Working backwards through this thread...
I hadn't read Dima's post until just now, and I see that at least a couple of his points (i.e., #2, #5, #6) apply to exposing the UUIDs externally. However, I think that a simple protocol (such as replacing spaces with "_", and avoiding characters that look the same but are different -- such as the Cyrillic 'a') could go a long way to mitigating those problems.
On the other hand, it really depends on what the identifier is for. The string "Danaus_plexippus_(Linnaeus_1758)" may be more friendly to our eyes, but "A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523" is definitely more friendly to a computer (Dima's points 1, 3 & 4, among others). My feeling is that the push for GUIDs is more about enabling computer-computer conversations, than it is about enabling human-human or human-computer interactions; and therefore we should not get bogged down in the "ugliness" of the identifiers. In the context of electronic data services, the "ugliness" potential of the "Danaus_plexippus_(Linnaeus_1758)" approach to identifiers is far greater than the ugliness potential of "A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523", when it comes to interlinking electronic biodiversity data. It is nothing for a computer to render relevant metadata of the object identified by "A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523" into "Danaus plexippus (Linnaeus_1758)" on a computer screen or piece of paper for human-eyeball consumption. But there are many pitfalls (some noted by Dima) for a computer to unambiguously resolve "Danaus_plexippus_(Linnaeus_1758)" back to a meaningful data object.
I guess my revised point is: GNI (and uBio/NameBank) are essentially the only taxonomic databases out there where a human-friendly persistent/actionable identifier of the sort being discussed is even plausible as an option. It may not even be wise in this context (as per Dima's points), but it *might* be, depending on the need for a human-friendly identifier.
Maybe the simplest thing to do would be to not regard "http://gni.globalnames.org/name_strings/Danaus_plexippus_(Linnaeus_1758)" as an identifier per se, but rather as a protocol for a web service. In other words, if you append a text string to the root URL "http://gni.globalnames.org/name_strings/", GNI would run that text string against its index and return whatever metadata based on a text-string match. This is not mutually exclusive with an "identifier" in the form of "http://gni.globalnames.org/name_strings/A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523", that would less ambiguously resolve a known record in GNI. At this point, the line between "identifier" and "service" gets fuzzy, of course. But the analogy is true in ZooBank:
The persistent "Identifer" looks like this: A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523
One way that this identifier can be represented as an *actionable* identifier is this: urn:lsid:zoobank.org:act:A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523
Another "actionable" form of the identifier might be this: http://zoobank.org/urn:lsid:zoobank.org:act:A9F435E0-8ED7-46DD-BAB4-EA8E5BF4...
or this: http://zoobank.org/A9F435E0-8ED7-46DD-BAB4-EA8E5BF41523
or even this(?): http://lsid.tdwg.org/urn:lsid:zoobank.org:act:A9F435E0-8ED7-46DD-BAB4-EA8E5B...
(all of which work, by the way)
However, the following are examples of what I would think of as *services*: http://www.google.com/search?q=Danaus+plexippus+(Linnaeus+1758) http://lsid.tdwg.org/summary/urn:lsid:zoobank.org:act:A9F435E0-8ED7-46DD-BAB... http://darwin.zoology.gla.ac.uk/~rpage/lsid/tester/?q=urn:lsid:zoobank.org:a...
But really, from the perspective of the end-user, does it matter if it's an identifier or a service? Ultimately, they ask the questions, and the answers appear on their computer screens.
