Get best of the week

Business process modeling, automatic diagram-text translation and CH-1 notation

Hello everyone.

By the nature of my activity, I pretty much had to model a lot of business processes of various organizations. Both existing companies (with the goal of systematizing and optimizing existing activities), as well as new ones, i.e. start-ups (designing activities from scratch). In this article I will try to briefly summarize the goals of such modeling ( section I ), the main types of models ( section II ), talk about my instrumental developments ( section III ), and also reflect on what is still missing , including and in terms of the course on import substitution ( section IV ).

(I) What is it and why is all this needed


Indeed, the first and natural question is what is it and why is it needed at all in an enterprise? Let's go and we canonically and start right from the beginning (your K.O.). So:

(1) The business processes of an enterprise are simply the totality of all its internal processes, that is, allegorically, this is the “ physiology of the enterprise ” (while the organizational structure is its “anatomy”). To manage something, you need, at a minimum, to know how it works .

It is important to understand that any business process (i.e. business process) is simply some kind of work technology that either already actually exists in the organization, or is supposed to be implemented (projected), butnot a document or “leaf with squares and arrows” . No, any process is a technology, the procedure for performing any actions necessary for a business (for an organization). Moreover, we will not forget that, ideally, such a procedure should be determined by internal documents in the official language - organization standards (STO) . And the graphical schemes themselves are a convenient means of designing a new order of work / visualization of an existing one (and they can be made out as an information and reference application to the service station).

(2) The availability of an up-to-date and detailed process model of work (in the form of a set of relevant service stations and / or a hierarchical set of graphic diagrams) of an already operating enterprise greatly simplifies:

  • making changes to the existing scheme of work,
  • redistribution of work and responsibility between various employees and departments,
  • setting the necessary control points (for example, “The secretary picks up the phone for 2-3 beeps”, “Orders the next day are accepted for processing until 16-00 the previous day,” etc.).

(3) In the case of designing a new enterprise, the development and visualization in the form of diagrams (“pictures”) of its main business processes allows :

  • imagine the future technology of work more fully visually,
  • identify any potentially difficult places, inconsistencies, set before yourself and then find the answer to various questions (“And what will happen ...?”, “And who will do ...?”, “And what time does this work need to be completed?” and etc.).
  • as well as in a graphical form, business process diagrams are more convenient to coordinate with all interested people and domain experts (visualization).

(4) The regulation of business processes allows them to be audited (“Does it all work as prescribed?”),

(5) Using graphical process diagrams facilitates the training of new employees and their adaptation, as well as avoids excessive dependence on The “know-how” of individual employees (“If he leaves, how can we figure it all out?”),

(6) The presence of formalized descriptions of internal processes is an important milestone on the way of implementing information systems and means of automation of activity.

Note. It is important not to confuse the primary and secondary. Yes, a computer program can only work on a clear algorithm. But one should not forget that “Automation is for business, not business is for automation”, and that many business processes of an enterprise are always more (certainly not less) areas of automation. So, first notes (technology) - then a tool (automation) , but not vice versa.

(7) Regional expansion : “having written” the technology of a successful enterprise into the model / system of internal standards, you can then replicate it to other, newly opened, or create a franchise.

(II) Process Models: Types


( II.1 ) From the point of view of relevance, the contents of the model are divided into:

(1) Model “As is” (Eng. “AS IS”) : reflects the REAL state of affairs at the time of the description, the actually existing, established technology of work.

(2) Model “As it should be” (English “TO BE”) : reflects the target state, which is supposed to be implemented in the future. For example, the model of work of a newly opened enterprise, or a new (completely new or improved old) procedure for performing any work.

(2) The “As It Should Be” Model (SHOULD BE): reflects the “idealized” state of affairs (according to regulatory documents), while the actual scheme of work in reality may be slightly different. In practice, the need to build such models is not common.

Note that the presented models of the same process can vary quite significantly. Example: a model of an adjustable pedestrian crossing, the traffic light switches automatically after a certain period of time.

“As is” : Some pedestrians are waiting for green and switch only to green. And some do not wait for the green signal, they look around and cross the road if, in their opinion, they are not in danger of getting into an accident. This is not worth doing, but in reality, unfortunately, it happens.

"As it should be"(because it is written in the SDA): All pedestrians are waiting for green and switch only to green.

In this case, the “As It Should Be” model could coincide with the “As It Should Be” model. However, they may not coincide if the model is “As It Should Be,” i.e. the one that will be recognized as the target will be one of the following:

“As it should be”: “Traffic light with a button” . Pedestrians approach the crosswalk and press the button - after a certain period of time, green lights up.
“As it should be” : No pedestrian crossing allowed.
“As it should be” : A pedestrian crossing will become elevated or underground.
“As it should be” : The street will become a pedestrian.

( II.1 ) From the point of view of the modeling methodand, accordingly, the scope of the result, we consider the following types of models:

(1) Functional models .

(2) Workflow models (worklow models) .

Functional models represent a “working principle diagram”. Those. that rye first sow, then reap, then thresh. Or that you first make parts, then assemble the product, then output quality control. Etc.

Today, perhaps one of the most popular functional modeling methodologies is IDEF0.. In fact, this methodology is a de facto “world standard” recognized both abroad and in the Russian Federation (see, for example, R 50.1.028-2001. Methodology of functional modeling). The description of the IDEF0 methodology is easy to find, including and on the web.

