Відео

It's here: ua-cam.com/video/U77SPPdfAos/v-deo.html

I'll post the info from Michael here: Stateless YAWL is a version intended for embedding into applications, and so doesn’t use a service architecture. It also doesn’t persist any data or process states (those responsibilities are passed to the client app). Client apps can listen for a variety of events while a process is executing and take appropriate action. Each call to the API to start a process or start and complete workitems returns the complete process state, which the client can persist if required. Any number of engines can be used concurrently (within hardware limits). Processes can be transferred between engines at any time, for load balancing purposes. Additionally, process cases can be monitored, and a time limit can be set for inactivity, upon which a client can choose to unload the process and reload it later as needed, saving memory resources. The resource perspective is ignored, as it is not part of stateless yawl; everything is a direct conversation between engine and client application. Hints on usage can be found in the file “YSExample.java”

Thanks, but are you referring to the pattern itself or the BPMN implementation?

@@yawlusergroup I meant to commend the YAWL implementation, because when you say the pattern is not directly supported in YAWL, I assume a complex workaround is needed.

I'm glad that you say that. In fact I wasn't quite so sure if this implementation is the best you can get. This pattern was much harder than most of the others.

If you are referring to the tennis videos: I didn't make them, they are stock footage. If you are referring to the Petri net: I made it with WoPeD: sourceforge.net/projects/woped

Yes, it's still an unsolved problem when it comes to process automation.

:-)

You are absolutely right. The video has been out since February and you are the first to have noticed. I corrected the title right away. Thanks!

@@yawlusergroup I was looking for précisions on how to properly use Joining Inclusive Gateways though, so if you plan on doing a video about that, keep me posted!😉

Upon your request I have just published this video which was already in the pipeline: ua-cam.com/video/GD-08--ARrE/v-deo.html Let me know what other aspects you are interested in that are not addressed there.

Oh, thanks, I didn't know that. I've never been there.

Thank you for your kind words. However, this channel gets many more viewers than expected, so the problem could only be solved with SEP.

@@yawlusergroup would you like to grow your UA-cam channel?

@@yawlusergroup would you like to increase your channel rich

Thanks - much appreciated!

Now fixed in YAWL 5.1

I am using WoPeD, the Workflow Petri net Designer. You can download it here: woped.dhbw-karlsruhe.de I have also made a video about WoPeD: ua-cam.com/video/LybCxtlRHpI/v-deo.html

Thank you. I am glad you like them.

I am glad that you like the content. It would certainly be interesting to further analyse the algorithms. If you want to perform some experiments with ProM, I can recommend the forum at promforum.win.tue.nl

BPMN and YAWL are not the same languages. You can use BPMN for modeling at a higher level of abstraction and then do the business process automation in YAWL. There exists a relatively simple way of translating from BPMN to YAWL as has been described here: www.researchgate.net/publication/221586372_Transforming_BPMN_Diagrams_into_YAWL_Nets

Glad it was helpful!

Oh, thank you.

easy to hack - probably, but what is a celar?

!!!!!!

If you want to find out yourself, you can find the Petri net on github: github.com/ahense/bpa Click on "Code"/Download zip to get the whole package. You also need the Petri net editor WoPeD: woped.dhbw-karlsruhe.de/

* I have made some experiments with the YAWL Web Service Invoker Service myself some time ago. The service wasn't able to handle my complex types at that time and I decided against using it. I would suggest that you invoke your web service using an ordinary YAWL codelet. * E-mail servers keep updating their protocols all the time because of security issues. I am afraid that I cannot help you here with the development. If you manage to solve the problem please let me know.

Thank you so much for your answer and your suggestion. I'll try and let you know!@@yawlusergroup

Thanks. You are right. The BPMN editor looks great and makes process definition a breeze.

These are actually two good questions. An intermediate timer event can be implemented in YAWL with a task that has a task timer. There is a video on task timers here: ua-cam.com/video/c0YewdYKnw8/v-deo.htmlsi=8UKfDXbQap8czCt8 Just make sure that the task isn't offered to anyone so that no one completes it before the task timer runs out. An interrupting timer boundary event in BPMN can be attached to simple tasks or subprocesses. Let's assume that we have it attached to a simple task A. You can use a YAWL task X with a task timer that also has a cancellation region. The cancellation region of X will include task A and the arc before it. There is a video on cancellation regions in YAWL here: ua-cam.com/video/srxbyaaCHKk/v-deo.htmlsi=K4PncY-D5d-vm720

Thanks a lot for your reply. I will indeed watch and study the two videos mentioned. "The cancellation region of X will include task A and the arc before it." Oh, that is the trick. Additionally, there's a need to set both minimum and maximum durations for a task. I'm exploring methods to implement this within YAWL. While I'm not sure how to create custom business logic for this yet, I'd really appreciate any suggestions or guidance for this case. Once again thanks for your reply. @@yawlusergroup

For a minimum duration you could put a task with a task timer parallel to your task and join the two afterwards with an AND-join. Then the control flow in this branch has to wait for the parallel timer before it can continue. However, I am not sure if it is this what you want.

