Understood. I'll try to remain vigilant about those little widgets. This was a subtle one. It really had me questioning whether I understood SHACL at all!
Steve On Tue, Oct 13, 2020 at 3:16 PM Holger Knublauch <hol...@topquadrant.com> wrote: > > Hi Steve, > > the UI needs to decide what kind of input widget it presents based on > what it can find out from the ontology. The case of sh:hasValue is > difficult, because it would need to make some assumptions about what > type of value is most likely needed. In your case this is extra > difficult because the property shape doesn't specify sh:datatype or > sh:class, and even worse, uses a property path. Due to lack of other > info it thus "guesses" you may want to use an xsd:string and presents > you with a string input field. > > Some things are better edited in source code, and I recommend taking a > look at the source code from time to time. The EDG web UI allows > arranging the source code and form panels side by side, making it easier > to spot such mistakes. > > Holger > > > On 10/14/2020 5:52 AM, Steve Ray wrote: > > Irene, > > Thanks for the feedback. I was working on a "double-hop" property > > shape as part of this exercise, and stumbled on what I would > > characterize as a UI bug in TBC. Note in the screenshot that when I > > chose "Add Empty Row" when setting a value for sh:hasValue, the > > datatype shows up as string. > > Autocompletion works as expected (see second screenshot) > > But here's the source code view (see third screenshot). I lost a > > couple of hours wondering why the validation was failing until Ralph > > helpfully pointed out the quotation marks in the source code, which of > > course should not be there. > > > > > > > > Steve > > > > > > On Mon, Oct 12, 2020 at 3:02 PM Irene Polikoff <ir...@topquadrant.com> > > wrote: > >> I am pretty sure this was a question about the most appropriate SHACL > >> modeling pattern as opposed to OWL vs SHACL. > >> > >> Regarding reasoning, TopBraid users typically use SHACL rules for > >> reasoning. SWRL is pretty obsolete. It will never be a standard and is > >> mainly supported by Protege. > >> > >> I was also assuming that PowerMeasurementDevice (if such class is created) > >> could be a subclass of a MeasurementDevice and instances would be the > >> actual devices - may be from different manufacturers, etc. Of course, the > >> decision on where to stop with classes and subclasses and start using > >> instances is also situational - it depends. > >> > >> On Oct 12, 2020, at 3:16 PM, Michael DeBellis <mdebelli...@gmail.com> > >> wrote: > >> > >> > >>> I'm looking for guidance on best practices for modeling the > >>> following kind of pattern: > >>> A PowerMeasurementDevice takes measurements of Power. > >> > >> In my opinion, you haven't said enough about your problem to really give a > >> good answer to the question. You can model this either with OWL or SHACL. > >> Actually, you will probably use OWL either way just to model that a > >> PowerMeasurementDevice is an instance of Device which is probably a > >> subclass of something else and that producesMeasurements is an object > >> property. So the real question is: is it better to use OWL alone or OWL + > >> SHACL? > >> > >> The standard answer to these kinds of questions is usually: "use OWL for > >> reasoning and SHACL for constraints" but of course things like the domain > >> and range of a property (like most things that can be modeled in SHACL) > >> can be modeled either way, as an axiom for reasoning or as a constraint on > >> data. > >> > >> I think what it really comes down to is what kind of data do you have? If > >> you are creating say an expert system ontology from scratch and are also > >> going to populate it with rules in SWRL or some other language then > >> probably OWL alone is better. If you have a lot of existing real world > >> data which may contain values of producesMeasurements that don't have the > >> appropriate values for their domain and/or range then you probably want to > >> use SHACL. Also, how volatile is your data? If there is constantly new > >> data coming in (which may not match the constraint) then SHACL is probably > >> better. If it is relatively static and your data will tend to be well > >> formed then OWL axioms are probably better. > >> > >> As you probably know, the issue with using OWL axioms is the first time > >> you encounter an Individual that doesn't have the proper domain or range > >> then your entire ontology is inconsistent which means it can't be used for > >> any reasoning until the bad data is fixed or removed. With SHACL you just > >> get a report (or some other user defined action) about the bad data but it > >> doesn't make your entire model inconsistent. > >> > >> Michael > >> michaeldebellis.com > >> > >> > >> On Mon, Oct 12, 2020 at 11:39 AM Steve Ray <st...@steveray.com> wrote: > >>> This question is directed at anyone in this group who is fluent in > >>> SHACL. I'm looking for guidance on best practices for modeling the > >>> following kind of pattern: > >>> > >>> A PowerMeasurementDevice takes measurements of Power. > >>> > >>> Is it better practice to put a SHACL constraint on the instances of a > >>> class called PowerMeasurement (that they should be measurements of > >>> Power), where: > >>> > >>> PowerMeasurementDevice producesMeasurements PowerMeasurement . > >>> > >>> > >>> or, should I just define a property shape for PowerMeasurementDevice > >>> such that the property producesMeasurements has the appropriate > >>> constraints? In this case, I don't need to create a special class > >>> called PowerMeasurement. > >>> > >>> Any thoughts one way or the other? I'm leaning toward the explicit > >>> definition of a PowerMeasurement class, but it's more intuition than > >>> principle. > >>> > >>> > >>> Steve > >>> > >>> -- > >>> You received this message because you are subscribed to the Google Groups > >>> "TopBraid Suite Users" group. > >>> To unsubscribe from this group and stop receiving emails from it, send an > >>> email to topbraid-users+unsubscr...@googlegroups.com. > >>> To view this discussion on the web visit > >>> https://groups.google.com/d/msgid/topbraid-users/CAGUep87ZvBUe0urm5V6rxFPJOVpYCzG3Aa0qz7DxFBH-9TR9Wg%40mail.gmail.com. > >> > >> -- > >> You received this message because you are subscribed to the Google Groups > >> "TopBraid Suite Users" group. > >> To unsubscribe from this group and stop receiving emails from it, send an > >> email to topbraid-users+unsubscr...@googlegroups.com. > >> To view this discussion on the web visit > >> https://groups.google.com/d/msgid/topbraid-users/CALGFikcT82t%3DeKh6bnEdiL8hkTCoOv89-R5my0WvfKrztTVibg%40mail.gmail.com. > >> > >> > >> -- > >> You received this message because you are subscribed to the Google Groups > >> "TopBraid Suite Users" group. > >> To unsubscribe from this group and stop receiving emails from it, send an > >> email to topbraid-users+unsubscr...@googlegroups.com. > >> To view this discussion on the web visit > >> https://groups.google.com/d/msgid/topbraid-users/1D78D8DF-B1ED-4122-B936-81D3046176C5%40topquadrant.com. > > -- > You received this message because you are subscribed to the Google Groups > "TopBraid Suite Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to topbraid-users+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/topbraid-users/e1cb2bd5-0763-5aa2-a166-8d52164e22da%40topquadrant.com. -- You received this message because you are subscribed to the Google Groups "TopBraid Suite Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to topbraid-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/topbraid-users/CAGUep87NUgi%2By-DYDYgshUA1s88-2FoHeL7dUVk8L6JmEcU8%3Dg%40mail.gmail.com.
