggest spelling out NIH
[edit]May I suggest spelling out NIH for the project and its categories so that it more clearly describes what the project is for? Of course we can leave the NIH as a redirect but I think it would help identify the project to folks who might be interested in participating. Kumioko (talk) 16:19, 4 June 2012 (UTC)
- I hope no one minds but I went ahead and move the project to the spelled out form of NIH. I also moved the user template to template space and spelled it out, fixed it on all the user pages and on the main project page. I also created a Todo list that links to the WikiProject US template as well as some maintenance categories when NIH is added as a parameter. If you want me to add that to the NIH template too let me know and I would be happy too. Kumioko (talk) 18:45, 5 June 2012 (UTC)
"Cite on Wikipedia" tool at National Center for Biotechnology Information website
[edit]I am in contact with the National Center for Biotechnology Information (who run web services like PubMed Central) over them providing references in a way that allows for easy copy-pasting into Wikipedia articles (similar to what Europeana does or the Biomedical citation maker). Where would be the best place to discuss what Wikipedia template formats (e.g. {{Cite web}}, {{Cite journal}}, {{Citation}}, {{Cite book}}) would be best to implement at what NCBI projects? Thanks for any pointers. -- Daniel Mietchen - WiR/OS (talk) 03:45, 18 July 2012 (UTC)
- There are already a number of tools for creating citations from PubMed data that can be copy and pasted into Wikipedia articles. For example, User:Diberri's Wikipedia template filling tool (instructions) takes a PubMed or PubMed Central ID as an argument and produces a fully formatted citation. This tool is widely used by WP:MCB and WP:MED projects. This tool is able to produce {{Cite web}}, {{Cite journal}}, {{Cite book}} templates and in addition {{chembox}} and {{protein}} templates. The ability to create {{drugbox}} templates is broken however. I suppose it would not hurt to have another tool for creating these templates, but Diberri's tool already does a very good job. Boghog (talk) 04:26, 18 July 2012 (UTC)
- One thing that would that would really be helpful is if the NCBI could produce {{Cite journal}} templates for NCBI Bookshelf entries. That would be a unique capability that as far as I know would not be available anywhere else. Boghog (talk) 04:57, 18 July 2012 (UTC)
- I think GeneReviews and some other bookshelf items are providing PMID values. Unfortunately the Diberri template filler doesn't yet cope with the XML returned to generate a cite journal. Arguably a cite book with a chapter reference may be more appropriate, (some attempts on my talk page). The template filler could also be enhanced to process an ncbi_acc like NBK1335 or equivalent URL. cite book is more appropriate for actual book chapters provided too.
- GeneReviews provides Dublin core (DC.) and some ncbi_ prefixed metadata, but not the PMID. Is metadata more generally useful or a can of worms? RDBrown (talk) 07:04, 18 July 2012 (UTC)
- I don't think GeneReviews have PMIDs, do they? Klortho (talk) 00:08, 5 November 2013 (UTC)
- Looking at this closer, I now see that they do. For the GeneReviews link above, the corresponding PMID is 20301510. This PMID is found near the bottom of the GeneReviews page near the copyright information. However when this pmid is feed to Diberri, many of the field in the template are not filled (e.g., {{cite journal | author = Pagon RA, Adam MP, Bird TD, Dolan CR, Fong CT, Stephens K, Dietz HC | title = | journal = | volume = | issue = | pages = | year = | pmid = 20301510 | doi = | url = }}). Strange. Boghog (talk) 03:55, 5 November 2013 (UTC)
- Yes, that is strange. I wonder where Diberri's tool gets the data. You can see the title (but no DOI) here. Klortho (talk) 12:51, 5 November 2013 (UTC)
- Looking at this closer, I now see that they do. For the GeneReviews link above, the corresponding PMID is 20301510. This PMID is found near the bottom of the GeneReviews page near the copyright information. However when this pmid is feed to Diberri, many of the field in the template are not filled (e.g., {{cite journal | author = Pagon RA, Adam MP, Bird TD, Dolan CR, Fong CT, Stephens K, Dietz HC | title = | journal = | volume = | issue = | pages = | year = | pmid = 20301510 | doi = | url = }}). Strange. Boghog (talk) 03:55, 5 November 2013 (UTC)
- I don't think GeneReviews have PMIDs, do they? Klortho (talk) 00:08, 5 November 2013 (UTC)
- Another thing that would be useful is that if when you search for review articles on a certain topic for a certain set of dates the url in the address bar was to link directly to that search (currently it does not change based on your search)
- Wondering if we should consider vcite or fcite do to their speed? Doc James (talk · contribs · email) (if I write on your talk page please reply on mine) 05:17, 18 July 2012 (UTC)