Aloha, Rich
-----Original Message----- From: tdwg-content-bounces@lists.tdwg.org [mailto:tdwg-content- bounces@lists.tdwg.org] On Behalf Of Dmitry Mozzherin Sent: Friday, June 03, 2011 4:34 AM To: David Remsen (GBIF) Cc: tdwg-content@lists.tdwg.org; Dmitry Mozzherin; Orrell, Thomas; Alan J Hampson; Nicolson, David; Gerald Guala Subject: Re: [tdwg-content] ITIS TSNID to uBio NamebankIDs mapping
In my opinion UUIDs have a few advantages over strings --
- It is uuid, so it will work with uuid tools (current and future ones)
- It is less ambiguous -- For example -- what is the difference between Betulа and
Betula for your eyes? (one of them has a Cyrillic 'a') 3. Database wise it is faster to search because it is just a 128bit number, while a name is at least 245 byte varchar -- it makes searching much faster because in relational databases the size of keys directly proportional to the search speed 4. UUID v. 5 (http://en.wikipedia.org/wiki/Universally_unique_identifier) allows to generate UUID algorithmically without looking up a database (no need for network connection) 5. Links like http://gni.globalnames.org/name_strings/Danaus_plexippus_(Linnaeus_1758) might be ambigous -- I can think of several ways I can represent name string part in the url and they will all resolve to the same thing in GNI. 6. Unescaped unicode characters in url containing literal name strings (people will forget to escape them) will depend on an implementation of a url resolver
Saying this links like http://gni.globalnames.org/name_strings/Danaus_plexippus_(Linnaeus_175 8) are definitely attractive and is it good to have them as another way to access a name! My personal preference would be not use them as main identifier because of the reasons 1, 2, 3 and 5.
Dima
On Fri, Jun 3, 2011 at 7:59 AM, David Remsen (GBIF) dremsen@gbif.org wrote:
Why not use the name as the basis for the resolvable identifier instead of a uuid. Isnt there a 1:1 cardinality between the name and the uuid in the GNI? Doesnt that mean that
http://gni.globalnames.org/name_strings/4ef223c4-0c3e-5e84-ace9-
755c34
c601ec and
http://gni.globalnames.org/name_strings/Danaus_plexippus_(Linnaeus_175
are equally unique? The latter is certainly more readable. In those cases where the namestring is a homonym like
http://gni.globalnames.org/name_strings/Oenanthe
couldn't you just return the addresses of the two globally unique forms of the name when you resolve it?
http://gni.globalnames.org/name_strings/Oenanthe_Smith_1899
http://gni.globalnames.org/name_strings/Oenanthe_Jones_1900
Wouldn't those be as globally unique and easier to read and adjust to? Or am I missing something. I always wanted to do that with ubio IDs after a back and forth with Gregor Hagedorn and wished we hadn't exposed those integers.
DR
Hi Steve,
I don't have time to go through this in detail, and I can't speak for the GNI, but I can tell you about how the GNI URI's work at least for now.
A while back Dima Mozzherin and I were looking into how triples etc. might be of use to the GNI.
We needed a way to generate unique URI's for each name.
We wanted to avoid having to keep these in sync and not require everyone to look each ID up through some service.
Dima came up with the following plan. We use the namestring as seed to generate a unique UUID.
Basically this is a shared algorithm which the GNI and TaxonConcept both use. But it could be used by anyone.
You feed the name string to the algorithm and it spits out a UUID. We append then append that to a URI and web service so it is resolvable.
So the name Danaus plexippus (Linnaeus 1758) => 4ef223c4-0c3e-5e84-ace9-755c34c601ec
So if the GNI and and another group have the same namestring they have the same UUID.
People can then can link their data set to the GNI with the following URI
http://gni.globalnames.org/name_strings/4ef223c4-0c3e-5e84-ace9-
755c3
4c601ec
RDF http://gni.globalnames.org/name_strings/4ef223c4-0c3e-5e84-ace9-
755c3
4c601ec.rdf
<http://gni.globalnames.org/name_strings/4ef223c4-0c3e-5e84-ace9-
755c
34c601ec.rdf>If you think of your data set as one table and the GNI as another, this URI serves as the foreign key that connects them together.
Some on the list don't like how these look, but there is a tremendous advantage in not having to worry about syncing two large data sets and determining if a given integer is already in use.
