AL/ppy/HC-06_15925_chat-transcript_21030330a.txt ------ Chat transcript from room: hc-06 2013-03-30 GMT-08:00 [PDT] ------ [3:09] AnatolyLevenchuk: We are started! [3:10] AnatolyLevenchuk: Please click on the link below to join my Mikogo session https://go.mikogo.com/?sp=&sid=723024095 If the above link does not work, you can follow these steps instead to join a session: 1. Go to http://go.mikogo.com 2. Enter the Session ID: 723-024-095 3. Enter your name 4. Click "Join Session" [3:11] anonymous morphed into VictorAgroskin [3:11] AnatolyLevenchuk: We will use this chat channel for synchronize all our activities today [3:11] VictorAgroskin: Hi everyone! [3:12] Martin Davtyan: Hello [3:13] PeterYim: Hello [3:14] anonymous1 morphed into Shishkin Dmitriy [3:15] AnatolyLevenchuk: Please join Victor in skype -- vic5784 [3:15] anonymous morphed into Alex Ivanov [3:15] PeterYim: Hello Anatoly, Hello Victor ... [3:16] Alex Ivanov: Hello from Russia! [3:16] AnatolyLevenchuk: we are preparing group call via skype now... [3:17] PeterYim: please have someone at the Hackathon-Clinics main chat-room too, and register access details there - http://webconf.soaphub.org/conf/room/summit_hackathon [3:20] VictorAgroskin: David, are you going to join us on voice? [3:22] AnatolyLevenchuk: Skype call in progress! [3:25] AnatolyLevenchuk: Now we have 4 of us here in our TechInvestLab office in Moscow and several people online -- 14:24 Moscow time now. [3:27] anonymous morphed into Pavel Selchukov [3:32] PeterYim: There will be an "open webcast" of this team's work at Sat 2013.03.30 - 5:00am PDT / 8:00am EDT / 4:00pm MSK / 1:00pm CET / 12:00 UTC [3:29] PeterYim: Connections: Main Chat-room (ALL Projects): http://webconf.soaphub.org/conf/room/summit_hackathon Dial-in for the "Open Webcast" segments of ALL Projects (ref. time schedule below) (3PRT) Phone (US): +1 (206) 402-0100 ... Conference ID: 141184# ; or (3PRU) Skype: joinconference ... Conference ID: 141184# (3PRV) see: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013_Hackathon_Clinics#nid3PRS . [3:44] VictorAgroskin: http://miuras.inf.um.es/evaluation/oquare/Metrics.html [3:48] AnatolyLevenchuk: We are started with metrics from http://miuras.inf.um.es/evaluation/oquare/Metrics.html (try to determine what this metrics means for ISO 15926). [3:54] AnatolyLevenchuk: We have representatives of 3 software environments but only 2 of them will be in work today. 4 of us here in office, 5 online -- 9 overall (14:54). [3:55] AnatolyLevenchuk: Discussing Lack of Cohesion in Methods (LCOMOnto) metric: how to adapt it to ISO 15926 from OWL-specific metric? [3:57] VictorAgroskin: igor2 who are you and do you want to be on voice? [4:03] AnatolyLevenchuk: We are started coding of LCOMO metric in 2 tools! [4:10] AnatolyLevenchuk: Presenter now Ilya Volkov (SPb). [4:14] igor2 morphed into Igor Katrichek [4:15] AnatolyLevenchuk: We have now Moscow, Saint-Petersburg, Surgut and Zurich in a team. [4:15] anonymous morphed into vvagr2 [4:16] vvagr2: >>> allcoi=find(type=part2.any.ClassOfIndividual) >>> allnonl=find(type=part2.any.Specialization, hasClassifier=out) >>> show(id=allcoi-allnonl) [4:23] AnatolyLevenchuk: items1 = find(type = part2.Specialization, hasSuperclass=out) items2 = find(type = part2.any.ClassOfIndividual) items3= items1-items2 show(id=items3) [4:25] AnatolyLevenchuk: Please, check your version of PCA RDL! Take latest! [4:30] AnatolyLevenchuk: Have 6 concurrent coding sessions (4 for .15926 Editor of TechInvestLab and 2 for DSL15926 of AgentLab)! [4:32] AnatolyLevenchuk: Currently we speak Russian, English will be since 16:00 for expert session (now 15:30 MSK). [4:33] AnatolyLevenchuk: Presenter in Mikogo now Victor (but he have ailev