Enterprise modeling is often recommended to begin with the formation of a functional model. However, it must be remembered that such models are “static”; they are not intended, for example, to describe the step-by-step implementation of any working procedure. And they are intended to display the overall picture, the concept of work. As for the detailed step-by-step models of the implementation of any activity , models of workflows (worklow-models) are intended for this . And about them - below.

Workflow models (worklow models)allow you to describe the process as an ordered sequence of various actions, emerging events, as well as objects involved in the implementation of this process. It is such models that need to be built when you want to describe / design a specific process at your enterprise, for example, “The procedure for receiving goods to a warehouse,” “Rules for submitting an application for transportation,” etc.

Workflow models can be formed as at the next step in constructing a functional models - with its further detailing - and independentlywhen there is a need to describe (design) any particular procedure. In practice, the second way is also often used when the work starts from the “pain points” of the enterprise, or even the method goes “from the register of enterprise processes”.

As for the choice of a specific notation of worklow modeling (i.e., the graphic language itself), here, unlike functional modeling, the choice is quite large . These are IDEF3, “swimming lanes”, and ARIS tools, and BPMN methodology, and others. And each of these methodologies has its advantages.

As for me, I use the notation of "my own manufacture" - CH-1 notation. Frankly, when I was just starting to build business processes, I didn’t even think of writing any language there. But: starting with one of the standard tools, it turned out in real work (I repeat, it is in my case) that the means of the language used are not enough for a brief and complete record "from the words", and without data loss, the other turned out to be too complicated for employees .... And so, introducing additional characters and some changes, without thinking - without guessing, in the middle of the "two thousandth" CH-1 appeared. A few words about her below.

(III) A few words about the CH-1 notation


So, there are many notations of workflow modeling . And this should be considered as a plus: the principle of “let a hundred flowers blossom” is most welcome here. The choice of modeling notation depends on the task (and tasks are very different). For example, for the modeling of production lines, completely different modeling methods not mentioned here are used. So such a “methodological pluralism” in modeling workflows is not accidental.

If we are talking about CH-1 notation, then it is originally intended to describe processes in the form of a sequence of actions with an indication of their executors, related events, significant parameters of the performed actions and / or the process as a whole (for example, the normative duration) and emerging flows: material and informational. The Russian-language version of its specification (with examples) is laid out here: https://drive.google.com/open?id=0B_wUAIgOErG8MTQzYzJhNGUtZGY1NC00OTE1LWFlMzgtMDEyZmFjYTFjMDk3 , for more details see also here. Https://ch1-notation.blogspot.com (personal page about the CH-1 )

If we talk about software, then the symbols used in this (as well as in any other) notation can be found in almost any software that has built-in libraries of graphic primitives. Moreover, both in commercial software and in software distributed free of charge. In addition, to date, a number of software products have special character libraries CH-1.

At the same time, if the used software product supports the mechanism of hyperlinks (which is quite common), then using them is useful:

  • “Link” decomposition diagrams to decomposable works;
  • “Attach” to the diagrams the templates of documents used in the implementation of the business process, examples of their filling, links to guidance documents, etc.

As for the use of more specialized software products for modeling processes (not “sketches” and editors), here opinions can be different. Personally, they are not very convenient for me due to the large number of additional restrictions imposed. Although for some they will be more convenient on the contrary. Here again - “to taste and color (and task)” - there is no comrade.

(IV) In dreams of the future: automatic generation of regulations (machine translation “diagram-text”)


A diagram (scheme) of a business process is, in fact, its (proto) regulation in graphical form. Well, or (proto) the standard of organization, to put it more officially. All information about “who, where, what and how” should be there. And, accordingly, the question arises: “Is it possible to form text on a diagram with the click of a single button”? Moreover, the task can be complicated: the regulation (standard) should be in the "human" language, and not in the style of the "report" from a series of headings: "Who?" "What?" "When?" etc. and automatically copied to the appropriate sections of non-declining text.

If we talk about CH-1 diagrams, an algorithm for their machine translation into “human language” into the form of a finished draft standard was developed for them. The algorithm - not tested to this day (due to the lack of implementation in the code) - is located here:https://drive.google.com/open?id=0B_wUAIgOErG8bTluc2xYSVI4NXc , see also here: https://ch1-notation.blogspot.com (personal page about CH-1). I could not cope with the implementation of this in the code alone. At the same time, the two-step version is optimal when we have export from a graphic editor to, for example, xml, and then the generation of text in a text editor from xml. Such an organization is open in the sense that it is possible to "cut" more and more new graphic and text editors to it sequentially. It should also ensure data confidentiality, and, from this point of view, the open source software option is beneficial.

And in conclusion - about what? - about import substitution. It is clear that, at least for strategic enterprises, government agencies and other organizations for which information security is important, trusting even the "rendering" of their processes to imported software with closed source code may be unacceptable. Indeed, the process model of the enterprise, covering its internal processes, - that in the graphical (hierarchical set of diagrams), that in the text (in the form of organization standards) form, contains almost all the information about its internal structure - “take and build”, as they say . Therefore, the creation of such a completely domestic software product (such as the synthesis of “drawing” (chart editor, graphic vector editor) and “text generator”) is of particular relevance.

More articles:


All Articles