Re: [Architecture] [Process center] Support for executable process models
Please add me to the tbd thread on processes and event streams :-) Best regards, Frank 2014-12-03 14:43 GMT+01:00 Chathura Ekanayake : > Hi Srinath, > > Yes, if a process is already deployed, it can just store the deployment > URL. > > An example scenario would be: > An employee wants to claim a travel expense. He logs in to the process > center deployed in his organization and search for travel expense claim > processes. This process can be in different formats. If only a textual > process or a non-executable BPMN is available, he just follows the > mentioned steps manually. If there is an associated executable model, he > can click the "Start" button to initiate an automated process. In that > case, he will be directed to the associated web app where he (and other > participants) can provide details via human tasks. > > As a next step, we can associate event streams with such executable > processes to facilitate runtime analysis. I will initiate another thread on > that after refining this idea bit more. > > Regards, > Chathura > > On Wed, Dec 3, 2014 at 5:49 PM, Nandika Jayawardana > wrote: > >> Hi Chathura, >> >> In this scenario, does the process center discover the process models and >> tasks deployed in BPS or a user manually configure / upload processes at >> process center which are deployed to BPS by a user action ? >> >> Regards >> NAndika >> >> On Wed, Dec 3, 2014 at 5:00 PM, Srinath Perera wrote: >> >>> Hi Chathura, >>> >>> I think one other case is it is deployed already, and we just get a URL. >>> >>> Also, I think we need to think how above can be useful in real life >>> scenario, basically we need to come up with a full story on how above can >>> be used. >>> >>> --Srinath >>> >>> >>> >>> On Wed, Dec 3, 2014 at 4:56 PM, Chathura Ekanayake >>> wrote: >>> Executable process model is a BPEL or BPMN model, which can be executed in a process runtime. We can support executable models in process center as follows: Executable model contains: - Process archive (e.g. BPEL archive) - [optional] Deployment URL (e.g. URL of a BPS server) - [optional] Webapp URL - This URL points to a web application which can be used to create a new process instance and work on human tasks defined in the process (i.e. this is the operational UI for the process). If a webapp URL is provided, process center provides a "Start" button which simply directs to that URL. For example, employees can search process models from the store side and start them if they are executable. If it is not provided, the human task web app shipped with BPS can be used to start and work on human tasks. If a deployment URL is provided, a "Deploy" button is shown, which deploys a process archive in the corresponding BPS server. Thoughts... Regards, Chathura >>> >>> >>> >>> -- >>> >>> Blog: http://srinathsview.blogspot.com twitter:@srinath_perera >>> Site: http://people.apache.org/~hemapani/ >>> Photos: http://www.flickr.com/photos/hemapani/ >>> Phone: 0772360902 >>> >>> ___ >>> Architecture mailing list >>> Architecture@wso2.org >>> https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture >>> >>> >> >> >> -- >> Nandika Jayawardana >> Senior Technical Lead >> WSO2 Inc ; http://wso2.com >> lean.enterprise.middleware >> > > > ___ > Architecture mailing list > Architecture@wso2.org > https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture > > ___ Architecture mailing list Architecture@wso2.org https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture
Re: [Architecture] [Process center] Support for executable process models
+1 Best regards, Frank 2014-12-04 4:40 GMT+01:00 Chathura Ekanayake : > Hi Chinthana, > > Process center is not used to interact with the BPS runtime or HT runtime. > Instead, end users use it to search/discover available processes and to > execute them if possible. If a user wants to execute a process, process > center delegates it to the associated webapp, which runs separately from > the process center. > > The idea here is that the process center connects all BPM related > products/runtimes in to a single platform. > > Regards, > Chathura > > On Thu, Dec 4, 2014 at 7:20 AM, Srinath Perera wrote: > >> +1, we need to make sure terminology and wording matches those scenarios. >> For example it should be "Start a Process" not "Create Instance" etc >> (latter is process jargon and user is not interested). >> >> --Srinath >> >> On Wed, Dec 3, 2014 at 7:13 PM, Chathura Ekanayake >> wrote: >> >>> Hi Srinath, >>> >>> Yes, if a process is already deployed, it can just store the deployment >>> URL. >>> >>> An example scenario would be: >>> An employee wants to claim a travel expense. He logs in to the process >>> center deployed in his organization and search for travel expense claim >>> processes. This process can be in different formats. If only a textual >>> process or a non-executable BPMN is available, he just follows the >>> mentioned steps manually. If there is an associated executable model, he >>> can click the "Start" button to initiate an automated process. In that >>> case, he