Also Rod Page has written a recently about UUID's. http://iphylo.blogspot.com/2011/05/zoobank-on-couchdb-uuids-replicati on.html
<http://iphylo.blogspot.com/2011/05/zoobank-on-couchdb-uuids-
replicat
ion.html>There may be a way to do something similar with bit.ly like identifiers that are shorter (mCcSp), but I think it the general idea is a good one.
If you recall from my talk at TDWG, I was able to use these to make statements that one namestring was a synonym etc. of another etc.
The algorithm we use is written in Ruby but I could be ported to many different languages since UUIDs are widely supported.
Respectfully,
- Pete
On Thu, Jun 2, 2011 at 11:41 PM, Steven J. Baskauf < steve.baskauf@vanderbilt.edu> wrote:
My email access has been sporadic since this thread developed, so at this point I'll respond to points made in several of the messages.
First, I should note that there has been previous discussion on this list on a similar topic from http://lists.tdwg.org/pipermail/tdwg-content/2011-January/002231.htm lthrough http://lists.tdwg.org/pipermail/tdwg-content/2011-
January/002231.html.
One can review what was said at that time rather quickly by starting on the first linked message and clicking on the "Next Message" link until you get to the end of the range I gave above.
My reason for the request for information that started this thread was that I wanted to link to a URI that would anchor the name portion of a name/sensu pair (TNU or Taxon Concept a la TCS if you prefer) as in this RDF snippet:
tc:nameStringQuercus rubra L.</tc:nameString> <tc:hasName
rdf:about="http://www.ubio.org/authority/metadata.php?lsid=urn:lsid:ubio .org:namebank:448439"
http://www.ubio.org/authority/metadata.php?lsid=urn:lsid:ubio.org:n amebank:448439/>
At this point in the discussion, I'm not actually talking about creating a link to a taxon concept but rather to a taxon name, so some of the issues Pete raised don't apply here (e.g. what's the "right" name for a concept
the question here is simply what's a stable identifier for the name) . In principle, I could probably just provide the name string and be done with it. However, having some degree of faith that Smart, Computer Savvy People might some day be able to use the metadata returned by the URI (or perhaps metadata which they already have in a triple store onsite) to do cool things like knowing that my name is the same as an orthographic variant or that "Quercus rubra L." is basically the same thing as "Quercus rubra", I would like to also provide a functional URI.
As an end -user who isn't very interested in the technical issues involving names, I don't really care what URI I use. I would prefer for it to be widely recognized and for it to "work" (i.e. be resolvable). In the earlier (January) thread, there was discussion about existing identifiers. There were a number of posts, but in particular http://lists.tdwg.org/pipermail/tdwg-content/2011-January/002258.htm l http://lists.tdwg.org/pipermail/tdwg-content/2011-January/002259.htm ldiscussed the relative merits of ITIS and uBio ID numbers. My take-home message from this was that uBio represented the largest single set of names with assigned identifiers (see http://gni.globalnames.org/data_sourcescited in Pete's email) and that uBio metadata provides useful references. Hence my interest in referencing uBio ids as a URI. However, as a practical matter, the organizations that I share images with either want ITIS TSNs (EOL and Morphbank) or just names (Discover Life). Nobody is asking for uBio identifiers or any other identifier.
I found Kevin's comment at http://lists.tdwg.org/pipermail/tdwg-content/2011-May/002486.html very thought-provoking: "My thoughts are that the most likely way this will be solved is by standard market type pressures - ie the best solution/IDs will be used the most and 'float' to the top." I'm not going to make a judgment about what is the "best" solution or ID. But I would say that in "computer" history, being the "best" doesn't necessarily mean that something will be used. Take for example, the FOAF vocabulary. What the heck is Friend of a Friend? I would venture to say that most of the people using the FOAF vocabulary don't know or care. The FOAF vocabulary was the one that people started to use and once that happened, people didn't switch even if there was something better. I'm not familiar with the history of other stuff like YouTube and Craig's List, but I would guess that they weren't necessarily "the best" systems - they were just the one that the most people started using first and once that happened, people didn't switch. I'm using ITIS IDs because they are easy to get and the people I communicate with want them. Whether they are the "best" or "done correctly" doesn't matter to me as much as the fact that that they are widely recognized and stable (and that thus far every name that I've looked for has been in their database).
