Hi Alex, In fact, this is because when I finished my phd, my prototype was almost broken due to the continuous changes in moose. So I decided that the next version of ConAn would be Moose independent to avoid to suffer with Moose evolution and also to be able to work with any domain (not necessary related to Moose). The implementation you have seen is something I started last year recovering my code from my prototype in Store in Bern. I know the advantages of FAMIXEntity, but it was a nightmare to repair ConAn in the times that Moose changes a lot. Anyway, and considering that moose seems to be stable, I think that we can make it a subclass of that class. I should think about it. cheers, gabi PS: When are you moving to Chile? > Hi Gabriela, > > I was wondering why your implementation of FCA isn't implemented > as an extension of FAMIX. Even if FCA has little to do with > Structural entities in OO languages, FAMIXEntity could be > inherited. For example, why FormalConcept is not a subclass of it? > > Cheers, > Alexandre

Hi Gabriela I would suggest going the Mondrian way, that is creating an independent, scriptable framework for formal concept analysis, which could then easily be plugged into Moose. This way you dont care if we break something here, and we still benefit from FCA on a case-by-case basis. Of course designing a good API is hard but in the end, it is much more valuable (because it gives a stable vocabulary in a unique location) and much easier to maintain than a mix of extensions interweaved in the system for tight integration. Doru did an excellent job with Mondrian, we should draw inspiration from that. What are the advantages of subclassing FamixEntity for a formal concept? Meta description for browsing and persistence? It is not necessary as we use pragmas to metadescribe. On 19 mai 09, at 11:25, Alexandre Bergel wrote: > Ok, it makes sense. I just loaded the last version of ConAn, but > some error popped up. What is the status ? Where you happy about > the implementation? > I can port SCG-Algorithm to Pharo, this shouldn't be a problem > since I am experienced in this kind of porting. > > I suggest to use the Moose mailing list for FCA related > discussion. > > Cheers, > Alexandre > > > > On 19 May 2009, at 05:13, Gabriela Arevalo wrote: > >> Hi Alex, >> >> In fact, this is because when I finished my phd, my prototype >> was almost broken due to the continuous changes in moose. >> So I decided that the next version of ConAn would be Moose >> independent to avoid to suffer with Moose evolution and also to >> be able to work with any domain (not necessary related to >> Moose). The implementation you have seen is something I started >> last year recovering my code from my prototype in Store in Bern. >> I know the advantages of FAMIXEntity, but it was a nightmare to >> repair ConAn in the times that Moose changes a lot. >> >> Anyway, and considering that moose seems to be stable, I think >> that we can make it a subclass of that class. I should think >> about it. >> >> cheers, >> >> gabi >> >> PS: When are you moving to Chile? >>> Hi Gabriela, >>> >>> I was wondering why your implementation of FCA isn't >>> implemented as an extension of FAMIX. Even if FCA has little >>> to do with Structural entities in OO languages, FAMIXEntity >>> could be inherited. For example, why FormalConcept is not a >>> subclass of it? >>> >>> Cheers, >>> Alexandre >> > > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > > > _______________________________________________ > Moose-dev mailing list > Moose-dev(a)iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev > -- Simon _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

The algorithm should have nothing to do with Moose, but instead it should work at the level of regular Smalltalk objects, just like Simon suggested. In this way you can use it in various contexts. This is what SCG Algorithms does. Of course, when you have a specific type of concept (like a design pattern or something) that should be a subclass of MooseEntity and can be linked with various FAMIX entities. Cheers, Doru On 19 May 2009, at 13:51, Alexandre Bergel wrote: > In a moose model, you can have histories, files, entities > associations among other things. Why not FCA concepts ? > One advantage to have FCA concepts as a subclass of FAMIXEntity > is to enable link between concepts and structural entities > (everything in a same moose model) and benefit from all > navigation tools. > > Cheers, > Alexandre > > > On 19 May 