DynaCASE

Hi

Thank you for CC-ing me, Peter, since I don't watch Pharo developers mailing list.

About petri nets... My main goal is to implement simulation for DynaCASE (i.e. "make it move") and Petri nets are for me more like a tool for trying simulation prototypes than goal itself. I'm currently stuck with them on developing simulation itself and on Roassal (visualisation engine from Moose which we use) problems with composite shapes.

Anyway, Serge, I will improve them in time and I would be glad to hear what do you expect from it - what should it support, do etc.
Can't promise anything, but it might help us both.

Jan

Peter Uhnák wrote

Hi Serge,

I didn't want to hijack the original thread (more than I've already done),
so I made a new one instead.

Looks really impressive Peter !
> I put some pictures on my twitter account:
> https://twitter.com/SergeStinckwich/status/582799867409526784

Thanks! I'll guess I will have to start using twitter now. :)

Apparently there is an icon bug when you don't use the Eclipse set of icons.
>
Yeah, that's my mistake. Thanks for the bugreport.


> I'm interested by Petri Nets in order to do Process Mining.
>
Petri Nets are being developed by my colleague Jan Blizničenko (
https://github.com/bliznjan?tab=repositories ) so he may have some more
info. I've CCed him.


> What kind of process models are you expected to implement ?
> Do you plan to draw activity diagrams ?
>
Right now I'm working on BORM (Business and Object Relation Modeling)
http://ccm.fit.cvut.cz/methodologies/borm/ (some of the materials are in
Czech, but you can see diagrams in the first link in our current
Eclipse-based tool); the following will be DEMO
http://ccm.fit.cvut.cz/methodologies/demo/

I can imagine that it wouldn't be that complicated to extend current FSM to
have activity diagrams (it's just two extra elements), so if you are
interested in that I can look into it.
Also I still have to migrate documentation to github so other people can
extend it if they want.

Thanks!
Peter

On Tue, Mar 31, 2015 at 4:02 PM, Jan B. <[hidden email]> wrote:
> Hi
>
> Thank you for CC-ing me, Peter, since I don't watch Pharo developers mailing
> list.

Hi Jan,

> About petri nets... My main goal is to implement simulation for DynaCASE
> (i.e. "make it move") and Petri nets are for me more like a tool for trying
> simulation prototypes than goal itself. I'm currently stuck with them on
> developing simulation itself and on Roassal (visualisation engine from Moose
> which we use) problems with composite shapes.
>
> Anyway, Serge, I will improve them in time and I would be glad to hear what
> do you expect from it - what should it support, do etc.
> Can't promise anything, but it might help us both.

At the moment, I'm quite interested to be able to do Process Mining:
http://www.processmining.org/
The idea is to be able to extract from logs a process.
I have already done some experiments some time ago. I'm using a basic
alpha algorithm and build a Petri net that represent the process:
https://www.youtube.com/watch?v=wdf80t32v-0

The repo is here: http://smalltalkhub.com/#!/~SergeStinckwich/Moose-XES
Maybe I can use your Petri Nets viz instead of mine.
How I can load your P/N viz code in my image ?

Regards
--
Serge Stinckwich
UCBN & UMI UMMISCO 209 (IRD/UPMC)
Every DSL ends up being Smalltalk
http://www.doesnotunderstand.org/

On Tue, Mar 31, 2015 at 10:08 AM, Peter Uhnák <[hidden email]> wrote:
> Hi Serge,

Hi Peter,

> I didn't want to hijack the original thread (more than I've already done),
> so I made a new one instead.
>
>> Looks really impressive Peter !
>> I put some pictures on my twitter account:
>> https://twitter.com/SergeStinckwich/status/582799867409526784
>
> Thanks! I'll guess I will have to start using twitter now. :)

Yes, I think so ;-)

With Alex and a student visiting my lab at the moment, we are working
on BPMN visualisation.

> I can imagine that it wouldn't be that complicated to extend current FSM to
> have activity diagrams (it's just two extra elements), so if you are
> interested in that I can look into it.

Yes, I'm quite interested to do that in order to specify in a visual
way to specify agent behaviour in a multi-agent system.
Have a look to the CORMAS announce I made on the mailing-list

> Also I still have to migrate documentation to github so other people can
> extend it if they want.

Yes ;-)

--
Serge Stinckwich
UCBN & UMI UMMISCO 209 (IRD/UPMC)
Every DSL ends up being Smalltalk
http://www.doesnotunderstand.org/

I'm sorry for not replying for such long time.

Thank you for sharing.

I recommend loading via my baseline from github, which automatically loads all prerequisites including DynaCASE itself:

Gofer new
    url: 'http://smalltalkhub.com/mc/Pharo/MetaRepoForPharo40/main';
    configurationOf: 'GitFileTree';
    loadDevelopment.

Metacello new
baseline: 'DynaCASEPetriNets';
repository: 'github://bliznjan/dynacase-petrinets/repository';
load

You can take a look. As I mentioned before, it is far from complete, but I keep improving it every day so it should be usable in few weeks (petri nets model and visualisation, i.e. without simulation, much sooner).

Jan

SergeStinckwich wrote

On Tue, Mar 31, 2015 at 4:02 PM, Jan B. <[hidden email]> wrote:
> Hi
>
> Thank you for CC-ing me, Peter, since I don't watch Pharo developers mailing
> list.

Hi Jan,

> About petri nets... My main goal is to implement simulation for DynaCASE
> (i.e. "make it move") and Petri nets are for me more like a tool for trying
> simulation prototypes than goal itself. I'm currently stuck with them on
> developing simulation itself and on Roassal (visualisation engine from Moose
> which we use) problems with composite shapes.
>
> Anyway, Serge, I will improve them in time and I would be glad to hear what
> do you expect from it - what should it support, do etc.
> Can't promise anything, but it might help us both.

At the moment, I'm quite interested to be able to do Process Mining:
http://www.processmining.org/
The idea is to be able to extract from logs a process.
I have already done some experiments some time ago. I'm using a basic
alpha algorithm and build a Petri net that represent the process:
https://www.youtube.com/watch?v=wdf80t32v-0

The repo is here: http://smalltalkhub.com/#!/~SergeStinckwich/Moose-XES
Maybe I can use your Petri Nets viz instead of mine.
How I can load your P/N viz code in my image ?

Regards
--
Serge Stinckwich
UCBN & UMI UMMISCO 209 (IRD/UPMC)
Every DSL ends up being Smalltalk
http://www.doesnotunderstand.org/