From christoph.lange at uni-bremen.de Wed Jan 18 19:38:15 2012 From: christoph.lange at uni-bremen.de (Christoph LANGE) Date: Wed, 18 Jan 2012 19:38:15 +0100 Subject: [Hets-devel] Minor revisions to Hets for CL users In-Reply-To: <4F167F0D.6090508@uni-bremen.de> References: <4F167F0D.6090508@uni-bremen.de> Message-ID: <4F171197.20500@uni-bremen.de> Now that I'm subscribed to the list, let me try again? On 2012-01-18 09:13, Christoph LANGE wrote: > Dear all, > > I made some minor revisions to the "Hets for CL users" document and felt > free to append myself as an author. > > I made it half way through the document (up to section 6) and will > continue today or tomorrow. > > I have some open questions (which I put into ednotes), and I'm sure that > you also had some things on your mind wrt. revising the document. > > What is the next step? > > Cheers, > > Christoph -- Christoph Lange, Universit?t Bremen http://kwarc.info/clange, Skype duke4701 My up-to-date presentations: http://www.slideshare.net/langec From till at informatik.uni-bremen.de Wed Jan 18 19:57:10 2012 From: till at informatik.uni-bremen.de (Till Mossakowski) Date: Wed, 18 Jan 2012 19:57:10 +0100 Subject: [Hets-devel] Minor revisions to Hets for CL users In-Reply-To: <4F171197.20500@uni-bremen.de> References: <4F167F0D.6090508@uni-bremen.de> <4F171197.20500@uni-bremen.de> Message-ID: <4F171606.6030504@informatik.uni-bremen.de> Hi Christoph, I have added an intro, a more compact logic graph, a better description of SoftFOL, short descriptions of more logics, and resolved your ednotes. Please have a look; once your are finished, we can send this out to the Common Logic mailing list. Best, Till Am 18.01.2012 19:38, schrieb Christoph LANGE: > Now that I'm subscribed to the list, let me try again? > > On 2012-01-18 09:13, Christoph LANGE wrote: >> Dear all, >> >> I made some minor revisions to the "Hets for CL users" document and felt >> free to append myself as an author. >> >> I made it half way through the document (up to section 6) and will >> continue today or tomorrow. >> >> I have some open questions (which I put into ednotes), and I'm sure that >> you also had some things on your mind wrt. revising the document. >> >> What is the next step? >> >> Cheers, >> >> Christoph > From christoph.lange at uni-bremen.de Fri Jan 20 17:50:14 2012 From: christoph.lange at uni-bremen.de (Christoph LANGE) Date: Fri, 20 Jan 2012 17:50:14 +0100 Subject: [Hets-devel] How do I reuse an OWL ontology (in an RDF/XML file) in a heterogeneous spec? Message-ID: <4F199B46.6090904@uni-bremen.de> Hi all, for the "Hets for CL users guide" I would like to work out that subset of my heterogeneous/distributed AAL ontology use case that is supported by Hets. My current attempt can be found in Hets-lib/Ontology/Examples/AAL.het. I would like to reuse the OpenAAL OWL ontology, which I have downloaded as an RDF/XML file OpenAALOntology.owl in the same directory. * What do I have to write in my heterogeneous specification, in order to be able to import that OWL ontology? * Or do I have to somehow pass it to Hets as an additional command-line argument? (BTW, I will be offline until shortly before midnight, but by then I will at least have written something about the current state in the guide document.) Another problem, which I will deal with separately, is that the OWL Manchester parser does not yet support some constructs that I'm using. I think I will file tickets about these: * that (same meaning as "and", makes class declarations more intuitive to read in some places) * inverse (e.g. "inverse part-of"), used in place of a property name, when the "inverse of some property" doesn't have its own name Note, the missing support for Manchester syntax is not so problematic at the moment, as the example doesn't work anyway, _semantically_. @Till, recall that we would like to revise/extend the translation from propositional logic to OWL, so that one propositional axiom applies to as many OWL individuals as desired. Cheers, and thanks, Christoph -- Christoph Lange, Universit?t Bremen http://kwarc.info/clange, Skype duke4701 My up-to-date presentations: http://www.slideshare.net/langec From till at informatik.uni-bremen.de Fri Jan 20 20:38:11 2012 From: till at informatik.uni-bremen.de (Till Mossakowski) Date: Fri, 20 Jan 2012 20:38:11 +0100 Subject: [Hets-devel] How do I reuse