[Taxacom] A new way to view taxonomic publications
Donat Agosti
agosti at amnh.org
Mon Jan 4 14:04:49 CST 2021
2021
Out of any reason, just came along this email from June 2013.
Quentin’s idea of starting a project à la human genome to tackle the huge knowledge hidden in the ca 500M printed, daily growing corpus is not forgotten, and the longer the more relevant to complement the global initiatives to sequence, to digitize collection of life one Earth, and image via iNaturalist et al global biodiversity. With GBIF in place we also have an aggregator and access point to this wealth making use of persistent identifiers to link all up.
At Plazi we now have tools in place to highly automatically extract data, especially taxonomic treatments, figures and to an increasing degree materials citations, and to make it findable, accessible, interoperable and reusable through the Biodiversity Literature Repository (BLR). GBIF is an interesting point by using all the liberated data by importing each article as dataset, and then display each treatment on a species page, including one to many treatments (similar to World Flora Online) including the cited figures, a collection of material citations as occurences (which hopefully will be changed to materials citation at some point). In the best case, the entire process from an article to showing up in GBIF and BLR occurs fully automatically, including al the many links.
This is about the last years output: 76,641 #taxontreatment including 6,650 new species described in 2020 liberated; Current total access to 391K treatments, 216K figures in @biolitrepo , from 41K articles. In the Biodiversity Literature Repository we have an increasing amount of FAIRizised published taxonomic data soon including 100K taxonomic treatments as DataCite objects with its own DOI https://zenodo.org/communities/biosyslit/search?page=1&size=20&subtype=taxonomictreatment&subtype=figure , and 32K datasets in @GBIF https://www.gbif.org/publisher/7ce8aef0-9e92-11dc-8738-b8a03c50a862 .
Happy read of the liberated, new described species in 2020: http://tb.plazi.org/GgServer/srsStats/stats?outputFields=doc.uuid+doc.doi+bib.author+bib.title+bib.year+bib.source+tax.name+tax.kingdomEpithet+tax.orderEpithet+tax.familyEpithet+tax.status&groupingFields=doc.uuid+doc.doi+bib.author+bib.title+bib.year+bib.source+tax.name+tax.kingdomEpithet+tax.orderEpithet+tax.familyEpithet+tax.status&orderingFields=tax.kingdomEpithet+tax.orderEpithet+tax.familyEpithet&FP-bib.year=2020&FP-tax.status=%22sp.%20nov.%22&format=HTML
This year will see a substantial increase through the processing of backlogs of journals we now process routinely. Comments and suggestions can be sent to https://github.com/plazi/community/issues
Glad to move ahead and to look forwards what is happening in 2021!
Donat
……………………
Let's give a Rod the +1 for this work.
However, there is an issue here. What Rod does is dealing with big data. That means, you accept an undisclosed amount of errors in your data in the assumptions that this will be corrected. It also has nothing to do with structural mark beyond name strings and probably bibliographic refernces, of which we already know that there are many version of the same around.
The vast numbers Rod joggles with 3.9 million names, and 400,000 articles is great, but it also has the effect that the funders think, that access to content of publication is easy and thus this conversion is of low priority. This happened for example in the GBIC draft.
It is also a question what the 3.9 million names mean? I assume they are all linked to the exact position in the article they come from
Mark up does not operate in a similar big data fashion. One reason why there are only small operation extracting more than just names is, because it is an extremely cumbersome process which tons of human interference.
From: Quentin Groom [mailto:quentin.groom at br.fgov.be]
Sent: Thursday, June 20, 2013 3:52 PM
To: Roderic Page
Cc: Donat Agosti; taxacom at mailman.nhm.ku.edu
Subject: Re: [Taxacom] A new way to view taxonomic publications
Dear Rod, Donat and others,
This is a little off topic, but it seems to me that we need a project on the scale of the Human Genome Project to get accurate transcriptions and markup of all the legacy text. At the moment there are many small scale trials, but to push the cost down and cover the vast corpus of literature we need to scale up. I'm sure that OCR could be improved and that automatic markup is possible, indeed we need a big project to give incentive to innovations on this topic. Perhaps we need to start pushing our funders in this direction.
Rod, what you're doing is useful as it, at very least, shows us what is and might be possible.
Regards
Quentin
Roderic Page wrote:
I guess I'm struggling to see what we're arguing about.
I have no issue with publishing structured documents going forward, and all power to ZooKeys (and sister journals) for pioneering this. The taxonomic literature viewer I built last night is possible because of Pensoft.
However, the vast majority of animal taxa have not been published in this way, and until we have described >10^6 new species in structured documents, this will always be the case. So, what do do about this legacy? How do we tackle that in a way that is scalable? How do we do this across all taxa, not just a few select groups (ants are cool, but there's a LOT more to life than ants).