I think that one reason why this question has been on my mind is that I've been waiting for GNUB (Global Name Use Bank) to come out. I'm not really up on how it is going to work, but my impression is that it was going to be based on the Global Name Index (GNI) which was mentioned in that earlier January thread. At that point, the GNI names didn't have any identifiers that were exposed to the public as permanent GUIDs. I'm assuming that if GNUB refers to GNI names, they will have some kind of identifiers. So if that happens how is the GUID recommendation 8 going to be followed? As Kevin said in http://lists.tdwg.org/pipermail/tdwg-content/2011-June/002499.html "What I take from recommendation 8 of the GUID applicability guide ... is that if you DON'T already have a record in your own database for a taxon name/concept, then reuse an existing one. " What we have here with GNI is a situation where none of the records have identifiers. In my mind, the "best practice" according to recommendation 8 would be for the GNI to reuse existing identifiers where they exist and NOT make up new ones. This is a bit more complicated because the ITIS identifiers (which are in common use) don't have an http URI version that is resolvable, and while the uBio identifiers have a resolvable http URI, it's in the form of a proxied LSID, which I've already complained is very ugly. So I'd like to hear some ideas about how to have "reused" identifiers in the GNI.
One thing that comes to my mind would be to have a "domain name" like "http://purl.org/gni/" http://purl.org/gni/ or "http://purl.org/tn/"http://purl.org/tn/("tn" for "taxon name") and to follow it with a namespace/id combination similar to what is done with lsids. So for example "itis/19408" and "ubio/448439" could be appended, creating http://purl.org/gni/itis/19408and http://purl.org/gni/ubio/448439 for "Quercus rubra L." Both URIs could point to the same RDF and that RDF could indicate that the two identifiers are owl:sameAs . I realize from what Bob Morris has cautioned in the past that there are problems with owl:sameAs when the two things aren't actually the same thing (e.g. if the uBio ID refers to a name string only but the ITIS TSN refers to the name plus an "accepted" status and a relationship to parent taxa). However, if there were an understanding that the GNI only refers to name strings, then one could still refer to http://purl.org/gni/itis/19408 as an identifier for the name string of the thing (whatever it is) that is referred to by an ITIS TSN of 19408. I don't think there would be a problem saying that and the ubio ID were "owl:sameAs". Some kind of solution like this would allow people to easily generate a resolvable URI for a name if they were using ITIS TSNs or uBio IDs. If the name that one wanted to use was so obscure that it was one of the 9.5 million names that uBio has that ITIS doesn't have, then that name would only have the ubio version. I have no idea whether this would be a good idea or not, but I was really cringing to think about 19 million newly minted UUIDs appended to "http://gni.globalnames.org/"http://gni.globalnames.org/and figuring out how to connect those horrid things to the names and ITIS TSNs that I'm already using. I think that I said this before, but using the purl.org domain rather than one like http://gni.globalnames.org/ would in the future allow somebody else to take over management of providing the metadata when the GUIDs
are
resolved without having to deal with issues of who "owns" the domain name.
Steve
Kevin Richards wrote:
Pete,
I’m not trying to say what you are doing is a waste of time/impossible. I actually think RDF + semantics are a good way forward, but this really implies that we need to rely on the semantics and linkages rather than having a SINGLE ID for a taxon name. (which is what I thought Steve was getting at). Each instance of a taxon name can have its own ID and then all these instances are connected via ontology defined semantic links. This seems more appropriate to me than insisting everyone uses the “Global Taxon Name ID X”.
In your example of *Aedes triseriatus* and *Ochlerotatus triseriatus* – these are two different names so they need two different IDs, they may be linked by a single taxon concept, but they are separate names. So which of these now 3 IDs do you expect people to use, and according to what source??