an OWL ontology (in an RDF/XML file) in a heterogeneous spec? In-Reply-To: <4F199B46.6090904@uni-bremen.de> References: <4F199B46.6090904@uni-bremen.de> Message-ID: <4F19C2A3.1030206@informatik.uni-bremen.de> Hi Christoph, > I would like to reuse the OpenAAL OWL ontology, which I have downloaded > as an RDF/XML file OpenAALOntology.owl in the same directory. > * What do I have to write in my heterogeneous specification, in order to > be able to import that OWL ontology? basically you should write: from OpenAALOntology get httpwwwopenaalorgSAMOntology but this does not work, because Hets then tries to interpret OpenAALOntology.owl as a HetCASL file. Christian, can you fix this? As a workaround, you can convert the OWL ontology to Manchester syntax, using either Prot?g? or Hets (with hets -g OpenAALOntology.owl, then click onto the node and select "Show theory" and then save the theory), and include it in some HetCASL file. > * Or do I have to somehow pass it to Hets as an additional command-line > argument? no. Best, Till From christoph.lange at uni-bremen.de Sat Jan 21 02:39:08 2012 From: christoph.lange at uni-bremen.de (Christoph LANGE) Date: Sat, 21 Jan 2012 02:39:08 +0100 Subject: [Hets-devel] How do I reuse an OWL ontology (in an RDF/XML file) in a heterogeneous spec? In-Reply-To: <4F19C2A3.1030206@informatik.uni-bremen.de> References: <4F199B46.6090904@uni-bremen.de> <4F19C2A3.1030206@informatik.uni-bremen.de> Message-ID: <4F1A173C.3000309@uni-bremen.de> Hi Till, I tried some further steps concerning the reuse of that OWL ontology in a heterogeneous spec, but I encountered further problems. Let me list all of them here. Note that not all of them are crucial for the _CL_ guide, but on the other hand the CL part of my spec won't work either, until I know a solution for at least some of the problems. I think the best solution would be a joint hacking session (maybe [also] involving Christian and/or Francisc?) in order to really get these files right: * OpenAALOntology.het (the OWL ontology to be reused) * AAL.het (the heterogeneous spec that reuses it) And in order to get the Java parts of Hets running. OWL is not working properly, and I have no idea how to install it. Could we maybe do something like that on Monday? 2012-01-20 20:38 Till Mossakowski: > basically you should write: > > from OpenAALOntology get httpwwwopenaalorgSAMOntology ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Seriously? In the HetCASL file into which I wrapped the OWL ontology according to these instructions ? > As a workaround, you can convert the OWL ontology to Manchester syntax, > and include it in some HetCASL file. ? I just gave the spec the non-namespaced name "OpenAAL", and it seems to work. By chance I found out that Hets is to some extent aware of namespaces and prefixes in OWL ontologies, e.g. Prefix: p: inside an OWL module is supported. Importing the OpenAAL ontology, which binds a number of prefixes, and then renaming using a prefixed identifier ? that works, e.g. OpenAAL with p:AP |-> AssistedPerson But I can't make sense of the error message that "hets -g AAL.het" is now giving me: > a:Appointment is not a predefined IRI I'm even getting that error when commenting all modules of my heterogeneous spec that are not urgently relevant wrt. the CL Guide, i.e. when I cut the spec to spec OurAAL = OpenAAL then logic CommonLogic : { %% the CL stuff } a:Appointment is not even used in AAL.het! Inside OpenAALOntology.het, the namespace prefix a: is bound, and a:Appointment is declared as a class. When I parse OpenAALOntology.het with Hets, I'm not getting any error message. One further problem that is not related to the CL part of the example: In my previous mail I claimed that OWL-Manchester's "that" and "inverse" keywords are not supported, but that was a false alarm. The actual problem is probably that mapping symbols (here: propositional variables) to complex OWL class expressions doesn't work (yet?). I'm trying to do that below my propositional ontology module. (So it is not of any urgent relevance wrt. the CL Guide, and I have commented that code for now.) > As a workaround, you can convert the OWL ontology to Manchester syntax, > using either Prot?g? or Hets (with hets -g OpenAALOntology.owl, then > click onto the node and select "Show theory" and then save the theory), With Hets it didn't work for