name there!). [4:36] AnatolyLevenchuk: Ilya presenter back! We have preliminary results for first metric now. [4:45] PeterYim: . == open webcast of the "HC-06. ISO 15926 Reference Data Validation" project will start in ~15 minutes (please dial-in a couple of minutes early) Phone (US): +1 (206) 402-0100 ... Conference ID: 141184# ; or (3PRU) Skype: joinconference ... Conference ID: 141184# (3PRV) * Main Chat-room: http://webconf.soaphub.org/conf/room/summit_hackathon * Project HC-05 chat-room: http://webconf.soaphub.org/conf/room/hc-05 . [4:46] List of members: Alex Ivanov1, AnatolyLevenchuk, DavidLeal1, Igor Katrichek, Martin Davtyan, PeterYim, Shishkin Dmitriy, VictorAgroskin, vvagr2 [4:47] AnatolyLevenchuk: Peter, our HC-06 chat room! [4:49] AnatolyLevenchuk: We will end our vic5784 skype session and join joinconference call in 10 minutes. [4:57] anonymous morphed into ilya_volkov [4:58] ilya_volkov: total = 0 count = 0 for v in result: next = v current = 0 while next: current += 1 next = find(type=part2.Specialization, hasSubclass=id, hasSuperclass=out) for v in next: if v in ref: count += 1 total += current avg = total/count [4:58] anonymous morphed into Pavel Selchukov [5:09] anonymous morphed into FQuattri [5:11] PeterYim: @FQuattri - please join us at the main chat-room and at the open webcast, Francesca - http://webconf.soaphub.org/conf/room/summit_hackathon [5:11] FQuattri: @Peter : Thanks! [5:14] FQuattri morphed into FrancescaQuattri [5:18] PeterYim: Hi Astrid ... please join us at the main chat-room and at the open webcast, Francesca - http://webconf.soaphub.org/conf/room/summit_hackathon [5:19] AstridDuqueRamos: Ok, thanks. [5:30] AnatolyLevenchuk: We are back. Staring to speedup calculation of first metric (LCOMO) on full PCA RDL dataset. [5:37] AstridDuqueRamos: Hello Victor and Anatoly In OQuaRE, we have some metrics that could be apply to ISO 15926. [5:38] AstridDuqueRamos: This metrics are defined in http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics [5:39] AstridDuqueRamos: One of them is LCOMOnto. [5:39] AnatolyLevenchuk: Astrid, we have it from http://miuras.inf.um.es/evaluation/oquare/Metrics.html [5:41] AnatolyLevenchuk: Now we have difficulties for find LOCOMOnto for 2800000 triples of PCA RDL. We already feformulated it from OWL to ISO 15926. [5:41] AnatolyLevenchuk: Kiev join us! , ! [5:42] AstridDuqueRamos: could you please send it me the uri and I Could calculate this metric [5:43] VictorAgroskin: Astrid, it is a very special representation of ontology [5:43] VictorAgroskin: it is not an OWL ontology, it is encoded in RDF/OWL [5:44] VictorAgroskin: This is one of the "OntologySummit2013_Hackathon_Clinics" Projects ... (3PI2) Project (HC-06): ISO 15926 Reference Data Validation (3PI3) Project Co-champions: VictorAgroskin (3PI9) Event Date / Time: (Day-1) Sat 2013.03.30 (possibly also Sun 2013.03.31) ... see: [ details] (3PI4) Participant Volunteers: (3PXR) Add your name below, along with any time constraints or other participation-related comments. Thank you! (3PXP) ... (3PXQ) Abstract (3PI5) This ontology clinic is aimed at the evaluation of publicly available ISO 15926 reference data, viewing it as an ontology for the engineering domain. We will look for compliance to upper ontology constraints, diagnose problems in reference data, evaluate ease of understanding and use of existing data, and make suggestions for ontology improvement. Another goal is to apply formal ontology quality metrics for data in question. (3PI6) The effort will develop rules and algorithms to support generic verification tests, and also attempt to invent and implement specialized checks and quality metrics for ISO 15926 reference data. (3PI7) see: [ [ Project document] ] (3PI8) Ontology involved (3PMO) The ISO 15926 is a standard for engineering data integration, sharing, exchange, and hand-over. The standard defines a generic data model as an upper ontology for an engineering domain. Extensive Reference Data Libraries (ontology data) for process plants are developed by community of users, which includes equipment manufacturers, engineering companies and owner/operator companies in oil and gas, nuclear power, petrochemical industries, and others. Standard development and application is jointly managed as JORD project by two industry associations: POSC Caesar Association and FIATECH. (3PMP) Upper ontology of the ISO 15926 is publicly available as an OWL representation of ISO 15926-2 at: (3PMQ) http://rds.posccaesar.org/2008/02/OWL/ISO-15926-2_2003 (3PMR) see https://www.posccaesar.org/wiki/ISO15926inOWL for more details (3PMS) JORD Reference Data Library is available as a reference data service: (3PMT) the human browser access & query page is http://posccaesar.org/endpoint/ (3PMU) the SPARQL endpoint is http://posccaesar.org/endpoint/sparql (3PMV) A snapshot file with full content of an endpoint is available for download at: (3PMW) http://rds.posccaesar.org/downloads/PCA-RDL.owl.zip (3PMX) [5:45] VictorAgroskin: sorry for big paste [5:45] VictorAgroskin: links are at the end [5:45] AstridDuqueRamos: ok [5:45] FrancescaQuattri: it's great, many thanks [5:50] AnatolyLevenchuk: Victor Agroskin pesenter, but his name is ailev now :-) [5:52] AnatolyLevenchuk: Two different algorithms for one metric: Victor+Martin vs. Ilya. Fight! [5:58] AnatolyLevenchuk: Ilya presenter now, he tried to adjust low-level code of .15926 Editor Scanner module. Still need more optimization. [6:12] AnatolyLevenchuk: Victor and Martin found 40837 of something (ClassesOfClasses included). Refine, please! But this is rather quick due to trick suggested by Martin. [6:16] AnatolyLevenchuk: At least we found about 1000 classes that not connected with Thing. These are candidates for error report that will be submitted to JORD. [6:18] vvagr2: env=set(['http://rds.posccaesar.org/2008/06/OWL/RDL#RDS398732751']) res=set() env=find(type=part2.Specialization, hasSubclass=out, hasSuperclass=env) res|=env-allsup [6:22] vvagr2: show('allsup', type=part2.any.Specialization, hasSuperclass=out) [6:26] AnatolyLevenchuk: First result: [6:27] ilya_volkov: items1= find(type = part2.Specialization, hasSuperclass=out) items2 = find(type = part2.any.ClassOfIndividual) res1 = items2-items1 expressstrings = find(id=items2-items1, type=part2.any.ClassOfInformationRepresentation) result=((items2-items1)-expressstrings) ref='http://rds.posccaesar.org/2008/06/OWL/RDL#RDS398732751' total = 0 count = 0 deep = 0 next = set([ref]) all = set() while next: deep += 1 print deep next = find(type = part2.Specialization, hasSubclass=out, hasSuperclass=next) next = next - all all |= next found = result & next print len(found ) for i in xrange(len(found )): total += deep count += 1 avg = total/count print avg [6:33] ilya_volkov: count += len(found) [6:34] ilya_volkov: [0, len) [6:36] AnatolyLevenchuk: LCOMOnto = 7 [6:38] VictorAgroskin: Astrid, we've made an estimate! [6:38] AstridDuqueRamos: ok, [6:39] AstridDuqueRamos: I was trying to measure, but it dont work with your ontologies [6:40] AstridDuqueRamos: How do yo estimate this value? [6:40] AstridDuqueRamos: and how can I help you?, [6:41] AnatolyLevenchuk: Astrid, is this a good ontology according this metrics in comparison with DOLCHE, CYC and others? 