- Just to note, fcite is currently being considered for deletion. In any case, changing cite to fcite or vcite is a fairly trivial modification that can be completed with the "search and replace" option of the advanced wikipedia editing tool bar. Therefore generating fcite or vcite templates doesn't necessarily need to be an option in the template generator. Boghog (talk) 06:09, 18 July 2012 (UTC)
- Of note, Wikipedia's builtin citation tools also already take pmid's and then autofill the details ("Cite" above the edit field and then cite journal, fill in pmid and click the search icon). --WS (talk) 08:22, 18 July 2012 (UTC)
- That is RefToolbar, which supports the major Citation Style 1 templates. ---— Gadget850 (Ed) talk 13:32, 18 July 2012 (UTC)
- Journal articles and book chapters or pages are only part of what NCBI has to offer, so what about the Nucleotide Database, OMIM, Taxonomy and all the other services already frequently cited here? I think having some standard way to cite such sources - and having this way indicated right at the source - could well facilitate work on our end, and it may even lure in as new contributors some of the users from NCBI. -- Daniel Mietchen - WiR/OS (talk) 02:08, 19 July 2012 (UTC)
- Yes, generation of citations to some of those resources would be very useful. For example:
- {{OMIM}}: <ref name="OMIM604715">{{OMIM|604715|Orthostatic Intolerance}}</ref>
- Also it would be very helpful if the NCBI could generate filled out {{Infobox protein}} and {{Infobox nonhuman protein}} templates. A majority of the links in these infoboxes are to the NCBI website. The Diberri template filler also creates protein boxes, but it is partially broken (not all the fields are filled out). Boghog (talk) 05:09, 19 July 2012 (UTC)
- Another advantage of having PubMed directly generate Wikipedia citation templates is that many including myself use PubMed to locate sources and then copy and past the PMID into a citation generation tool. If PubMed could directly generate the citation template, that would save one copy and paste operation. Furthermore Wikipedia relies on volunteers to maintain many of these tools. The NCBI is probably in a better position to maintain these tools long term. Boghog (talk) 06:02, 19 July 2012 (UTC)
- Yes, generation of citations to some of those resources would be very useful. For example:
- Journal articles and book chapters or pages are only part of what NCBI has to offer, so what about the Nucleotide Database, OMIM, Taxonomy and all the other services already frequently cited here? I think having some standard way to cite such sources - and having this way indicated right at the source - could well facilitate work on our end, and it may even lure in as new contributors some of the users from NCBI. -- Daniel Mietchen - WiR/OS (talk) 02:08, 19 July 2012 (UTC)
- That is RefToolbar, which supports the major Citation Style 1 templates. ---— Gadget850 (Ed) talk 13:32, 18 July 2012 (UTC)
- In answer to the original question of where to host the discussion, there is already centralized discussion of those templates at Help talk:Citation Style 1. I am curious about the following:
- Before we the Wikipedia community do outreach to a large distributor of information and ask them to adopt a citation style for us to copy and paste, are we sure that our current citation model is adequate? Is there any need for reform before we make the citation system more difficult to change by entering this external collaboration?
- Is this a bigger issue than just collaboration with NCBI? Do other organizations have a stake in this?
- If there is to be a request from the Wikipedia community to some organization that this organization facilitate the citing of their papers on Wikipedia, is there anything about the negotiation process which ought to be documented so that other organizations can replicate this kind of citation-partnership in the future?
- In response to the notion of reforming citing for easy use amongst many languages and Wiki's there is already such a proposal in Wikicite which makes each citation an object that can be transcluded. Also I think on the French Wikipedia citations are pages which allows them to be transcluded onto many different pages (but only in frWP).
- On the issue of whether people should be able to make Wikitext citation calculators on other sites, I definitely can't see any harm in it. Maximilianklein (talk) 15:21, 18 July 2012 (UTC)
- I recommend they generate standard COinS metadata. Then a reference management system like Zotero can be used to read and store the data, then output it as a Wikipedia citation template. This will allow their references to be used across multiple platforms, not just Wikipedia. If they create a Wikipedia reference generator, then we have to keep ties to maintain it through any changes. ---— Gadget850 (Ed) talk 11:24, 19 July 2012 (UTC)
- Unless I am missing something, if you supply Zotero a PMID, it already is able to import data from PubMed. Furthermore I think for most people, having to go through a third party program just to be able to generate a citation template is unnecessarily complicated. The RefToolbar is far simpler. Boghog (talk) 14:08, 19 July 2012 (UTC)
Updating infoboxes automatically with data from NCBI?
