[tdwg-tag] RDF/OWL Good Practices Task Group

Steve Baskauf steve.baskauf at vanderbilt.edu
Mon Sep 26 03:34:15 CEST 2011


Thanks for the explanation, Hilmar.  I think I agree with you (at least 
mostly).  I certainly don't want the group to start off getting bogged 
down with disagreements over a domain model.  However, I also don't 
think that considering the implications of differing approaches to 
modeling should be stricken from the charter.  I don't think that it is 
likely that we would be successful in replacing current publishing and 
exchange standards even if that were a desirable thing (and it probably 
isn't in situations where it's working).  However, from my admittedly 
limited point of view, the "successful" data publishing is happening 
mostly in the traditional turf of TDWG: specimen and taxon metadata that 
can be handled by Darwin Core Archives and similar "flat" data transfer 
structures.  It is not happening in more complicated, non-traditional 
circumstances like mark-recapture (e.g. 
http://lists.tdwg.org/pipermail/tdwg-content/2010-December/002200.html), 
bioiblitzes, image metadata of the type I provide, the chains of 
connected resources (specimens, identifications, tissue samples, DNA 
samples like that being tracked by BiSciCol, etc. which could be handled 
very nicely by data exchange in the form of RDF if there were a common 
model.  In my case, using RDF for data transfer is my primary interest 
in the subject, so I guess the importance one places on it depends on 
one's perspective.  As you suggest, this is something for the 
participants in the group to grapple with.  I just don't want it to be 
declared "out of bounds".

Steve

Hilmar Lapp wrote:
> Hi Steve,
>
> I'm mostly in agreement with you, in that I understand your "first 
> floor" (or perhaps more aptly named ground floor) as defining the 
> domain data model, and the second floor as communicating data that 
> instantiate the domain model.
>
> I guess my two principle concerns, where I think it should be left to 
> the participants and not the charter to put their foot down, can be 
> summed up as follows. I think the value propositions of RDF/OWL isn't 
> in enabling us to publish our biodiversity data, compared to what we 
> can do already without RDF/OWL. Therefore, the domain model we want to 
> instantiate in RDF or OWL should as much, if not entirely, be driven 
> by what we want to do with the instance data, rather than by 
> unrealistic expectations of replacing current data publishing and 
> exchange standards with an RDF representation. In that, I'm much with 
> Joel's earlier stressing the "schema last" power of RDF - we can add 
> properties later to the instance data if and when we need them to 
> accomplish something, all we need in the first floor are the 
> weight-carrying walls and a staircase to the 2nd floor, and the pipes 
> and electrical lines we can add when and where we'll need them for the 
> higher floors in your picture. The potential of publishing data in RDF 
> is, IMHO, really in enabling reuse of the data that would not be 
> easily obtainable with the ways we publish data currently. By and 
> large, this will be integrative and aggegrating reuse applications - 
> ones that are seriously hampered by data that, albeit available 
> online, are in silos.
>
> So I think it would be a mistake and missed opportunity if the group 
> started with debating at length the RDF domain model for the sake of 
> getting the domain model "right".
>
> -hilmar
>
> Sent with a tap.
>
> On Sep 25, 2011, at 2:28 PM, Steve Baskauf 
> <steve.baskauf at vanderbilt.edu <mailto:steve.baskauf at vanderbilt.edu>> 
> wrote:
>
>> Hilmar,
>> I've had time to look over the revisions that you made to the 
>> charter.  I think that the changes you made were useful for 
>> clarifying the issues faced by the group and have improved the 
>> document - thanks for that.  The only change that I have questions 
>> about is one between r45 and r46, showing up as line 42 in the source 
>> comparison.  (It's possible that this may be viewable at this URL: 
>> http://code.google.com/p/tdwg-rdf/source/diff?path=/wiki/CharterOfTG.wiki&format=side&r=46 
>> <http://code.google.com/p/tdwg-rdf/source/diff?path=/wiki/CharterOfTG.wiki&format=side&r=46>).  