For example if we have a name, eg the Robin, Erithacus rubecula, mentioned in IT IS (TSN : 559964) and also in EOL (www.eol.org/pages/1051567), also in GBIF (http://data.gbif.org/species/21266780), also in avibase ( http://avibase.bsc-eoc.org/species.jsp?avibaseid=C809B2B90399A43D), which ID are you hoping people will use?? Would you put the IT IS ID in your own dataset as the ID for that name – unlikely. Or would it be better to link them up with semantic linkages.
What I take from recommendation 8 of the GUID applicability guide (as Steve puts is "stop making up new identifiers when somebody else already has one for the thing you are talking about”) is that if you DON’T already have a record in your own database for a taxon name/concept, then reuse an existing one. NOT ditch all your current IDs and adopt someone else’s (especially hard considering it is so hard to work out which if the multitude of names ad concept IDs that directly relates to your taxon name).
I am all for limiting the number of IDs for the “same” thing, but in some cases it is more useful to build linkages than force this tight integration of data and IDs. Especially for taxon names and concepts, where it is complex to define if you are even talking about the “same” thing or not.
Kevin
*From:* Peter DeVries [mailto:pete.devries@gmail.compete.devries@gmail.com]
*Sent:* Wednesday, 1 June 2011 12:38 p.m. *To:* Kevin Richards *Cc:* Steve Baskauf; tdwg-content@lists.tdwg.org; Gerald Guala; Nicolson, David; Alan J Hampson; Orrell, Thomas *Subject:* Re: [tdwg-content] ITIS TSNID to uBio NamebankIDs mapping
Hi Kevin,
I forgot one mention some other things that are different about my project.
You can write a simple SPARQL query to get a list of all the TaxonConcept's that have ITIS ids, or all those that have ITIS and NCBI ID's etc.
You can do this on any SPARQL endpoint that hosts the data.
You can download the entire data set and run the queries on your own endpoint.
You can write a script that runs the query and downloads the ITIS numbers and exports them to CSV etc.
- Pete
On Tue, May 31, 2011 at 5:16 PM, Peter DeVries
wrote:
Hi Kevin,
On Tue, May 31, 2011 at 3:27 PM, Kevin Richards < RichardsK@landcareresearch.co.nz> wrote:
This is exactly why this problem still exists and will be very complex to solve - everyone says "we should have a single ID for a specific taxon name, there seems to be several IDs 'out there' that refer to the same taxon name, so Im going to create another ID to link them all up" - yet another ID that no one will particularly want to follow - you would have to get everyone to agree that your combinations/integration of taxon names is the best one and hope everyone follows it - unlikely in this domain.
Isn't this kind of what the The Plant List, and eBird already do?
A difference being that they tie these to a specific name and specific classification.
The Plant list is not really even open so it is difficult to people to adopt it in mass.
For instance, if I manage a herbarium, how do I easily reconcile my species list with the entities represented in the Plant List?
eBird has millions of records which implies that they have been able to convince the observers in the field to adopt their system. You are correct in that there are probably a lot of taxonomists that don't like their list.
It differs from many of the other classifications, but remember the system rewards them for not agreeing. Note the difference between the
microbial
taxonomists and other taxonomists. In the case of the microbial
workers, the system rewards them for solving problems not debating alternatives. Also, if a good idea comes out that will make it easier for the microbiologists to solve the problems they are rewarded for solving, they are less likely to care whose idea it is.
Like the microbiologists, there are lots of biologists that work with species with the goal of addressing some non-taxonomic problem.
They don't really care if the name is *Aedes triseriatus* or *Ochlerotatus triseriatus, *but they do care that the identifier that they connect their data to is stable.
