Opasnet:Guidance system development

<accesscontrol>Members of projects</accesscontrol>

Guidance system development contains specification and other information about the information system to be developed for carrying out and managing risk/impact assessments and to storing the products of the assessments as well as information that is useful as input for the assessments. As issues become worked out their implications should be transferred or built-in to the relevant content pages of guidance system, i.e. Guidebook, Toolbox or Data center.

Dimensions and hierarchies

The guidance system has two distinct dimensions. It can be confusing to browse through different parts. This is an attempt to clarify these issues. The hierarchy levels are listed from bottom to top for each dimension. The hierarchy list has dramatically changed; the original list was much longer and more complex. However, the old "dimensions" are captured in the new two dimensions. "Class", "real world", and "work environment" are actually examples of structures in the complexity dimension. "Meta" still exists but in a more rich form. "Guidebook articles" about glue (summary) articles, structured (according to the method) articles, and background articles is a practical grouping of objects inside and outside the method. "Resource Centre" is an application of the meta dimension. "Tools/Models" is an application of the complexity dimension.

Dimension: Meta (or Abstraction level)
1. Zero meta level is the real world: the real objects and events.
2. First metal level contains the descriptions of the real world (such as variables and assessments) and the collection and synthesis processes of the first meta level information.
3. Second meta level contains the general structures of the first-level objects, and the descriptions of these objects (such as process descriptions).
4. Third meta level describes the development processes and the descriptions of the second-level objects.
- For objects in the Resource Centre
  1. (Nothing exists from the zero meta level)
  2. Tools/models (software or programs) and core data sets
  3. Descriptions of tools, models, and core data sets, including links to those that are not within the Resource Centre.
  4. Descriptions and links to tool, model, and databases (meta-metadata).

Dimension: Complexity (or composition level)
- For product objects
  1. Variable: this level looks at individual product objects (typically variables) that describe particular pieces of reality.
  2. Assessment: this level looks at groups of variables that together form a synthesis of a particular problem at hand.
  3. Policy context: this level looks a particular problem as its relation to the policy context in which it is assessed. This level is usually not described explicitly.
- For process objects
  1. Method: a procedure for manipulating information (about the real world) into a useful form for a particular place in an assessment.
  2. Assessment framework: A distinct set of methods that, when used together, form a scientifically coherent and efficient way for manipulating all information that is needed to perform a particular kind of an assessment.
  3. Scientific context: this level looks at the assessment frameworks in their scientific context in which they are used. This level is usually not described explicitly.
- For Classes
  1. Item: an object (typically a variable or a class) that belongs to a particular class. For example, a pollutant concentration in a particular location at a particular time.
  2. Class: a set of all objects that share a key property, defined in the scope of the class. For example a class of "dioxin concentration objects" (in any spatio-temporal location) forms a class that can be useful when utilising the full chain approach for a dioxin assessment.
- For tools:
  1. Tools/Models that help to specify the result of the variables (e.g. Chimère, WATSON)
  2. Tools that help to perform the assessment as such (so dealing with the process of an assessment). These would be e.g. a tool for stakeholder involvement.

What are processes and products?

Fundamentally, a process is an event, or a composite of events, that produces something; a product. Product is an object which has been produced by some process, and has a potential of being used for something built in its structure. So, process produces a product, which in turn can initiate its use process. Processes thus use products (as their input) and produce products (as their output). Products have a history of being produced by a product and bear in them the potential of initiating its use process. Processess and products can be seen as different sides of the same coin. The same thing could (often) be considered from the point of view of a product: what is it, what produced it, what can it be used for - or from the point of view of a process: what is it, what inputs does it need, what does it produce. A process can produce several products and products can be produced by several processes. Similarly, a product can be used in several processes and a process can use several products (there does not necessarily have to be a 1-1 relationship between processes and products). It is merely a question of the chosen level of observation/description (could be also be referred to as granularity or resolution), that defines how the boundaries of the processes and products are set.

Variable is a typical product object. It can be used as an input of a process, or it can be an output of a process. It can – at least in theory – be validated against reality. In a variable, there is the Definition/Formula attribute, which tells how the result of the variable can be derived or computed. If the formula is a large piece of text, or generally usable in many variables, it might be useful to extract that piece out of the variable and make it a separate process object. This process of creating new objects from within an existing object is called budding. In this way, models can be seen as ancient Formula attributes that were created through budding from a specific variable in the evolution of the system.