Furthermore, how do we integrate the taxonomic literature with the broader biodiversity literature (e.g., ecology, genomics, etc.). How do we make taxonomic literature as findable and as accessible? When people publish articles in Nature, why is the taxonomic literature cited not linked in the same way as the other papers? Why is most of the taxonomic literature effectively invisible in the digital age?
These are the things that motivate me to build things like BioNames, which currently has about 3.9 million names, and 400,000 articles (of which about a quarter I've linked to some form of digital identifier). Is it complete, obviously not. Is it a ridiculous thing to attempt, of course it is. Is it useful, I hope so.
Regards
Rod
On 20 Jun 2013, at 10:52, Donat Agosti wrote:
May be Chuck will now – that's how I recapitulate my exchanges along the history of BHL.
I don't say treatment only – we have most of the names already done in ants (see HNS), the literature has been refenced, scanned and linked to the citations well before BHL existed (Smithsonian supported the scanning as their first grant by the Atherton Seidall foundation, parallel to Biologia Centrali Americana and the mosiqto group), so the next logic step is to dig into content. And that is treatment in the content of names (even Linnaeus didn't just supply binomen, but they all are linked to a treatment), and materials citations, which are children of the treatment (and even Linnaeus had them).
May be one should turn it around. You work like crazy to discover articles in a massive body of legacy, and then dig out names. Essentially we in the ant world have that and want more, and we have learned a lesson: This approach is extremely inefficient. So, the questions I am really interested in is: how can we avoid make you work that much? And the only way is to promote the publishing of structured (eg semantically enhanced linked publications), that also include the elements the Linnaeus foresaw in his minimalist Systema: names AND treatments with some structure (description, ecology/behavior, distribution/ citations). And that's why there is taxpub and needs being promoted by good implementations like those of Pensoft.
Donat
From: Roderic Page [mailto:r.page at bio.gla.ac.uk]
Sent: Thursday, June 20, 2013 2:03 PM
To: Donat Agosti
Cc: taxacom at mailman.nhm.ku.edu<mailto:taxacom at mailman.nhm.ku.edu>; 'Lyubomir Penev'
Subject: Re: [Taxacom] A new way to view taxonomic publications
Hi Donat,
Quick comment on "BHL deliberately allows access to single pages, because their scientists wanted to be able to link directly to the page of the treatment, protologue"
BHL allows access to pages because that's the physical unit they scan, and the obvious thing to expose on their web page. They also didn't have an easy way of locating articles (which has been one of the biggest complaints about BHL). I seriously doubt it was a decision to enable people to link to protologues. It's the standard way you expose scanned literature.
I get that linking to the actual treatment/nomenclatural event is desirable, all I'm arguing is that article-level linking is more tractable, and enables a bunch of things that are bigger than simply providing access to taxonomic names. The prize is much bigger than that.
Regards
Rod
On 20 Jun 2013, at 10:15, Donat Agosti wrote:
From: Roderic Page [mailto:r.page at bio.gla.ac.uk]
Sent: Thursday, June 20, 2013 12:34 PM
To: Donat Agosti
Cc: taxacom at mailman.nhm.ku.edu<mailto:taxacom at mailman.nhm.ku.edu> com; Lyubomir Penev
Subject: Re: [Taxacom] A new way to view taxonomic publications
Hi Donat,
In the demo you skip one of the nice elements, the treatment element. And in my view this is a loss of quality, which you can easily see, when you start looking up one of the new descriptions, eg . Paedophryne dekot. If you look it up on the right hand in Biostor, you finally end up up at the article level and the treatment itself, which is the relevant information that I need, not the publication. Similar the pdf that you show in biostor has lost all the links that are in the original pdf.
It would be straightforward to list treatments in the "Contents" section. The articles discovered in the reference are typically in PDF or image form, but obviously the next step is to display them in the same way, where possible.
DA: this is the huge problem: How to get them out of the pdf (text born, or images) – this right now does not scale to your 10^6. And it might be also questionable whether this underlies the big data paradigm, since there are not plenty of treatment covering the same but rather singletons, and thus errors in the conversion play a different role?
I know, you are not interested in the treatment but names and publications. I as a taxonomist in the world you describe, a publication being a database, which happens when you convert it into XML, and even more so an XML that is domain specific, don't want to be stuck in the world or articles, just because we grew up with it. The goal is customized information, and that means treatment (in the context of a scientific publication). And with Zookeys at hand you have all to make this happen.
Finally, most of the functionality is already in the Zookeys article itself when you look at it in the html version instead of the pdf your link to (http://www.pensoft.net/journals/zookeys/article/1963/at-the-lower-size-limit-for-tetrapods-two-new-species-of-the-miniaturized-frog-genus-paedophryne-anura-microhylidae-) , and even more so in their species profiles http://ptp.pensoft.eu/external_details.php?type=1&query=Paedophryne . The interesting step then is to get this done for other journals and see, how it looks like. There, the production of clean OCR or text extraction from PDF, the semantic mark-up is not done, that is served on a silver platter
by Pensoft, with a pretty overhead in the production of the article.