me. I installed Hets from the sources according to the instructions in README; just that I didn't understand what this concretely means: > Further environment variables for provers are: > > ... > HETS_OWL_TOOLS Trying to run "hets -g OpenAALOntology.owl" without having that variable set gives me the error "OWL2Parser.jar not found". Indeed there is no such JAR, so I have to build it. I found out that there is an undocumented target "make java-libs", but that doesn't work. It expects an OWL API directory in some directory that I don't have: > $ make java-libs > ... > ant -q java-libs > BUILD FAILED > /mnt/data/svn/agbkb.informatik.uni-bremen.de/Hets/build.xml:4: The following error occurred while executing this line: > /mnt/data/svn/agbkb.informatik.uni-bremen.de/Hets/OWL2/java/build.xml:106: Basedir /mnt/data/svn/agbkb.informatik.uni-bremen.de/Hets/OWL2/java/OwlApi/antbuild does not exist BTW, another final thing I have not yet fully understood is that library paths seem to be "semi-absolute" in a way, i.e. I made that part of AAL.het work by writing library AAL from OpenAALOntology get OpenAAL and then calling "hets -g AAL.het". However I can't call Hets from other directories. Other examples in Hets-lib use the whole directory path below "Hets-lib" as library identifiers, and then you have to run Hets from the "Hets-lib" directory. Cheers, Christoph -- Christoph Lange, Universit?t Bremen http://kwarc.info/clange, Skype duke4701 My up-to-date presentations: http://www.slideshare.net/langec From christoph.lange at uni-bremen.de Sat Jan 21 13:07:37 2012 From: christoph.lange at uni-bremen.de (Christoph LANGE) Date: Sat, 21 Jan 2012 13:07:37 +0100 Subject: [Hets-devel] Several minor inconsistencies in COLORE Message-ID: <4F1AAA89.1030009@uni-bremen.de> Hi Michael, dear Hets developers, for an example in the "Hets for CL Users" guide, I did some searches in COLORE ? more exactly, in the dump that is part of the Hets library: https://svn-agbkb.informatik.uni-bremen.de/Hets-lib/trunk/CommonLogic/colore. I noticed some inconsistencies, but I'm not sure who is to blame: COLORE, the Hets library, or both. 1. Filenames: Most of the *.clif filenames are lowercase, whereas in the CLIF sources, the modules are named in CamelCase. Mac OS and Windows ignore the case of filenames, but on Linux it makes a difference. Also, some filenames do include uppercase characters, e.g. lat/complementedLattice.clif. 2. Import syntax: I mentioned in an earlier mail that COLORE uses the non-standard cl-foo syntax instead of cl:foo. For Hets that's no problem. But some imports are declared using "cl-import" instead of "cl-imports". You can verify that using find . -name '*.clif' | xargs grep 'cl-import[^s]' 3. Module names and import references: In some cl-import and cl-module declarations, the argument is enclosed in parentheses. You can verify that using find . -name '*.clif' | xargs egrep 'cl-(module|imports?) \(' Cheers, Christoph -- Christoph Lange, Universit?t Bremen http://kwarc.info/clange, Skype duke4701 My up-to-date presentations: http://www.slideshare.net/langec From Mudcat at mie.utoronto.ca Sat Jan 21 15:42:59 2012 From: Mudcat at mie.utoronto.ca (Michael Gruninger) Date: Sat, 21 Jan 2012 09:42:59 -0500 Subject: [Hets-devel] Several minor inconsistencies in COLORE In-Reply-To: <4F1AAA89.1030009@uni-bremen.de> References: <4F1AAA89.1030009@uni-bremen.de> Message-ID: <1327156979.4f1acef3e59cd@webmail.mie.utoronto.ca> Hello Christoph, I am in still in the process of completing an overhaul of the ontologies in COLORE for Version 2.0 of the repository. It is addressing all of the issues that you raise in your message. Everything should be finished by the end of the month. Michael Quoting Christoph LANGE : > Hi Michael, dear Hets developers, > > for an example in the "Hets for CL Users" guide, I did some searches in > COLORE ??? more exactly, in the dump that is part of the Hets library: > https://svn-agbkb.informatik.uni-bremen.de/Hets-lib/trunk/CommonLogic/colore. > > I noticed some inconsistencies, but I'm not sure who is to blame: > COLORE, the Hets library, or both. > > 1. Filenames: > > Most of the *.clif filenames are lowercase, whereas in the CLIF sources, > the modules are named in CamelCase. Mac OS and Windows ignore the case > of filenames, but on Linux it makes a difference. > > Also, some filenames do include