will be directed to the associated web app where he (and other >>> participants) can provide details via human tasks. >>> >>> As a next step, we can associate event streams with such executable >>> processes to facilitate runtime analysis. I will initiate another thread on >>> that after refining this idea bit more. >>> >>> Regards, >>> Chathura >>> >>> On Wed, Dec 3, 2014 at 5:49 PM, Nandika Jayawardana >>> wrote: >>> Hi Chathura, In this scenario, does the process center discover the process models and tasks deployed in BPS or a user manually configure / upload processes at process center which are deployed to BPS by a user action ? Regards NAndika On Wed, Dec 3, 2014 at 5:00 PM, Srinath Perera wrote: > Hi Chathura, > > I think one other case is it is deployed already, and we just get a > URL. > > Also, I think we need to think how above can be useful in real life > scenario, basically we need to come up with a full story on how above can > be used. > > --Srinath > > > > On Wed, Dec 3, 2014 at 4:56 PM, Chathura Ekanayake > wrote: > >> >> Executable process model is a BPEL or BPMN model, which can be >> executed in a process runtime. We can support executable models in >> process >> center as follows: >> >> Executable model contains: >> - Process archive (e.g. BPEL archive) >> - [optional] Deployment URL (e.g. URL of a BPS server) >> - [optional] Webapp URL - This URL points to a web application which >> can be used to create a new process instance and work on human tasks >> defined in the process (i.e. this is the operational UI for the process). >> >> If a webapp URL is provided, process center provides a "Start" button >> which simply directs to that URL. For example, employees can search >> process >> models from the store side and start them if they are executable. If it >> is >> not provided, the human task web app shipped with BPS can be used to >> start >> and work on human tasks. >> >> If a deployment URL is provided, a "Deploy" button is shown, which >> deploys a process archive in the corresponding BPS server. >> >> Thoughts... >> >> Regards, >> Chathura >> >> >> > > > > -- > > Blog: http://srinathsview.blogspot.com twitter:@srinath_perera > Site: http://people.apache.org/~hemapani/ > Photos: http://www.flickr.com/photos/hemapani/ > Phone: 0772360902 > > ___ > Architecture mailing list > Architecture@wso2.org > https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture > > -- Nandika Jayawardana Senior Technical Lead WSO2 Inc ; http://wso2.com lean.enterprise.middleware >>> >>> >> >> >> -- >> >> Blog: http://srinathsview.blogspot.com twitter:@srinath_perera >> Site: http://people.apache.org/~hemapani/ >> Photos: http://www.flickr.com/photos/hemapani/ >> Phone: 0772360902 >> > > > ___ > Architecture mailing list > Architecture@wso2.org > https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture > > ___ Architecture mailing list Architecture@wso2.org https://mail.w
Re: [Architecture] Chevron diagram editor/viewer for the Process Store
Hi Himasha, thanks for your comments - they helped :-) I still have comments: I am still not so sure about the use of "Thumbnails": in the create_chevron.png you show the ability to add a thumbnail to a chevron (!), the edit_chevron.png depicts that I can add a thumbnail to a complete chevron diagram. The list_chevron.png then uses these thumbnails to present the chevrons as well as chevron diagrams that have been created before - did I get it now? If the latter is true, the pull-down menue should support to list "composite" chevron diagrams only, as well as "atomic" chevrons, i.e. chevrons that are (or may become) part of some chevron diagram. The value of the process property in the create_chevron.png is a "flat name". If it is a link (in the sense of a URL), the figure should show this. Otherwise, my question about the scope of uniqueness of names is still valid. I was not precise enough about "annotations on the canvas": users like to see textual annotations of individual chevrons on the canvas; these are some sort of descriptions of a chevron that are rendered close to the chevron on the canvas (see examples in my next mail that I'll send to Chathura). Best regards, Frank 2014-12-01 6:31 GMT+01:00 Himasha Guruge : > Hi Frank, > >> >> >>1. create_chevron: >> 1. Is the list of properties of a chevron fixed? Or does "add >> properties" mean that I can add more properties? I suspect that the >> list of >> property types of a chevron is pre-defined... >> >> >- Yes, the property list is fixed. Here 'add properties' mean that >the user can add the values for the provided properties. > > 2.What is the purpose of the Thumbnail? > >> >>- do we support to change the graphical appearance of a chevron? >> Especially supporting to change the color of individual chevrons will >> be >> helpful in practice. >> >> >- Thumbnail image will be displayed for that respective chevron >diagram in the list page. ( as shown in the list_chevron screenshot) > Currently we do not support change for the graphical appearance of a >chevron. If it is to be implemented , how would you suggest it to be ?