Processes and products in risk/impact assessments

An assessment can be seen as consisting of the process of making the assessment and the product(s) that are produced by the process. The products are produced according to some needs (use purpose of the product), which defines the intended (desired?) structure (contents, appearance, form, etc.) of the product. The use purpose further defines, or at least sets requirements for, the structure of the process that produces, or is intended to produce, the product (which has the structure that meets the requirements set by the use purpose). In general it can be said that the the purpose of the process is to produce the product, whose purpose furthermore is to initiate (become a part of, fulfill its task within, ...) its use process.

When talking about risk or impact assessments the level of observation is such that the term product refers to the certain kind of description of reality within a certain scope. It is thus a composite information object synthesized of smaller chunks of information. The term process then refers to the (organized?) compilation of activities or events that produce this composite information object. They can be of types such as observation, information collection and synthesis. Furthermore, the term process can also refer to the activities of managing, or organizing, these above-mentioned activities.

The term process object may be used to refer to the structural parts of the process or their composites, including also the whole process. The term product object may be used to refer to the structural parts of the product or their composites, including also the whole product.

The risk/impact assessment product is information. The product and its parts are descriptions of reality, information about the reality. The risk/impact assessment process is producing, collecting and synthesizing information and management of these activities (let us call them sub-processes). The target of information collection and synthesis sub-processes is information about reality, the target of observation is the reality itself. The target of management sub-process is the composite of the other sub-processes. It can also be argued that management is a meta-process of risk/impact assessment process, that observation is a sub-process of risk/impact assessment process, and that only collection and synthesis sub-processes actually belong to the risk/impact assessment process as its internal structural parts. This is merely a boundary definition issue, and in any case it is easy to see that the above-mentioned structural parts of the process are hierarchically related.

Risk/impact assessment is thus (almost?) all about information. It deals with descriptions of phenomena, not directly the phenomena themselves. A variable (a part of the assessment product, a product object) is a description of a part of reality, not the part of reality itself. The process objects, such as editing a particular variable, on the other hand could be argued to be reality itself.

What does this all mean when thinking of risk/impact assessment information systems and their contents? Naturally, information products can be easily captured within such systems, but processes not (directly). It is obvious that the objects that can be stored within an information system are of the type product and that processes can only be built-in indirectly. However, the contents of these product type objects can of course be about either products or processes - descriptions of something.

A guidebook of risk/impact assessment should contain:

generalized description(s) of the product, and its structural parts, of risk/impact assessment (2nd meta)
descriptions of the risk/impact assessment process, and its structural parts, that produces such product (2nd meta)

A resource center should contain:

unorganized or untargeted descriptions of reality (1st meta, result of information collection and/or observation)

A result database should contain:

organized, targeted descriptions of reality (1st meta, result of information synthesis)

The above-mentioned parts of a risk/impact assessment information system are storages of product type objects. Perhaps, however, for clarity it might be good to reserve the term product only for the result database contents(?).

If we wish a risk/impact assessment information system to not only be a storage of information, but to make it functional, we should to try to build in some processes into it. This can, as mentioned above, only be done indirectly, i.e. by providing support of some kind for carrying out the needed parts of the process. In part the guidebook is intended provide this kind of support, but the system would also require a workspace where the process can be carried out, in part or wholly, and specific tools for information synthesis as required.

Information objects included in the guidance system

The content of the guidance system will be composed of:

pages describing the theoretical and contextual foundations of risk/impact assessment
pages describing processes relevant to risk/impact assessment
pages describing the products that these processes produce
pages containing data and information that can be used in/by these processes, or providing or access to them
pages providing access to tools
glue pages, such as lists of other pages according to different purposes and introductory texts

The guidance system content will be composed of structured and unstructured information objects. Structured information objects, i.e. those having a pre-defined format, include process descriptions and product objects, all other information objects will be unstructured, i.e. not having a pre-defined format.

The guidance system contents can also be divided into a guidebook, a toolbox and a data center. The guidebook contains the pages describing the theoretical and contextual foundations of risk/impact assessment and the pages describing the processes relevant to risk/impact assessment. The toolbox contains pages providing access to tools that can be used in/by these processes. The data center contains pages describing the products that these processes produce and pages containing data and information that can be used in/by these processes, or providing or access to them.

Structured information objects

Processes relevant to risk/impact assessment are captured into the guidance system as descriptions of these processes, possibly also including (descriptions of and/or access to) the tools relevant to the process and descriptions of how to manage the process. A process object can be e.g. in the form of a method description (calculation, formal discussion, etc.) describing also the tools, models and other means that are used in applying the method. Process descriptions belong to the Guidebook.

