Why won't The Plant List won't let me do this?

In my last post I discussed why I thought the decision of The Plant List to use a restrictive license (CC-BY-NC-ND) was such a poor choice. CC-BY-NC-ND states that

You may not alter, transform, or build upon this work.

To make this point more concrete, I've created this site:

Experiments with The Plant List

to show the kinds of things that The Plant List's choice of license prevents the taxonomic community from doing. As a first step I'm exploring linking the names in the list to the primary scientific literature, as this video demonstrates:

The Plant List from Roderic Page on Vimeo.

For example, we can take a name like Begonia zhengyiana Y.M.Shui, parse the bibliographic citation provided by The Plant List (via IPNI), and locate the actual paper online, in this case it's freely available as a PDF:



Now we can see a drawing of the plant, and instead of simply trusting that the compilers of The Plant List have correctly interpreted this paper, we can see for ourselves. Down the track, we could imagine mining this paper for details about the plant, such as its morphology and geographic distribution. This requires the link to the original literature, which The Plant List lacks.

A good chunk of the recent plant taxonomic literature has DOIs, for example journals such as the Kew Bulletin and Novon. Playing with some scripts I've managed to associate nearly 9000 accepted names with a DOI, and that's by looking at only a few journals. There are lots more DOIs to be found, but because of the way botanical nomenclators record references (see my post Nomenclators + digitised literature = fail) it can be something of a challenge to find them. This task isn't helped by the fairly lax way some publishers enter data in CrossRef (Cambridge University Press I'm looking at you). The other obvious source of digitised literature is, of course, BHL, and that's next on the list of resources to play with.

Experiments with The Plant List is very crude, and I've barely scratched the surface of linking names to primary literature. That said, given that there are exactly zero links between names and digital literature in The Plant List, I'd argue that my site adds value to the data in that The Plant List. And that's my point — by making data available for others to play with, you enable others to add value to that data. By choosing a CC-BY-NC-ND license, The Plant List has killed that possibility.

So, my question for The Plant List is "why did you do that?"

The Plant List: nice data, shame it's not open

The Plant List (http://www.theplantlist.org/) has been released today, complete with glowing press releases. The list includes some 1,040,426 names. I eagerly looked for the Download button, but none is to be found. You can grab download individual search results (say, at family level), but not the whole data set.

OK, so that makes getting the complete data set a little tedious (there are 620 plant families in the data set), but we can still do it without too much hassle (in fact, I've grabbed the complete data set while writing this blog post). Then I see that the data is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) license. Creative Commons is good, right? In this case, not so much. The CC BY-NC-ND license includes the clause:

You may not alter, transform, or build upon this work.

So, you can look but not touch. You can't take this data (properly attributed, or course) and build your own list, for example with references linked to DOIs, or to the Biodiversity Heritage Library (which is, of course, exactly what I plan to do). That's a derivative work, and the creators of the Plant List don't want you to do that. Despite this, the Plant List want us to use the data:

Use of the content (such as the classification, synonymised species checklist, and scientific names) for publications and databases by individuals and organizations for not-for-profit usage is encouraged, on condition that full and precise credit is given to The Plant List and the conditions of the Creative Commons Licence are observed.

Great, but you've pretty much killed that by using BY-NC-ND. Then there's this:

If you wish to use the content on a public portal or webpage you are required to contact The Plant List editors at editors@theplantlist.org to request written permission and to ensure that credits are properly made.

Really? The whole point of Creative Commons is that the permissions are explicit in the license. So, actually I don't need your permission to use the data on a public portal, CC BY-NC-ND gives me permission (but with the crippling limitation that I can't make a derivative work).

So, instead of writing a post congratulating the Royal Botanic Gardens, Kew and Missouri Botanical Garden (MOBOT) for releasing this data, I'm left spluttering in disbelief that they would hamstring its use through such a poor choice of license. Kew and MOBOT could have made the Plant List available as open data using one of the licenses listed on the Open Definition web site, such as putting the data in the public domain (for example, or using a Creative Commons CC0 license). Instead, they've chosen a restrictive license which makes the data closed, effectively killing the possibility for people to build upon the effort they've put into creating the list. Why do biodiversity data providers seem determined to cling to data for dear life, rather than open it up and let people realise its potential?

BHL and OCR

Some quick notes on OCR. Revisiting my DjVu viewer experiments it really struck me how "dirty" the OCR text is. It's readable, but if we were to display the OCR text rather than the images, it would be a little offputting. For example, in the paper A new fat little frog (Leptodactylidae: Eleutherodactylus) from lofty Andean grasslands of southern Ecuador (http://biostor.org/reference/229) there are 15 different variations of the frog genus Eleutherodactylus:

• Eleutherodactylus
• Eleutheroclactylus
• Eleuthewdactyliis
• Eleiitherodactylus
• Eleuthewdactylus
• Eleuthewdactylus
• Eleutherodactyliis
• Eleutherockictylus
• Eleutlierodactylus
• Eleuthewdactyhts
• Eleiithewdactylus
• Eleutherodactyhis
• Eleiithemdactylus
• Eleuthemdactylus
• Eleuthewdactyhis

Of course, this is a recognised problem. Wei et al. Name Matters: Taxonomic Name Recognition (TNR) in Biodiversity Heritage Library (BHL) (hdl:2142/14919) found that 35% of names in BHL OCR contained at least one wrong character. They compared the performance of two taxonomic name finding tools on BHL OCR (uBio's taxonFinder and FAT), neither of which did terribly well. Wei et al. found that different page types can influence the success of these algorithms, and suggested that automatically classifying pages into different categories would improve performance.

Personally, it seems to me that this is not the way forward. It's pretty obvious looking at the versions of "Eleutherodactylus" above that there are recognisable patterns in the OCR errors (e.g., "u" becoming "ii", "ro" becoming "w", etc.). After reading Peter Norvig's elegant little essay How to Write a Spelling Corrector, I suspect the way to improve the finding of taxonomic names is to build a "spelling corrector" for names. Central to this would be building a probabilistic model of the different OCR errors (such as "u" → "ii"), and use that to create a set of candidate taxonomic names the OCR string might actually be (the equivalent of Google's "did you mean", which is the subject of Norvig's essay). I had hoped to avoid doing this by using an existing tool, such as Tony Rees' TAXAMATCH, but it's a website not a service, and it is just too slow.

I've started doing some background reading on the topic of spelling correction and OCR, and I've created a group on Mendeley called OCR - Optical Character Recognition to bring these papers together. I'm also fussing with some simple code to find misspellings of a given taxonomic names in BHL text, use the Needleman–Wunsch sequence alignment algorithm to align those misspellings to the correct name, and then extract the various OCR errors, building a matrix of the probabilities of the various transformations of the original text into OCR text.

One use for this spelling correction would be in an interactive BHL viewer. In addition to showing the taxonomic names that uBio's taxonFinder has located in the text, we could flag strings that could be misspelt taxonomic names (such as "Eleutherockictylus") and provide an easy way for the user to either accept or reject that name. If we are going to invite people to help clean up BHL text, it would be nice to provide hints as to what the correct answer might be.