>> In that revision, you removed the section about "meeting the needs 
>> for standardized data exchange".  My concern to some extent reflects 
>> Donald Hobern's comments at: 
>> http://lists.tdwg.org/pipermail/tdwg-tag/2011-September/002396.html
>> namely that RDF can play an important role as a data-transfer 
>> technology.  
>>
>> In my somewhat simplistic view, the problem the RDF group faces is 
>> something like constructing a building with four floors.  (My 
>> apologies in advance for oversimplification and errors in use of 
>> terms.)  Building the bottom floor involves coming to a consensus 
>> about kinds of things we need to talk about in the biodiversity 
>> informatics world and their basic properties.  (This would involve 
>> examining existing vocabularies and ontologies within and outside of 
>> TDWG and choosing, and if necessary creating classes for the 
>> resources which are of interest to our constituency and at least 
>> suggesting some of the critical datatype properties [sensu OWL; 
>> http://www.w3.org/TR/owl-ref/#Restriction] for those classes.)  The 
>> second floor involves figuring out how we communicate information 
>> about these things and the relationships among them.  (This would be 
>> working out the object properties [sensu OWL] that describe the 
>> relationships among classes and working out the practical details of 
>> how users at one institution can make use of information provided by 
>> or housed at another.)  The third floor involves aggregating 
>> information and preforming queries on the aggregated data.  (This 
>> would be figuring out how to make it possible to create triple stores 
>> and query them with SPARQL; roughly the goals of Linked Data).  The 
>> topmost floor would involve doing more sophisticated reasoning and 
>> making discoveries by drawing inferences about the things we have 
>> discovered (roughly the goals of the Semantic Web).  It seems to me 
>> like we are asking for trouble if we start building the upper floors 
>> without constructing the lower ones first.  We need something to 
>> support the upper floors!  It also seems like we would also be asking 
>> for trouble if we start building the bottom floor without putting 
>> some thought in what we plan for the upper floors to look like. 
>>
>> So my concern about the change is that we don't remove the first and 
>> second floors from the RDF/OWL building.  The core member list of the 
>> proposed group is well populated with people who are comfortable with 
>> living in floors three and four.  But floors one through three are 
>> really the most directly related to the core mission of TDWG (see 
>> http://www.tdwg.org/about-tdwg/).  So I feel strongly that they 
>> deserve a lot of attention, at least in the early part of the TG's work.
>>
>> Steve
>>
>> Hilmar Lapp wrote:
>>> Hi Steve -
>>>
>>> Bob suggested too that I changed the charter into an Interest Group 
>>> charter. So I'm sorry if rather than moving anything forward I 
>>> created mostly confusion. 
>>>
>>> Having said that, the changes I made are a reflection of the context 
>>> and scope of charge in which I think this group, whether it is now 
>>> an IG or TG, should be operating, i.e., that I feel would make the 
>>> most sense. I feel pretty strongly that producing and practically 
>>> validating RDF/OWL data publishing and consumption practices will 
>>> consist of more than one task, and the idea that this is a task we 
>>> can do once and for all is rather concerning to me. In fact, IMHO it 
>>> isn't even worth attempting - the technology landscape in this area 
>>> is evolving so rapidly, anything we produce now is virtually 
>>> guaranteed to be obsolete in one year if no group feels committed to 
>>> maintain it.
>>>
>>> I'll also admit that I'm actually surprised to find that the TAG is 
>>> an IG similar to all others. I would think the TAG ought to be a 
>>> cross-cutting group that integrates the output from all IGs, and has 
>>> no TGs of its own other than those devoted to accomplishing this 
>>> cross-IG integration.
>>>
>>> Finally, if TGs are devoted to accomplishing one task and then they 
>>> dissolve, I don't understand why anyone should be bothered with 
>>> creating, and then approving a charter to start with - shouldn't 
>>> they rather have an agenda (or possibly a proposal preceding that)? 