uppercase characters, e.g. > lat/complementedLattice.clif. > > 2. Import syntax: > > I mentioned in an earlier mail that COLORE uses the non-standard cl-foo > syntax instead of cl:foo. For Hets that's no problem. > > But some imports are declared using "cl-import" instead of "cl-imports". > > You can verify that using > > find . -name '*.clif' | xargs grep 'cl-import[^s]' > > 3. Module names and import references: > > In some cl-import and cl-module declarations, the argument is enclosed > in parentheses. You can verify that using > > find . -name '*.clif' | xargs egrep 'cl-(module|imports?) \(' > > Cheers, > > Christoph > > -- > Christoph Lange, Universit??t Bremen > http://kwarc.info/clange, Skype duke4701 > > My up-to-date presentations: http://www.slideshare.net/langec > > From christoph.lange at uni-bremen.de Mon Jan 23 12:03:29 2012 From: christoph.lange at uni-bremen.de (Christoph LANGE) Date: Mon, 23 Jan 2012 12:03:29 +0100 Subject: [Hets-devel] [CL] Announcing 'Hets for Common Logic Users' guide; your feedback appreciated In-Reply-To: References: <4F1AFD8A.7010601@uni-bremen.de> Message-ID: <4F1D3E81.10908@uni-bremen.de> Chris, thanks for your feedback; please allow me to share it with the hets-devel list for further discussion. 2012-01-21 20:40 Chris Mungall: > Thanks for making this available - the ability to integrated OWL2 and CLIF in this way will be very useful. > > I have one question - the spec refers to 0.99 That is probably a mistake, ? > Most of the downloads here are for 0.98 and lower > http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/installation_e.htm ? as, indeed, the current version of Hets is 0.98. I need to clarify with the other authors, but if we intended to say that the user guide _itself_ has version 0.99 we should add some clarifying notes that that number does not refer to the version of Hets. Or drop the version number of the user guide altogether. Cheers, Christoph -- Christoph Lange, Universit?t Bremen http://kwarc.info/clange, Skype duke4701 My up-to-date presentations: http://www.slideshare.net/langec From gruninger at mie.utoronto.ca Wed Jan 25 16:36:02 2012 From: gruninger at mie.utoronto.ca (Michael Gruninger) Date: Wed, 25 Jan 2012 10:36:02 -0500 Subject: [Hets-devel] Several minor inconsistencies in COLORE In-Reply-To: <4F1AAA89.1030009@uni-bremen.de> References: <4F1AAA89.1030009@uni-bremen.de> Message-ID: <4F202162.8070705@mie.utoronto.ca> Hello Christoph, On 12-01-21 7:07 AM, Christoph LANGE wrote: > Hi Michael, dear Hets developers, > > for an example in the "Hets for CL Users" guide, I did some searches > in COLORE ? more exactly, in the dump that is part of the Hets > library: > https://svn-agbkb.informatik.uni-bremen.de/Hets-lib/trunk/CommonLogic/colore. > > I noticed some inconsistencies, but I'm not sure who is to blame: > COLORE, the Hets library, or both. > > 1. Filenames: > > Most of the *.clif filenames are lowercase, whereas in the CLIF > sources, the modules are named in CamelCase. Mac OS and Windows > ignore the case of filenames, but on Linux it makes a difference. The COLORE goup has decided that the naming convention in COLORE will use lowercase underscores, rather than CamelCase. This will appear in COLORE 2.0 > > Also, some filenames do include uppercase characters, e.g. > lat/complementedLattice.clif. > > 2. Import syntax: > > I mentioned in an earlier mail that COLORE uses the non-standard > cl-foo syntax instead of cl:foo. For Hets that's no problem. I am attaching the Defects Report that Cameron Ross completed. If you look at Defect 24707/012, you will see that cl:foo is replaced by cl-foo. The parser that Cameron wrote (and which we use) follows this revision. > > But some imports are declared using "cl-import" instead of "cl-imports". These must have been from files that somehow missed the parser check. > > You can verify that using > > find . -name '*.clif' | xargs grep 'cl-import[^s]' > > 3. Module names and import references: > > In some cl-import and cl-module declarations, the argument is enclosed > in parentheses. You can verify that using > > find . -name '*.clif' | xargs egrep 'cl-(module|imports?) \(' Thanks for finding these errors. - michael -------------- next part -------------- A non-text attachment was scrubbed... Name: 24707-defect-report.pdf Type: application/pdf Size: 89599 bytes Desc: not available URL: