Create a Sample Simple Invoice for Quality Assurance Effortlessly

Sample simple invoice for Quality Assurance

chapter 5 section 4 process model quality assurance we need to make sure that our process models are of good quality we can distinguish three different quality aspects their syntactic quality and verification techniques help us to check them their semantic quality and validation methods help us to check them and there's pragmatic quality we use certification let's start with syntactic quality the structural correctness and behavior of correctness validation we will look into validity and completeness and for pragmatic quality we will have a look at understandability maintainability and learning structural correctness is the first matter that we need to consider here we see an example of a bpmn model that is not structurally correct it is obvious that elements should be all connected we have the reject order that is dangling apparently an arc is missing from the gateway there's a mid invoice and the arc pointing away from it is obviously not correctly connected with the gateway that follows the result is a structurally not correct bpmn model we need to fix all these structural incorrectnesses to make our model nice a bit more difficult to check is behavioral correctness it is also referred to as soundness soundness is a criterium that was defined for petrinet models it requires three conditions to hold there must be an option to complete that means any case that is processed must have the chance to reach one of the output events the second property says we need to have proper completion this means when we reach the end event it should not be the case that anything can still be done in the process and there should not be any dead activities it means all activities that we represent should be potentially been able to be executed these are technical requirements that can be checked with automatic formal tools for process verification and these verification tools search in the bpmn model for patterns like these let's start on the left hand side top there is an example of what is called a lack of synchronization on the left hand side is an end split the end split activates both branches on the right hand side is an x or gateway that merges these two branches that means both branches can be executed and it means that we potentially execute anything that is later than the xor xor join it might be executed double on the left hand side bottom is an example of a deadlock on the left hand side is an exo split and that makes sure that only one of the subsequent options is activated the end join on the right-hand side waits for both branches to complete but as only one of these branches is activated it is never possible that both of them would complete that means the process gets stuck it gets deadlocked at this end join on the right hand side we see an example of a potential lack of synchronization which is also a potential deadlock the example has an or split gateway on the left hand side this may activate one or the other or both of the subsequent branches if you have both branches activated you have the same problem as with the lack of synchronization if you activate only one branch and on the right hand side is not an xor but an and join you have also the situation similar to the deadlock finally on the right hand side bottom is an example of a so-called live lock this is a never-ending loop you start this loop by entering the loop through the x or join gateway on the left hand side on the right hand side you have an end split that means anytime you reach this end split you activate the branch outside the loop and also the branch inside the loop and this way you can never stop processing this loop all these patterns here are patterns that do not make much sense from a logical perspective verification tools help us to automatically identify such incorrect behavior the second aspect of quality is semantic quality semantic quality cannot be checked automatically we need to validate that means we need to build on the expertise of the domain expert in our validation by interacting and asking the domain expert we want to find out if the model that we're trying to model is valid that means everything that is represented in the model is actually correct and relevant and we also want to make sure that our model is complete that means everything that should be represented is actually represented the process analyst has to carefully ask questions in order to check these two aspects key is that validation cannot be done based on the model alone validation builds on the relationship between the model and the actual real world process as domain experts understand them so let's consider again our example of the orders that are forwarded as sub-orders to the amsterdam and the hamburg warehouse remember that was the example that we used to introduce the or join and the ore split the idea of that process is that we may have three situations it may be that we forward some orders to amsterdam it may be that we forward some orders to hamburg and it may be the case that we forward sub orders to both warehouses however it can never be the case that we order forward no sub order at all to any warehouse so what can we say about the semantic quality of this model there is an issue it's invalid because we potentially may bypass any of the forwarding activities that means after the and split we may choose to select the path order does not contain absolute products as well as the path order does not contain hamburg products if that is the case we do not produce any sub-orders apparently this is not what is really happening that means this model is semantically not correct ing to the description of the process what about this model here we see that there is an exer split we have to choose to either produce sub-orders for amsterdam or for hamburg this model is not complete apparently we cannot represent the fact that it is also possible to have orders for amsterdam and for hamburg this means this model is also semantically not correct but for a different reason it may be the case that we have sub-orders produced for both of these warehouses now let's look at pragmatic quality it is not only that our models have to be syntactically and semantically correct they also have to be efficient and effective for the people working with them there are different concerns of usability models should be understandable maintainable and facilitate learning different model characteristics such as size structure and layout influence the usability let's first look at the labeling of the activities and events in the process there are some conventions that have been established to make the labels effective a recommendation is that activities should be written with an imperative verb and a corresponding noun events should also refer to a noun and a past participle verb we should also write conditions on the outgoing arcs of or and xor splits with a reference to a particular object of the process furthermore we should also make sure that the layout is nice models should run from top left to bottom right and we should use direct arcs with no crossings look at the two models at the bottom of this picture the left hand side is exactly the same as the one on the right hand side with the only difference the layout is effective on the right hand side you see what a difference it makes in the impression and reading of this process model beauty is not an option make sure that your models look nice various guidelines have been formulated to help making models nice one set of guidelines are the seven process modeling guidelines guideline 1 says we should use as few elements as possible in the model in this way we can make sure that the model is compact and easy to understand we should also try to minimize the routing paths per element in this way the model is also nicer and easier to read in many cases we can represent a choice as a yes and no choice between two options third we should use one start event and one end event if that is possible it has to be clear what triggers the process and what outcomes the process has sometimes we want to represent that there's a positive and a negative outcome guideline 4 says that we should create models that are structured structured models have corresponding gateways of the same type this means that an x or split has a corresponding x xor join an n split has a corresponding and join and an or split has a corresponding or join in this way we can nicely structure the model as opening and closing brackets rule 5 says we should avoid or gateways where possible sometimes in very specific situations or gateways are very helpful but we should use them scarcely because for some readers of the model they're quite complex guidelines 6 says we should use verb object activity labels and finally guideline 7 says we should decompose a model that has more than 30 elements this is a rule of thumb the model should not be too big and too crowded otherwise it is difficult to understand let's have a look at an exercise that we have in our book in relation to the seven process modeling guidelines it is about a complaint a complaint is triggered by a phone call by a complaining customer it is decided whether the complaint can be handled or whether it has to be referred to an internal or external party an external referral leads to a telephone confirmation to the external party an internal referral is added to the incident agenda if no referral is needed a complaint analysis is conducted and the complainant is conducted in either case the complaint is archived and the case is closed what you see below is a bpmn model that domain experts have created this is a real world example you may be surprised how complicated this model looks like you can use the seven process modeling guidelines to check what can be done to redesign this process and make it easier to understand you may be surprised how easy this model is to understand after revision have a look into the book to see how this model looks like let's recap we recap we have seen that process discovery involves defining the setting gathering information conducting modeling and assuring model quality there are various discovery methods such as document analysis observation automatic process discovery interviews and workshops a modeling method proceeds by identifying process boundaries then activities and events then resources and their handoffs then the control flow and finally additional elements quality assurance relates to syntactic semantic and pragmatic quality