[edit]Following up on the previous section, another point of interest is how NCBI content could be used to update infoboxes (or even article text), similar to what PBB does - what is the best place to discuss such matters? -- Daniel Mietchen - WiR/OS (talk) 04:02, 18 July 2012 (UTC)
- Updating article text would be very tricky. One thing for sure is that we don't want to overwrite any text that was added by human editors. Furthermore we would need to make sure that any added text is in the public domain. When the PBB creates new articles, it includes the Entrez gene summary text. The idea was that the Entrez text would provide seed material for humans to expand on. The original plan was for the PBB to also update the Entrez text, but this was abandoned. The PBB does update the {{GNF Protein box}} templates. Having two bots updating the same set of templates is redundant and could cause unintentional edit wars. I suggest that you contact the PBB maintainers (User:AndrewGNF and User:Pleiotrope) directly about this proposal. Boghog (talk) 04:46, 18 July 2012 (UTC)
- I have left a note to both of them, pointing to one I had left earlier at the Gene Wiki Portal. -- Daniel Mietchen - WiR/OS (talk) 02:22, 19 July 2012 (UTC)
- I have no idea where to discuss this but I want to be part of any discussion which anyone organizes. User:Boghog says this would be tricky but I think there are enough uncontroversial applications that if there is any potential to plan long-term for pulling data and sharing it automatically that I want to be around to support any group who raises such proposals. Blue Rasberry (talk) 14:24, 18 July 2012 (UTC)
- I'm fully in support of updating more than just the {{GNF Protein box}} templates. There are at least a few other template sets I can think of that could benefit from information retrieved from NCBI. However, I don't think updating article text is a very good idea unless the bot is appending things on the bottom, such as "related information" or "further reading", for instance, for the reasons Boghog stated above. For example, PBB adds citation information when it creates article stubs under the "further reading" sections; perhaps any NCBI bot could add similar links to pages that don't have them, or have very few? Finally, if anybody does want to develop a bot on this idea, our code for PBB is available here and should be of some help. Pleiotrope (talk) 17:12, 19 July 2012 (UTC)
NCBI traffic data
[edit]If you have ideas or suggestions as to which data NCBI should make available about traffic they get from Wikimedia servers, please list them below. They will look into making that data available if they have it and it is technically and legally feasible. They are currently redesigning significant parts of their infrastructure and also interested to hear about feature requests, bug reports, existing and potential use cases and such from the Wikimedia community for their services and platforms. Thanks! -- Daniel Mietchen (talk) 21:11, 25 October 2013 (UTC)
- Stats requests
- Stats request – I would be interested in a break down of which NCBI databases are accessed/linked/queried most frequently from Wikipedia. I would imagine by far the most common would be PubMed followed by PubMed Central. But the breakdown for the other databases (Books, OMIM, MeSH, Gene, Protein, HomoloGene, Structure, etc. ) would be interesting. Boghog (talk) 21:20, 25 October 2013 (UTC)
- Stats request – I would like an article level microdata from NCBI on total traffic by day, and sources of that traffic including Wikipedia. So, it would be useful to see how traffic to NCBI pages changes when Wikipedia page incorporate that information, link to them, or when Google algorithms change. Dalek2point3 (talk) 15:31, 26 October 2013 (UTC)
- Stats request – I would be interested to see a top 10 listing of incoming links from Wikipedia projects, including different language-pedias. So listing the most popular subjects linked is interesting (see above), but also, the "most linked Wiki projects". Jane (talk) 08:46, 28 October 2013 (UTC)
- Basic traffic numbers – I want to know how much traffic they get in general so that I can compare their traffic for a few popular topics to Wikipedia's traffic to the same topics. Blue Rasberry (talk) 13:49, 28 October 2013 (UTC)
- Feature requests
- Feature request – Diberri's citation filler tool has been down for some time. I am trying to get it back up, but it would be really nice if the NCBI could provide a similar tool as proposed above. Boghog (talk) 21:51, 25 October 2013 (UTC)
- Rather than a Wikipedia specialized citation format generator, could the NCBI provide an output format for a singleton result from PubMed and other searches that maps the NLM/DTD fields to the Dublin Core and or PRISM metadata. Better journals do this, having the NLM set a precedent could help move institutions like the national libraries, WorldCat and Google Books and Scholar into doing so, allowing the use of more generally applicable tools.