>>> Either there are people willing to do the task or there are not - I 
>>> don't see the point of the chartering/approval process here.
>>>
>>> So, apparently the TDWG process just confuses the hell out of me. 
>>> And apparently it's really only an IG that would be in line with 
>>> what I think is  the most useful way to do this. Can we still change 
>>> to IG?
>>>
>>> -hilmar
>>>   
>>> On Sep 23, 2011, at 5:03 PM, Steve Baskauf wrote:
>>>
>>>> Hilmar,
>>>> I've been in class all afternoon so I haven't had time to look 
>>>> carefully at your edits yet.  But I wanted to make one comment 
>>>> about what you said in your second paragraph.  It is part of the 
>>>> nature of a task group that it have a limited lifespan: the amount 
>>>> of time that it takes to complete the task which it has been 
>>>> assigned.  After that, the job of maintaining the standard which 
>>>> the task group creates reverts to the interest group which 
>>>> chartered it (I am paraphrasing here from my understanding of 
>>>> http://www.tdwg.org/about-tdwg/process/).  So if the RDF group is 
>>>> actually a Task Group chartered by the TAG, then after its task is 
>>>> completed, it will fall to the TAG to maintain the product that it 
>>>> creates.
>>>>
>>>> The concerns that you raise below include some of the reasons why 
>>>> we had initially suggested that the group be an Interest Group 
>>>> rather than a Task Group.  An interest group does not have a 
>>>> defined lifespan - it exists as long as the interest exists.  
>>>> Unlike a Task Group, it does not have to produce a defined product 
>>>> which http://www.tdwg.org/about-tdwg/process/ implies (but does not 
>>>> explicitly state) would be a standard of one of the flavors 
>>>> described in http://www.tdwg.org/standards/status-and-categories/ 
>>>> (Technical Specification, Applicability Statement, Best Current 
>>>> Practice, or Data Standard). 
>>>>
>>>> The reason why we are currently proposing that the group be a Task 
>>>> Group is primarily because several members of the TAG felt that was 
>>>> the most appropriate thing.  I think that I agree with them.  
>>>> However, I am still uneasy about several aspects of chartering the 
>>>> group as a Task Group, namely:
>>>> 1. I don't really understand exactly who the TAG is (i.e. 
>>>> specifically, who are the particular people to whom the RDF TG 
>>>> would be accountable?).
>>>> 2. What precisely is the task whose completion will signal the end 
>>>> of the life of the Task Group?  We have put some benchmarks in the 
>>>> charter, but none of them include the creation of a standard of any 
>>>> of the forms I listed above.  Is that OK for a Task Group?  I don't 
>>>> know.
>>>>
>>>> I certainly don't want to put a damper on the forward progress of 
>>>> the group by asking these questions, because I'm excited about the 
>>>> prospect of getting the group off the ground and because the TDWG 
>>>> meeting is only weeks away.  But at the moment we are engaging in a 
>>>> discussion within the chartering group and I think it would be 
>>>> appropriate for some of the TAG members to weigh in on these 
>>>> concerns.  If it turns out that there isn't really any answer to 
>>>> the question "who exactly is the TAG?" and "what is our task?" then 
>>>> maybe chartering an Interest Group would be more appropriate than a 
>>>> Task Group.
>>>>
>>>> Steve
>>>>
>>>> Hilmar Lapp wrote:
>>>>> Joel -
>>>>>
>>>>> I've made a number of edits. These are in part to put the motivation  
>>>>> into a larger beyond-TDWG context, and in part to make it a little  
>>>>> more future-proof. The charter in places read (to me) more like a  
>>>>> workshop agenda than a charter, thus preempting decisions that the TG  
>>>>> participants might (want to) make to a degree that I wasn't fully  
>>>>> comfortable with. I've tried to make it take a step back.