( >ex: user selects a color from a provided list) > >3. Do we support multiple predecessors and successors? > >- For the initial requirements we did not support multiple >predecessors/successors, but this will be implemented in the next phase. > >4.Is a value of the Process property a link? If it is > only a name, what is the scope of its uniqueness (aka namespace)? > >- Yes, a value of the process property is linked to that relevant >process model. > > 5.Do we support drawing vertical chevron diagrams, i.e. > diagrams the arrows of which point from top to bottom? > >- Currently only horizontal diagrams are supported. > > 6.Do we support annotations of individual chevrons that > are shown on the canvas? > >- Yes, if a user clicks on an individual chevron that is placed on >the canvas, those annotations will be listed. ( as shown in create_chevron >screenshot). > > 2.list_chevron: > >> >>1. Please call the type "Chevron Diagram" (instead of just >>"Chevron"): A chevron is a single "arrow" in a chevron diagram. >> >> >- +1.Will change the type name to chevron diagram instead of chevron. > > 2.What's the purpose of the "Statistics" button? > >- 'Statistics' page comes by default for an asset type that is created >in ES. It contains charts that display rated assets, top assets etc. > > 3.edit_chevron: > >> >>1. Can we change "Edit Asset" into "Edit Chevron Diagram"? >> >> >-+1.Will change the title to Edit Chevron Diagram instead of Edit >Asset. > >2.The properties now contain an "owner" and a > "description" property, i.e. more properties than you show in > create_chevron; and the "process" property is > now missing: why? > >- There are two sets of properties that a user should fill. A >property list for the entire diagram as well as properties for each >chevron. > > For the entire chevron diagram ( ex : product payment) following > properties[1] will be included. > (name,owner,predecessor,successor,provider,version,description) > > Whereas for each individual chevron (ex:check payment type etc) , > following properties[2] will be included. > (name,owner,predecessor,successor,process) Therefore in chevron_edit page, > if the user clicks on the canvas, properties for the entire diagram will be > displayed.while if the user clicks on a specific chevron that relevant > properties will be listed. This is the same mechanism followed in > create_chevron page. > > 4.store_view: > >> >>1. Again, I recommend to speak about Chevron Diagrams and not only >>about "Chevron". Btw the
Re: [Architecture] [Process center] Support for executable process models
Hi Chathura, ​what is the semantics of the deployment URL (second bullet): is this the endpoint where a BPEL archive can be deployed? Or is this the URL of an archive that contains all the artifacts to create a BPEL engine where I can deploy and execute the BPEL file? I suspect the first alternative is what you mean - just double-check. I don't understand your third bullet: when you deploy a BPEL archive you (nearly always) deploy all roles of all partner links. After that an instance of the process model can be instantiated by any external source by sending messages to the ports that correspond to initiating activities of the process model. The same is true in case a BPEL file uses the BPEL4People extension, i.e. in case some activities are implemented by human tasks. Continuing on the "Webapp URL" (it must be something implementation specific that we have done at WSO2): A process model may be initiated by activities that have not user interaction at all (i.e. that is no peopleActivity activity kicking-off a new instance of the process model). What is the Webapp URL all about in this case? Pushing the "deploy" button is done by non-business-oriented users because it will require quite some technical information, right? I.e. do we support in the Process Store corresponding role models to avoid "frustration" by business users when pushing the "deploy" button? Best regards, Frank 2014-12-03 12:26 GMT+01:00 Chathura Ekanayake : > > Executable process model is a BPEL or BPMN model, which can be executed in > a process runtime. We can support executable models in process center as > follows: > > Executable model contains: > - Process archive (e.g. BPEL archive) > - [optional] Deployment URL (e.g. URL of a BPS server) > - [optional] Webapp URL - This URL points to a web application which can > be used to create a new process instance and work on human tasks defined in > the process (i.e. this is the operational UI for the process). > > If a webapp URL is provided, process center provides a "Start" button > which simply directs to that URL. For example, employees can search process > models from the store side and start them if they are executable. If it is > not provided, the human task web app shipped with BPS can be used to start > and work on human tasks. > > If a deployment URL is provided, a "Deploy" button is shown, which deploys > a process archive in the corresponding BPS server. > > Thoughts... > > Regards, > Chathura > > > ___ Architecture mailing list Architecture@wso2.org https://mail.wso2.org/cgi-bin/mailman/listinfo/architecture