7 is for PCA RDL. This was counted with algorithm posted a couple replics up in this chat. [6:42] VictorAgroskin: Astrid, we've made certain adaptations of formula, selecting specific classes and relationships [6:42] VictorAgroskin: from our ontology representation [6:43] VictorAgroskin: the algorithm above runs in a specific environment developed for ISO 15926 reference data libraries [6:44] AstridDuqueRamos: Yes the scaled value is 4 [6:45] VictorAgroskin: Astrid, are all metrics required together to measure a quality? [6:45] AnatolyLevenchuk: Astrid, what is "scaled value"? [6:45] VictorAgroskin: or single measurment has sense also? [6:46] AstridDuqueRamos: you can see here. http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics [6:46] AnatolyLevenchuk: Astrid, where we can see examples of these metrics values for other ontologies for comparison? [6:47] AnatolyLevenchuk: Astrid, you have no "scale" in this page. What is "scaled value"? [6:48] AstridDuqueRamos: yes, we have some metrics online [6:48] AstridDuqueRamos: http://miuras.inf.um.es/evaluation/oquare/UOM/ContMetricsUOM.html [6:48] AstridDuqueRamos: yes in this page you can se metric/Score [6:48] AstridDuqueRamos: Score is refered to the scale. [6:52] Sent transcript to: peter.yim@cim3.com [6:52] AnatolyLevenchuk: Thus "7" is 2 score for LCOMOnto? [6:52] List of members: Alex Ivanov1, AnatolyLevenchuk, AstridDuqueRamos, FrancescaQuattri, Igor Katrichek, ilya_volkov, Martin Davtyan, PeterYim, Shishkin Dmitriy, VictorAgroskin, vnc2, vvagr2 [6:52] Sent transcript to: peter.yim@cim3.com [6:53] AstridDuqueRamos: No, this score is not updated, the updated score are in http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics [6:53] AstridDuqueRamos: and this value is 4 [6:55] AstridDuqueRamos: you can see, some explain about the score in my paper, which is this: [6:55] AstridDuqueRamos: First, it should be noted that the different metrics generate quantitative values in different ranges. Metrics such as LCOMOnto or WMCOnto produce an absolute value, whereas metrics such as RROnto and AROnto generate relative ones. Second, the SQuaRE scores of the quality characteristics and subcharacteristics are in the range 1 to 5, where 1 means not acceptable, 3 is minimally acceptable, and 5 is exceeds the requirements (ISO25000, 2005). Consequently, in OQuaRE, a mapping between the range of values of the metrics and the range 1 to 5 was needed, and such mapping must take into account that high values in the metrics might not correspond to a high quality score [6:56] AnatolyLevenchuk: Metric/Score 1 2 3 4 5 LCOMOnto >8 (6-8] (4-6] (2-4] <=2 [6:57] AnatolyLevenchuk: How you see 7 is for 4? [6:57] AstridDuqueRamos: mm sorry, [6:58] AstridDuqueRamos: yes is 2 [6:58] AstridDuqueRamos: is a low value. [6:58] AnatolyLevenchuk: Yes, we guess that :-) [6:59] AstridDuqueRamos: you can see in my paper http://ws.acs.org.au/jrpit/JRPITVolumes/JRPIT43/JRPIT43.2.159.pdf, how LCOMOnto has a good o bad influence in some subcharacteristics [6:59] AnatolyLevenchuk: Meanwhile we discussing second metric now: WMCOnto. [6:59] AstridDuqueRamos: ok [7:01] AnatolyLevenchuk: We now figured how to interpret most of the metrics in ISO 15926 case and how have fast calculation for it. [7:03] AnatolyLevenchuk: We will calculate second metric and will have a break to relocate to our home and have second part of hackathon in 20:00 MSK (to permit to join for people from USA). [7:15] AnatolyLevenchuk: Have difficulties in speed of calculations once more. Will have break, then in 20:00 resume work with some other tasks (not metrics). [7:16] AnatolyLevenchuk: Till 20:00 MSK (same communication channels)! Thank you for good work! [7:16] VictorAgroskin: We'll try to work with taxonomies and check for cros-taxonomy inconsistencies during this second part. [9:00] AnatolyLevenchuk1: Starting second half of our hackaton/clinic! Please