>>>>>
>>>>> I also removed the sentence about handing off to the TAG after one  
>>>>> year - while that may be what the participants indeed decide to do  
>>>>> after one year, it's not what I'd want ingrained in the charter, and  
>>>>> also a one-off mindset isn't necessarily what I'd like to start with.  
>>>>> More to the point, if the TG (or whatever its successor(s)) doesn't  
>>>>> maintain those documents, I'm afraid nobody will, and there is plenty  
>>>>> of empirical evidence around the TDWG site to support that.
>>>>>
>>>>> 	-hilmar
>>>>>
>>>>> On Sep 19, 2011, at 3:46 PM, joel sachs wrote:
>>>>>
>>>>>   
>>>>>> Greetings everyone,
>>>>>>
>>>>>> After some back and forth amongst Steve Baskauf, myself, Greg  
>>>>>> Whitbread,
>>>>>> and the executive, we've decided to move forward with an RDF/OWL task
>>>>>> group, convened under the TAG. Our task will be to deliver a document
>>>>>> comprising
>>>>>> i. use cases and competency questions;
>>>>>> ii. well documented examples of addressing those use cases via rdf and
>>>>>> sparql; and
>>>>>> iii. discussion of advantages and disadvantages of the approaches
>>>>>> illustrated by the examples.
>>>>>>
>>>>>> Our draft charter is at
>>>>>> http://code.google.com/p/tdwg-rdf/wiki/CharterOfTG
>>>>>> and we welcome comments, suggestions, and better ideas. One area where
>>>>>> we're still open is the question of whether or not our deliverable  
>>>>>> should
>>>>>> be an official Best Current Practice document [1]. The charter  
>>>>>> reflects
>>>>>> our current feeling that it should not. After we deliver our "book  
>>>>>> of use
>>>>>> cases and examples", options would include being re-chartered by the  
>>>>>> TAG
>>>>>> to produce a best practices document, spinning off as a "Semantic Web
>>>>>> Interest Group", or disbanding (either in triumph or despair).
>>>>>>
>>>>>> When we were planning to convene as an Interest Group, several of you
>>>>>> accepted our invitation to serve as core members, and we hope that
>>>>>> convening as a Task Group does not change your willingness to do so.  
>>>>>> If
>>>>>> you would like to be a core member of the group, and we haven't yet
>>>>>> contacted you, there's a good chance that we will. But don't wait!  
>>>>>> Feel
>>>>>> free to volunteer for core membership. (And recall that you don't  
>>>>>> have to
>>>>>> be a "core member to" contribute.)
>>>>>>
>>>>>> In regards timeline, I'd like to incorporate any feedback we  
>>>>>> receive, and
>>>>>> submit the charter to the executive at the end of this week, in  
>>>>>> hopes of
>>>>>> being chartered by New Orleans.
>>>>>>
>>>>>> Many thanks!
>>>>>> Joel.
>>>>>>
>>>>>> 1. http://www.tdwg.org/standards/status-and-categories/
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> tdwg-tag mailing list
>>>>>> tdwg-tag at lists.tdwg.org <mailto:tdwg-tag at lists.tdwg.org>
>>>>>> http://lists.tdwg.org/mailman/listinfo/tdwg-tag
>>>>>>     
>>>>>
>>>>>   
>>>>
>>>> -- 
>>>> 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) 343-6707
>>>> http://bioimages.vanderbilt.edu
>>>>     
>>>
>>> -- 
>>> ===========================================================
>>> : Hilmar Lapp  -:- Durham, NC -:- informatics.nescent.org 
>>> <http://informatics.nescent.org> :
>>> ===========================================================
>>>
>>>
>>>
>>
>> -- 
>> 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) 343-6707
>> http://bioimages.vanderbilt.edu
>>     

-- 
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) 343-6707
http://bioimages.vanderbilt.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.tdwg.org/pipermail/tdwg-tag/attachments/20110925/177d30b1/attachment-0001.html 


More information about the tdwg-tag mailing list