2009, at 13:42, Simon Denier wrote: > >> Hi Gabriela >> >> I would suggest going the Mondrian way, that is creating an >> independent, scriptable framework for formal concept analysis, >> which could then easily be plugged into Moose. This way you >> dont care if we break something here, and we still benefit from >> FCA on a case-by-case basis. >> Of course designing a good API is hard but in the end, it is >> much more valuable (because it gives a stable vocabulary in a >> unique location) and much easier to maintain than a mix of >> extensions interweaved in the system for tight integration. >> Doru did an excellent job with Mondrian, we should draw >> inspiration from that. >> >> What are the advantages of subclassing FamixEntity for a formal >> concept? Meta description for browsing and persistence? It is >> not necessary as we use pragmas to metadescribe. >> >> >> On 19 mai 09, at 11:25, Alexandre Bergel wrote: >> >>> Ok, it makes sense. I just loaded the last version of ConAn, >>> but some error popped up. What is the status ? Where you happy >>> about the implementation? >>> I can port SCG-Algorithm to Pharo, this shouldn't be a problem >>> since I am experienced in this kind of porting. >>> >>> I suggest to use the Moose mailing list for FCA related >>> discussion. >>> >>> Cheers, >>> Alexandre >>> >>> >>> >>> On 19 May 2009, at 05:13, Gabriela Arevalo wrote: >>> >>>> Hi Alex, >>>> >>>> In fact, this is because when I finished my phd, my prototype >>>> was almost broken due to the continuous changes in moose. >>>> So I decided that the next version of ConAn would be Moose >>>> independent to avoid to suffer with Moose evolution and also >>>> to be able to work with any domain (not necessary related to >>>> Moose). The implementation you have seen is something I >>>> started last year recovering my code from my prototype in >>>> Store in Bern. >>>> I know the advantages of FAMIXEntity, but it was a nightmare >>>> to repair ConAn in the times that Moose changes a lot. >>>> >>>> Anyway, and considering that moose seems to be stable, I >>>> think that we can make it a subclass of that class. I should >>>> think about it. >>>> >>>> cheers, >>>> >>>> gabi >>>> >>>> PS: When are you moving to Chile? >>>>> Hi Gabriela, >>>>> >>>>> I was wondering why your implementation of FCA isn't >>>>> implemented as an extension of FAMIX. Even if FCA has little >>>>> to do with Structural entities in OO languages, FAMIXEntity >>>>> could be inherited. For example, why FormalConcept is not a >>>>> subclass of it? >>>>> >>>>> Cheers, >>>>> Alexandre >>>> >>> >>> -- >>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >>> Alexandre Bergel http://www.bergel.eu >>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Moose-dev mailing list >>> Moose-dev(a)iam.unibe.ch >>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>> >> >> -- >> Simon >> >> >> >> _______________________________________________ >> Moose-dev mailing list >> Moose-dev(a)iam.unibe.ch >> https://www.iam.unibe.ch/mailman/listinfo/moose-dev > > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > > > _______________________________________________ > Moose-dev mailing list > Moose-dev(a)iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev -- www.tudorgirba.com "Yesterday is a fact. Tomorrow is a possibility. Today is a challenge." _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Gabi I know that adrian kuhn implemented also an FCA engine in VW probably in the SCG algo package. If you need help to migrate from VW to pharo please let us know. Stef On May 20, 2009, at 5:26 AM, Gabriela Arevalo wrote: > Hi Doru, > > What you propose seems to be reasonable for me. I added in this > discussion Diego Kogan because he is trying to help me in > recovering all my tools related to FCA. > One problem that I have here is that if I want to load Moose in > a VW, it is extremely low. > I tried to replicate here some versions of Moose when I was > trying to recover all my FCA stuff that I had implemented, and > so far this was a solution. > But maybe you can help us telling me if there is a way to speed > up the connection with Store in Bern from my university. > > Cheers, > > Gabi >> The algorithm should have nothing to do with Moose, but instead >> it should work at the level of regular Smalltalk objects, just >> like Simon suggested. In this way you can use it in various >> contexts. This is what SCG Algorithms does. >> >> Of course, when you have a specific type of concept (like a >> design pattern or