click on the link below to join my Mikogo session https://go.mikogo.com/?sp=&sid=761769109 If the above link does not work, you can follow these steps instead to join a session: 1. Go to http://go.mikogo.com 2. Enter the Session ID: 761-769-109 3. Enter your name 4. Click "Join Session" [9:02] AnatolyLevenchuk1: Skype call will be from Victor Agroskin -- vic5784 [9:05] VictorAgroskin: ? [9:06] AnatolyLevenchuk1: We are ready, you can call. [9:06] anonymous morphed into ilya_volkov [9:06] ilya_volkov: hello [9:06] AnatolyLevenchuk1: I will call :-) [9:12] ilya_volkov: NOCOnto ~1 [9:19] AnatolyLevenchuk1: Now discussing NOCOnto meaning and how count it. [9:25] VictorAgroskin: Astrid, can you explain NOCOnto metrics [9:25] VictorAgroskin: why there is |C(leaf)i| in the formula [9:26] VictorAgroskin: ? [9:28] AnatolyLevenchuk1: http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics [9:34] ilya_volkov: for v in result: total += len(find(type=part2.Specialization, hasSubclass=v, hasSuperclass=out)) [9:37] AnatolyLevenchuk1: This was code of number of superclasses of leafs. We have 7 hachaton/clinic participants now. [9:45] AnatolyLevenchuk1: Discussing what meaning can be of all metrics and computational complexity of it. [9:53] anonymous morphed into iya_volkov [9:53] AnatolyLevenchuk1: NACOnto - Number of Ancestor Classes [9:54] AnatolyLevenchuk1: NOCOnto - Number of Children [9:58] AnatolyLevenchuk1: Please, use zoom in mikogo (screen shearing of 2560*1600 now). [9:59] VictorAgroskin: NOCOnto=1.2 [10:00] AstridDuqueRamos: is the average number of ancestor classes [10:00] AnatolyLevenchuk1: There will be detailed report of this hackathon finding. [10:00] iya_volkov: NACOnto=1.2 too [10:02] AstridDuqueRamos: it is important to see that NACONto is refering to leaf classes [10:02] AstridDuqueRamos: Mean number of ancestor classes per leaf class [10:03] AstridDuqueRamos: but NOCOnto is refering to the total of classes. Mean number of direct subclasses [10:03] AstridDuqueRamos: look up the definitions here http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics [10:04] VictorAgroskin: yes, the code above has a cycle by all leaf classes [10:04] AnatolyLevenchuk1: Stops with metrics now. Discussing of error detection for PCA RDL now. [10:08] AstridDuqueRamos: I have to leave now, if anything is needed, please contact me by email. Thanks. [10:09] AnatolyLevenchuk1: Astrid, thank you! [10:27] iya_volkov: [21:26:35] Ilya V: result=set() for v in items: r = pattrns.Classification.find(classifier=out, classified=v) if len(r) > 5: result.add(v) [10:28] AnatolyLevenchuk1: We try algorithmically to find nontrivial classification hierarchies in PCA RDL. [10:41] VictorAgroskin1: We are finishing our seccion for today [10:41] VictorAgroskin1: Our results are: [10:42] VictorAgroskin1: 3-4 metrics from http://miuras.inf.um.es/oquarewiki/index.php5/Quality_metrics calculated [10:45] VictorAgroskin1: New problems identified: element unconnected to the main taxonomy, elements probable forming cycle (which is not strictly against rules as ISO 15926 based non well founded set, but requires further investigation). [10:48] VictorAgroskin1: Thanks to everyone! Our detailed report will be published early next week. [10:51] List of attendees: Alex Ivanov, Alex Ivanov1, AnatolyLevenchuk, AnatolyLevenchuk1, AstridDuqueRamos, AstridDuqueRamos1, DavidLeal, DavidLeal1, FQuattri, FrancescaQuattri, Igor Katrichek, Martin Davtyan, Martin Davtyan1, MatthewWest, Pavel Selchukov, PeterYim, Shishkin Dmitriy, Shishkin Dmitriy1, VictorAgroskin, VictorAgroskin1, anonymous, anonymous1, igor, igor1, igor2, ilya_volkov, iya_volkov, vnc2, vvagr2 [10:53] AnatolyLevenchuk1: We finished! Thanks to all! ------