Two points, one minor, one not so. The minor point is that, yes Pensoft displays marked-up HTML, but it is surrounded by lots of publish-specific stuff. This is one reason why PDFs are still so popular, publishers can't resist surrounding HTML with junk (logos, links, etc.).
The second point is that Pensoft has a platform for their journals. Great, but I want platform that is publisher-agnostic. I don't care who publishes the stuff, I want a consistent way to explore it. I chose ZooKeys because yes, it is essentially pre-processed, so I can focus on just the rendering. But I want PLoS, BMC, SciElo journals looking like this, I want Zootaxa to look like this, etc.
DA: I agree, we want to have an independent site, but we also need to have a business model behind (University pays, EU-pays, somebody pays). The consistent way is clearly the goal.
Then this needs to be compared to other similar sites, like Species-ID http://species-id.net/wiki/Paedophryne_dekot , that allows the crowd to edit and add content or Plazi, the treatment repository http://tinyurl.com/nku2rdd , which allows to get back to the treatment and not end up in the article.
So, the challenge has to be that Bionames does not loose granularity, and operates at the level we cite (not article but a page within an article, essentially linking to the treatment), and to show mechanisms to read in pdfs that have no XML, not text in fact, and make this fly, something you have demonstrated to find articles within the BHL body. This time just one level down.
I realise you want treatments, and I'm not saying I don't, but I have to choose the level of granularity that scales across 10^6 names and 10^5 publications. I need an infrastructure that enables me to make links between names and publications, and that means working at the level of articles. This is also where citation networks operate, and where links to other kinds of data operate (e.g., links between sequences, phylogenies, and publications). The reality is the publication is the fundamental unit we keep track of.
DA: I would argue, that this is an artifact. BHL deliberately allows access to single pages, because their scientists wanted to be able to link directly to the page of the treatment, protologue. All the references in HNS are linked to the particular page, not the article. Yes we have to cite, and we measure citations, but this is just a historical legacy, similar that BHL works on journals (and not articles, that what you are digging out with a lot of pain) is because of constraints outside the taxonomists control (the libraries have them in the stacks this way). All the microcitations point to pages, not articles.
Now, there are interesting developments in making publications more granular. PLoS has had DOIs for figures for some time, and BMC figures are retrospectively having DOIs assigned by figshare (e.g., http://dx.doi.org/10.6084/m9.figshare.34256 ).
I would argue that if you are serious about treatments being citable and discoverable, you'd give them DOIs, so that, for example, the treatments within a ZooKeys article would have their own DOIs. It's time we started playing the bigger game. It's not about treatments per see, it's about linking citable entities.
DA: We are serious, just not as fast as you. We also decided for the time being to make use of stable http URI for identifiers instead of DOIs, not least because the goal is getting this content into the semantic web. This follows the venue CETAF is going by using this system for their specimen data (see also the discussion on stable identifiershttps://plus.google.com/u/1/117201190352607228695/posts/RXEwpGWu18o)
Cheers
Donat
Regards
Rod
Cheers
Donat
-----Original Message-----
From: taxacom-bounces at mailman.nhm.ku.edu<mailto:taxacom-bounces at mailman.nhm.ku.edu> [mailto:taxacom-bounces at mailman.nhm.ku.edu] On Behalf Of Roderic Page
Sent: Thursday, June 20, 2013 12:50 AM
To: taxacom at mailman.nhm.ku.edu<mailto:taxacom at mailman.nhm.ku.edu> com
Subject: [Taxacom] A new way to view taxonomic publications
I've developed a somewhat experimental viewer for articles in the journal ZooKeys which might be of interest. There is a blog post here http://iphylo.blogspot.co.uk/2013/06/a-new-way-to-view-taxonomic-publications.html
You can try a live example here: http://bionames.org/labs/zookeys-viewer/?doi=10.3897/zookeys.154.1963
This viewer is one of the motivations behind http://bionames.org I'm aiming for a platform where we can embed the taxonomic literature and have names and publications seamlessly linked together, enabling us to navigate through the primary taxonomic literature in a single place.