In regards to the issue of market forces,I suspect (but have no knowledge of) that there were probably decisions made in devising these lists that have more to do with appeasing certain personalities that creating best list. With the way this system rewards people it is likely that the "correct" version will float to the top only after that person has passed away. I don't have much faith that the best system will always float to the top, That has a lot to do with the personalities and how the system rewards are setup. Theoretically, it is possible for one strong personality or group to force others to adopt their less than optimal solution - at least this seems to happen in other environments.
Also, there are all sorts of ways that people can use the publication record to rewrite history. Simply cite the review paper that cites the original paper. Or don't cite it at all.
I would have used only the ITIS TSN but if the name changes the ID changes. This isn't "wrong", it just does not solve my problem.
- ITIS also should add the spiders from the World Spider Catalog.
Another issue that I think has inhibited adoption of a common list is that people can't agree on a particular name or a particular classification.
Since you can model a species concept as having many names and many classifications why not do so?
If this idea was originally accepted, I would not have needed to create TaxonConcept.org.
My plan has aways been to get something that works to solve some problems and then let some larger group take it over.
In a sense, I am more like the microbiologists in that I am not being paid to solve this or debate this problem.
I am doing it because I think something like this is needed, and it is an interesting and personally rewarding puzzle.
- Pete
My thoughts are that the most likely way this will be solve is by stnadard market type pressures - ie the best solution/IDs will be used the most and "float" to the top. It is easy to say that the global taxon name data is a mess, but if you think about it 30 years ago taxon name data were very disparate, duplicated, unconnected, many with NO IDs at all. So I beleive we are making progress and that we will continue to do so albeit at a fairly slow rate.
Kevin
"I agree. This was one of the reasons that I setup TaxonConcept the way I did. It attempts to connect both the LOD entities and the foreign key based entities."
Please consider the environment before printing this email Warning: This electronic message together with any attachments is confidential. If you receive it in error: (i) you must not read, use, disclose, copy or retain it; (ii) please contact the sender immediately by reply email and then delete the emails. The views expressed in this email may not be those of Landcare
Research
New Zealand Limited. http://www.landcareresearch.co.nz
--
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 Email: pdevries@wisc.edu TaxonConcept http://www.taxonconcept.org/ & GeoSpecieshttp://about.geospecies.org/ Knowledge Bases A Semantic Web, Linked Open Data http://linkeddata.org/ Project
--
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 Email: pdevries@wisc.edu TaxonConcept http://www.taxonconcept.org/ & GeoSpecieshttp://about.geospecies.org/ Knowledge Bases A Semantic Web, Linked Open Data http://linkeddata.org/ Project
Please consider the environment before printing this email Warning: This electronic message together with any attachments is confidential. If you receive it in error: (i) you must not read, use, disclose, copy or retain it; (ii) please contact the sender immediately by reply email and then delete the emails. The views expressed in this email may not be those of Landcare
Research
New Zealand Limited. http://www.landcareresearch.co.nz
-- Steven J. Baskauf, Ph.D., Senior Lecturer Vanderbilt University Dept. of Biological Sciences
postal mail address: VU Station B 351634 Nashville, TN 37235-1634, U.S.A.
delivery address: 2125 Stevenson Center 1161 21st Ave., S. Nashville, TN 37235
office: 2128 Stevenson Center phone: (615) 343-4582, fax: (615) 343-6707http://bioimages.vanderbilt.edu
--
Pete DeVries Department of Entomology University of Wisconsin - Madison 445 Russell Laboratories 1630 Linden Drive Madison, WI 53706 Email: pdevries@wisc.edu TaxonConcept http://www.taxonconcept.org/ & GeoSpecieshttp://about.geospecies.org/ Knowledge Bases A Semantic Web, Linked Open Data http://linkeddata.org/ Project
tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content
David Remsen, Senior Programme Officer Electronic Catalog of Names of Known Organisms Global Biodiversity Information Facility Secretariat Universitetsparken 15, DK-2100 Copenhagen, Denmark Tel: +45-35321472 Fax: +45-35321480 Skype: dremsen
tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content
tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content