something) that should be a subclass of >> MooseEntity and can be linked with various FAMIX entities. >> >> Cheers, >> Doru >> >> >> On 19 May 2009, at 13:51, Alexandre Bergel wrote: >> >>> In a moose model, you can have histories, files, entities >>> associations among other things. Why not FCA concepts ? >>> One advantage to have FCA concepts as a subclass of >>> FAMIXEntity is to enable link between concepts and structural >>> entities (everything in a same moose model) and benefit from >>> all navigation tools. >>> >>> Cheers, >>> Alexandre >>> >>> >>> On 19 May 2009, at 13:42, Simon Denier wrote: >>> >>>> Hi Gabriela >>>> >>>> I would suggest going the Mondrian way, that is creating an >>>> independent, scriptable framework for formal concept >>>> analysis, which could then easily be plugged into Moose. This >>>> way you dont care if we break something here, and we still >>>> benefit from FCA on a case-by-case basis. >>>> Of course designing a good API is hard but in the end, it is >>>> much more valuable (because it gives a stable vocabulary in a >>>> unique location) and much easier to maintain than a mix of >>>> extensions interweaved in the system for tight integration. >>>> Doru did an excellent job with Mondrian, we should draw >>>> inspiration from that. >>>> >>>> What are the advantages of subclassing FamixEntity for a >>>> formal concept? Meta description for browsing and >>>> persistence? It is not necessary as we use pragmas to >>>> metadescribe. >>>> >>>> >>>> On 19 mai 09, at 11:25, Alexandre Bergel wrote: >>>> >>>>> Ok, it makes sense. I just loaded the last version of ConAn, >>>>> but some error popped up. What is the status ? Where you >>>>> happy about the implementation? >>>>> I can port SCG-Algorithm to Pharo, this shouldn't be a >>>>> problem since I am experienced in this kind of porting. >>>>> >>>>> I suggest to use the Moose mailing list for FCA related >>>>> discussion. >>>>> >>>>> Cheers, >>>>> Alexandre >>>>> >>>>> >>>>> >>>>> On 19 May 2009, at 05:13, Gabriela Arevalo wrote: >>>>> >>>>>> Hi Alex, >>>>>> >>>>>> In fact, this is because when I finished my phd, my >>>>>> prototype was almost broken due to the continuous changes >>>>>> in moose. >>>>>> So I decided that the next version of ConAn would be Moose >>>>>> independent to avoid to suffer with Moose evolution and >>>>>> also to be able to work with any domain (not necessary >>>>>> related to Moose). The implementation you have seen is >>>>>> something I started last year recovering my code from my >>>>>> prototype in Store in Bern. >>>>>> I know the advantages of FAMIXEntity, but it was a >>>>>> nightmare to repair ConAn in the times that Moose changes a >>>>>> lot. >>>>>> >>>>>> Anyway, and considering that moose seems to be stable, I >>>>>> think that we can make it a subclass of that class. I >>>>>> should think about it. >>>>>> >>>>>> cheers, >>>>>> >>>>>> gabi >>>>>> >>>>>> PS: When are you moving to Chile? >>>>>>> Hi Gabriela, >>>>>>> >>>>>>> I was wondering why your implementation of FCA isn't >>>>>>> implemented as an extension of FAMIX. Even if FCA has >>>>>>> little to do with Structural entities in OO languages, >>>>>>> FAMIXEntity could be inherited. For example, why >>>>>>> FormalConcept is not a subclass of it? >>>>>>> >>>>>>> Cheers, >>>>>>> Alexandre >>>>>> >>>>> >>>>> -- >>>>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >>>>> Alexandre Bergel http://www.bergel.eu >>>>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> Moose-dev mailing list >>>>> Moose-dev(a)iam.unibe.ch >>>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>>>> >>>> >>>> -- >>>> Simon >>>> >>>> >>>> >>>> _______________________________________________ >>>> Moose-dev mailing list >>>> Moose-dev(a)iam.unibe.ch >>>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >>> >>> -- >>> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >>> Alexandre Bergel http://www.bergel.eu >>> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Moose-dev mailing list >>> Moose-dev(a)iam.unibe.ch >>> https://www.iam.unibe.ch/mailman/listinfo/moose-dev >> >> -- >> www.tudorgirba.com >> >> "Yesterday is a fact. >> Tomorrow is a possibility. >> Today is a challenge." >> >> >> >> _______________________________________________ >> Moose-dev mailing list >> Moose-dev(a)iam.unibe.ch >> https://www.iam.unibe.ch/mailman/listinfo/moose-dev > > _______________________________________________ > Moose-dev mailing list > Moose-dev(a)iam.unibe.ch > https://www.iam.unibe.ch/mailman/listinfo/moose-dev >