- See Wikipedia:Scripts/Perl scripts/hdump-head.pl for a sketch of metadata => citation parameters RDBrown (talk) 11:43, 27 October 2013 (UTC)
- To play the devil's advocate, what does Dublin Core and PRISM metadata have to do with Wikipedia? Adding an additional layer of metadata (and software) between PubMed and Wikipedia citation templates seems unnecessarily complicated. This metadata may eventually be useful for WikiData, but there are no immediate plans to include general citations in Wikidata. Boghog (talk) 22:24, 1 November 2013 (UTC)
- To clarify, for PubMed and PMC data, the citation tool requests the data as XML then maps the data into appropriate citation template fields. The reason GeneReviews data doesn't work is because it is returned as a PubmedBookArticle. The Bibliographic Ontology which subsumes the PRISM and Dublin Core stuff could be used to provide a mapping to the Document type definition if the NLM hasn't already done so. Mapping citation template parameters to the BIBO resources is Wikipedia's business which should allow easier citation generation from Semantic Web Resource Description Framework resources. Maybe the NIH already provides that. Then maybe we could enhance the citation filler so that Cite Web can instead generate cite journal or book or whatever, by using metadata, or recognizing Pubmed or PMC URLs. Make it mostly a case of copying in a URL and having the template filler behave appropriately. RDBrown (talk) 13:00, 9 November 2013 (UTC)
- The tools described in this document EutilsRDF Web Service: an RDF interface to NCBI Entrez Utilities seems to provide a lot of what is needed. The citation template parameters to COinS stuff may give use enough of a mapping to go back the other way. RDBrown (talk) 21:50, 10 November 2013 (UTC)
- I've been very busy with other tasks the last couple of weeks, but hope to start working on this in earnest this week. I guess the first task is to really scope out what NCBI can provide, on our site, as a direct user affordance. But, I'm curious as to how you happened onto that document - it's about an API of course, which could be used by something like Diberri's tool - i.e., an external tool. But right now that web service is live, but not officially supported in any way. If you are interested in NCBI providing RDF data, by all means let me know about that, or just go ahead and submit a pull-request here, and if it looks good I can pull it in. If there is enough demand, perhaps the interface can become "official". Klortho (talk) 05:46, 13 November 2013 (UTC)
- Somewhat off topic, Diberri's citation filler tool is again available at this link. Nevertheless I think the NCBI is better positioned to provide long term support for this feature than Wikipedia volunteers. Boghog (talk) 22:24, 1 November 2013 (UTC)
- To play the devil's advocate, what does Dublin Core and PRISM metadata have to do with Wikipedia? Adding an additional layer of metadata (and software) between PubMed and Wikipedia citation templates seems unnecessarily complicated. This metadata may eventually be useful for WikiData, but there are no immediate plans to include general citations in Wikidata. Boghog (talk) 22:24, 1 November 2013 (UTC)
- Feature request – PubMed shows images from papers. It would be nice to enable Wikipedian's to copy those with compatible licenses into Wikimedia Commons for reuse in Wikipedia. I don't think it makes sense to do this in bulk, but allowing us to do it on a case by case basis would be good. Alexbateman (talk) 06:23, 26 October 2013 (UTC)
- Feature request – Diberri's citation filler tool has been down for some time. I am trying to get it back up, but it would be really nice if the NCBI could provide a similar tool as proposed above. Boghog (talk) 21:51, 25 October 2013 (UTC)
Editing guide
[edit]![](http://upload.wikimedia.org/wikipedia/commons/thumb/0/08/National_Library_of_Medicine_toolkit_for_Wikipedia_editors.pdf/page1-220px-National_Library_of_Medicine_toolkit_for_Wikipedia_editors.pdf.jpg)
I just shared these slides from the spring 2020 editing event which I also logged in the events section. Blue Rasberry (talk) 18:38, 20 August 2020 (UTC)
Project-independent quality assessments
[edit]Quality assessments by Wikipedia editors rate articles in terms of completeness, organization, prose quality, sourcing, etc. Most wikiprojects follow the general guidelines at Wikipedia:Content assessment, but some have specialized assessment guidelines. A recent Village pump proposal was approved and has been implemented to add a |class=
parameter to {{WikiProject banner shell}}, which can display a general quality assessment for an article, and to let project banner templates "inherit" this assessment.
No action is required if your wikiproject follows the standard assessment approach. Over time, quality assessments will be migrated up to {{WikiProject banner shell}}, and your project banner will automatically "inherit" any changes to the general assessments for the purpose of assigning categories.
However, if your project has decided to "opt out" and follow a non-standard quality assessment approach, all you have to do is modify your wikiproject banner template to pass {{WPBannerMeta}} a new |QUALITY_CRITERIA=custom
parameter. If this is done, changes to the general quality assessment will be ignored, and your project-level assessment will be displayed and used to create categories, as at present. Aymatth2 (talk) 17:19, 12 April 2023 (UTC)