That is what we need. Many thanks. Let's assume a with min and max durations of 2 and 10 days, respectively. If takes only 1 day, the AND-join will cause the flow to wait for the parallel 2-day timer (wait for another day) before it can continue. Thus, enforce a minimum of two days. It is helpful to be notified that task did not initially meet the minimum 2-day requirement. In other words, AND-join shouldn't be waiting, and now it is. I'd really appreciate any suggestions. @@yawlusergroup

You could try this. Define a task A with a task timer set to 2 days followed by a task B that informs that the minimum of two days has been respected. Then parallel to that Define your task T with a task timer of 10 days and a cancellation region that includes tasks A and B. Then join the two parallel branches with an OR-join.

Yes, you are right. This is an example for a very primitive user interface with just 3 buttons. I had something like a digital watch in mind, where you have just 2 buttons. You can insert the tokens in each of the three digit places in any order. The limited number of tokens in the "tokens" place guarantees that you can only open the safe with the correct code. The safe video at the beginning is just for demonstration purpose and may be a little misleading.

Glad you liked it!!

Petri nets are easy to implement in any programming language. You can use an array of non-negative integers for the marking and backward and forward matrices for the transitions. That being said, there are many Petri net implementations out there. It really depends on which programming language you prefer and what your requirements are.

@@yawlusergroup well I understood this, it's more like the mathematical way of describing it. What I acutely need is a practical example with a microcontroller, PLC or something similar like the whole thing from modelling to implementing

Petri nets abstract from time. They have nice mathematical properties like the decidability of the coverability problem and therefore you can automatically prove safety conditions in complex systems that are modelled with Petri nets. So a real system could consist of a Petri net where the microcontroller or PLC is firing transitions in real time. The Petri net "guarantees" that the safety conditions hold. I don't know any concrete example like the one you are looking for. Sorry if I cannot help you here. If you find something I would like to know.

@@yawlusergroup ok I see, thank you anyway sir

Thanks for the flowers :-)

@@yawlusergroup I have been playing with the product and please do not take this the wrong way. It appears to be very well thought out on the back end and as an overall architecture. The one spot that could do with some 2025 type upgrade is the ui for the editor and other items. To that end I might be interested in helping out with a state of the art react front end if there is interest? I am offering as an individual but I do have a very large client that I might be able to pitch on this effort l

You are right. The YAWL engine is rock solid. However, the user interface of the editor could use some upgrade. If you can help here you are more than welcome. You can write me an e-mail to the address in the channel info and we can have a discussion on how we can organise this.

No worries!

It's planned roughly every fortnight. Is this too slow?

@@yawlusergroup That’s perfect 👍

Petri nets by Wolfgang Reisig is the classic book.

Sure thing

Have you been everywhere else to know this?

yes

Here ec.europa.eu/transport/road_safety/going_abroad/czech_republic/traffic_lights_en.htm it says: red and green FLASHING together... In Germany they don't flash.

Yes, and the driving assist systems of modern cars only make things worse.

Why do you euros need to be warned what the next light is? Not dissing. In the US we just use yellow before red to tell you to slow down. Warning for a green just doesn't make sense to me

Oh, good to know. I'll keep my eyes open next time I drive there.

Flashing amber in the UK is at pedestrian crossing traffic lights as an additional time for pedestrians to finish crossing before the light turns green. The crossing signal for pedestrains turns to "Don't cross" the same time the lights start flashing Amber (or near enough the same time) so no new pedestrians should step onto the road. It also means that you may drive through a flashing amber traffic light if there are no pedestrians on the crossing. We also have the solid red + amber in UK to show the light is turning green. The reason for this is that if you're driving down the road and see a traffic light ahead, if it shows amber, you need to start slowing/stopping, if it shows red+amber you can keep going since it's turning to green (assuming no obstructions ofc, use normal driving sense lol)

@@humanhammer101 OK, duly noted.

​@@tanneroneill4160 as said in the short. Its useful since you can take out the gear or shut of the engine. Then with red yellow you get ready to drive again

yes

thanks

Sorry, I didn't get notified on your reply some days ago. UA-cam doesn't have the right notification setting or I don't see how to configure it. I have never seen duplicate case IDs in YAWL before. I have seen the "cached workitem... removed" messages however, but in all the cases that I have seen so far, the removal was correct. I am afraid I cannot help you here. I have experts in my company who could investigate this but we would need a service contract for this, so perhaps no option for you if you are just evaluating YAWL.