Regards
Rod
---------------------------------------------------------
Roderic Page
Professor of Taxonomy
Institute of Biodiversity, Animal Health and Comparative Medicine College of Medical, Veterinary and Life Sciences Graham Kerr Building University of Glasgow Glasgow G12 8QQ, UK
Email: r.page at bio.gla.ac.uk<mailto:r.page at bio.gla.ac.uk>
Tel: +44 141 330 4778
Fax: +44 141 330 2792
Skype: rdmpage
Facebook: http://www.facebook.com/rdmpage
Twitter: http://twitter.com/rdmpage
Blog: http://iphylo.blogspot.com
Home page: http://taxonomy.zoology.gla.ac.uk/rod/rod.html
Wikipedia: http://en.wikipedia.org/wiki/Roderic_D._M._Page
Citations: http://scholar.google.co.uk/citations?hl=en&user=4Z5WABAAAAAJ
ORCID id: http://orcid.org/0000-0002-7101-9767
_______________________________________________
Taxacom Mailing List
Taxacom at mailman.nhm.ku.edu<mailto:Taxacom at mailman.nhm.ku.edu>
http://mailman.nhm.ku.edu/mailman/listinfo/taxacom
The Taxacom Archive back to 1992 may be searched with either of these methods:
(1) by visiting http://taxacom.markmail.org
(2) a Google search specified as: site:mailman.nhm.ku.edu/pipermail/taxacom your search terms here
Celebrating 26 years of Taxacom in 2013.
---------------------------------------------------------
Roderic Page
Professor of Taxonomy
Institute of Biodiversity, Animal Health and Comparative Medicine
College of Medical, Veterinary and Life Sciences
Graham Kerr Building
University of Glasgow
Glasgow G12 8QQ, UK
Email: r.page at bio.gla.ac.uk<mailto:r.page at bio.gla.ac.uk>
Tel: +44 141 330 4778
Fax: +44 141 330 2792
Skype: rdmpage
Facebook: http://www.facebook.com/rdmpage
Twitter: http://twitter.com/rdmpage
Blog: http://iphylo.blogspot.com
Home page: http://taxonomy.zoology.gla.ac.uk/rod/rod.html
Wikipedia: http://en.wikipedia.org/wiki/Roderic_D._M._Page
Citations: http://scholar.google.co.uk/citations?hl=en&user=4Z5WABAAAAAJ
ORCID id: http://orcid.org/0000-0002-7101-9767
---------------------------------------------------------
Roderic Page
Professor of Taxonomy
Institute of Biodiversity, Animal Health and Comparative Medicine
College of Medical, Veterinary and Life Sciences
Graham Kerr Building
University of Glasgow
Glasgow G12 8QQ, UK
Email: r.page at bio.gla.ac.uk<mailto:r.page at bio.gla.ac.uk>
Tel: +44 141 330 4778
Fax: +44 141 330 2792
Skype: rdmpage
Facebook: http://www.facebook.com/rdmpage
Twitter: http://twitter.com/rdmpage
Blog: http://iphylo.blogspot.com
Home page: http://taxonomy.zoology.gla.ac.uk/rod/rod.html
Wikipedia: http://en.wikipedia.org/wiki/Roderic_D._M._Page
Citations: http://scholar.google.co.uk/citations?hl=en&user=4Z5WABAAAAAJ
ORCID id: http://orcid.org/0000-0002-7101-9767
---------------------------------------------------------
Roderic Page
Professor of Taxonomy
Institute of Biodiversity, Animal Health and Comparative Medicine
College of Medical, Veterinary and Life Sciences
Graham Kerr Building
University of Glasgow
Glasgow G12 8QQ, UK
Email: r.page at bio.gla.ac.uk<mailto:r.page at bio.gla.ac.uk>
Tel: +44 141 330 4778
Fax: +44 141 330 2792
Skype: rdmpage
Facebook: http://www.facebook.com/rdmpage
Twitter: http://twitter.com/rdmpage
Blog: http://iphylo.blogspot.com
Home page: http://taxonomy.zoology.gla.ac.uk/rod/rod.html
Wikipedia: http://en.wikipedia.org/wiki/Roderic_D._M._Page
Citations: http://scholar.google.co.uk/citations?hl=en&user=4Z5WABAAAAAJ
ORCID id: http://orcid.org/0000-0002-7101-9767
_______________________________________________
Taxacom Mailing List
Taxacom at mailman.nhm.ku.edu<mailto:Taxacom at mailman.nhm.ku.edu>
http://mailman.nhm.ku.edu/mailman/listinfo/taxacom
The Taxacom Archive back to 1992 may be searched with either of these methods:
(1) by visiting http://taxacom.markmail.org
(2) a Google search specified as: site:mailman.nhm.ku.edu/pipermail/taxacom your search terms here
Celebrating 26 years of Taxacom in 2013.
--
Dr. Quentin Groom
(Botany and Information Technology)
National Botanic Garden of Belgium
Domein van Bouchout
B-1860 Meise
Belgium
ORCID: 0000-0002-0596-5376
Landline; +32 (0) 226 009 20 ext. 364
FAX: +32 (0) 226 009 45
E-mail: quentin.groom at br.fgov.be<mailto:quentin.groom at br.fgov.be>
Skype name: qgroom
Website: www.botanicgarden.be<http://www.botanicgarden.be>
More information about the Taxacom
mailing list