@@yawlusergroup Thanks for the info. The system is in production. I believe these problems started when we changed to dockerized tomcat with yawl services, as well as AWS RDS SQLServer backend. Shortly after, we started getting reports of missing cases. And at least once, a workflow specification disappeared. The system was working fine (as far as I know) when deployed in a regular Windows tomcat instance connected to an enterprise DB server. I suspect something in the startup/initialization sequence is failing. I will keep investigating.

OK, thanks for the info. I know that we did have some issues in the past with the docker of the database not being ready when the YAWL docker looked for its persistence layer. It seems you really need to check if the persistence layer is ready before starting the engine. But as already said, I'm not the expert on the technical level. I hope you can find the problem.

Just to close the loop on this: it appears to have been caused by a defective process specification. The org.yawlfoundation.yawl.engine.YEngineRestorer.restoreRunners() method was throwing an NPE on finding a null net when trying to restore busy tasks. By adding log statements, I found that one particular subnet was causing this. Upon close analysis of the subnet, it appeared that a net variable should have been set to I/O, but was only set to Input. Fixing this problem eliminated the engine restoration problem. I did not analyze why this defect caused the net persistence problem. The engine restore process swallows unknown exceptions and just emits a log warning about possible inconsistent state. But the ResourceManager initialization process is more severe. If it doesn't find a cached item in the engine, it deletes it completely. Because the null net appeared in several workitems, it caused a catastrophic data loss. I am working with a slightly modified fork of Yawl 4.2 codebase. I did not test or analyze 5.0 to see if this problem could happen there.

Thank you for sharing this thorough analysis with us. As far Yawl 5, my guess would be that it behaves in the same way here because it was mainly the UI that was charged.

Dear Bertus, Yawl needs a single XSD declaration in the editor for each workflow specification in a project. The best way I know - which is not very convenient - is to create the expanded file containing everything outside of the Yawl editor and copy it from there into each specification.

@@yawlusergroup thank you very much for the quick confirmation! It won't be the first time to do it and it won't be the last either ;-)

There is of course my playlist on workflow nets: ua-cam.com/play/PL4BZgFsmRzfTxESOUTVgoTml0K_DznehP.html Then there is an article by my colleague Thomas Freytag - the creator of WoPeD: woped.dhbw-karlsruhe.de/wp-content/uploads/WoPeD-PN-Newsletter-FINAL.pdf If these don't answer your questions, just ask me here. I'd be happy to help.

​@@yawlusergroup this pdf and playlist have nothing link to this petri net, i need something documentation and description about this "traffic light petri net" for my better understanding. I got the source file but couldn't understand even a single point of thid petri net all.

@@developerism_ On the left-hand side you find one traffic light with the states Red (Re), Yellow (Ye), and Green (Gr). States are depicted with circles in Petri nets and the current state with a token in the circle. On the right-hand side you find the other traffic light with the states Red (Re), Yellow (Ye), and Green (Gr). The traffic light on the left-hand side is red at the beginning and the other one is green. Transitions in Petri nets are depicted as squares. Here in the so-called token game in WoPeD, the active transitions are green and clicking on them fires the transisiton. When a transition is fired it consumes the token in the places of the incoming arcs and it produces new tokens in the places of the outgoing arcs. Does this answer your question?

Practically, there are two poles containing signals and this petri net shows the sequence of signals (red, yellow,green)? Is it?

Yes exactly, you can interpret it like this. You can also interpret it as two pairs of traffic lights - one pair for each road. I'll update the description of the video with this explanation.

You can find supplementary material on github: github.com/ahense/bpa Click on "Code"/Download zip to get the whole package. Let me know if it works.

@@yawlusergroup Thanks for sharing source file, but I am facing difficulties to set WopeD. I also installed JRE's latest version.

Try with an Oracle JDK v 11 and let me know if that works.

Do you mean the mean or average time that the execution of a task takes? What is called service time in the capacity planning? This is simply an estimate.

Sorry, I was confused because your question is not about this video but on the one with quantitative simulation. This "mean lambda" is the average number of cases that occur in the period given next to it. So if you simulate a working day with 8 hours you would put in mean lambda 50 to say that there are 50 cases within 8 hours. Does this answer your question?

but why exactly 50, is there a specific method to calculate it ?

no, it can be any number. If in our scenario we expect that there are 80 clients every day, then put in 80. It's just an estimated number. Capacity planning and also the quantitative simulation need this number as an input.

It's called YAWL - my whole channel is basically about it.