Product objects in the guidebook are pieces of information that are descriptions of reality produced by the risk/impact processes. Product objects belong to the data center.

Article templates

Template: process description (a guidebook information object)

Summary

The summary of a process is a very short overview of the process, and may contain all types of information that are considered relevant for this specific process (max words: 250?)

Name

The name of the process should be unique (there should not be two objects (processes, products or product descriptions) with identical names). The name should be chosen so that it is descriptive, unambiguous and not easily confused with other processes, products or product descriptions. (max words: 20?)

Purpose

(This is the same as Scope in the general object attribute list.) The general purpose of every process is to manipulate information in aim to produce a particular information product. The process-specific purpose describes the intended product of this particular process. This should contain all relevant information needed to distinguish the process from other processes. (max words: 400?)

Structure of the process

(This is the same as the Result (of the process description) in the general object attribute list.) The result section has three sub-attributes. (max words: 3000?)

Input format

The input data, variable or parameter types, their syntax, and other relevant information is described here.

Procedure

The actual process – how it works and how to do – is described here. It consists of e.g. the mathematical formula to calculate the result. The procedure uses algebra or other explicit methods if possible.

Management: A complex process may require that also the management process of the procedure is described. The procedure may for example be a mathematical algorithm. The management process is a computer software that runs the algorithm, so that when you have the software, the management is a trivial task. If you do not have the management process - well, good luck to you and your calculator. So, actually models and software are procedures that have the management process packed into the same neat package. An other example of a management process is a guidance by the U.S.EPA about how to plan and organise for stakeholder meetings. The actual procedure here is to obtain information and feedback from the stakeholders.

Output format: shortly describes the format of the product of the process, which should fulfil the purpose as described in the scope. It furthermore links – if applicable – to the product object.

Rationale: (This is the same as the Definition (of the process description) in the general object attribute list.) This attribute answers to the following questions: What is known about a good process for this purpose? How do you know that the procedure described is good? (max words: 2000)
See also: See also links to pages (both internal guidebook pages and external) which relate to the process subject. All subjects that could be relevant for readers of this page can be listed here. (max links: 20?)
References: All references, as used in the texts above. (max references: 30?)

Template: product (a structured data center information object)

Summary

The summary of a product is a very short description of the product, and may contain all types of information that are considered relevant for this specific product (max words: 250?)

Name

The name of the product should be unique (there should not be two products or processes with identical names). The name should be chosen so that it is descriptive, unambiguous and not easily confused with other products or processes. (max words: 20?)

Scope

The scope of the product gives a research question that this product object aims to answer. The scope includes – if applicable – its spatial, temporal, or other limits (system boundaries). A product object may also be a generic one applying to any spatio-temporal locations. Whether these are just variables, and how they are used in the Guidebook is under study. (max words: 400?)

Definition

The definition describes the data and reasoning that tells us what the answer to the question in the scope is. In the case of a variable, it has four sub-attributes:

Data: What data or observations are available about this object?
Causality: What variables affect the result of this object when they change (i.e., which variables are causally related to this one)?
Unit: What is the measurement unit of the result?; Formula: How is the result derived or computed?.

Definition may link to the process(es) that lead to this product, explaining shortly why these processes are relevant. (max words: 3000?)

Results

Contains the answer to the question presented in the scope. Usually the text is short (max 500 words), but the result tables or figures may be extensive; there is no upper limit.

See also

See also links to pages (both internal resources centre pages and external) which relate to the product. All subjects that could be relevant for readers of this page can be listed here. (max links: 20?)

References

All references, as used in the texts above. (max references: 30?)

Practical information for both templates

----#(number):: . General information on how the sentences should be structured etc. (see also my first version and the project that RIVM already have conducted. --Alexandra Kuhn 12:50, 14 January 2008 (EET) (type: truth; paradigms: science: comment)

Any section may contain a "more" button to more detailed information that is hidden by default to increase readability. The content is still directly relevant for the section. In addition, if there is a lot of further background information available about the object, e.g. its history, current practice, etc.) which is useful but not required to utilise the object, a background article can be created and linked to from the object. A background article can be freely structured, but an established encyclopedia article structure is recommended.

Process/ product structure in Intarese

Elements of the Intarese method as developed in SP1 relate to either a process, or a product, or both. The graph below gives an idea of processes and products as developed in the Intarese project (graph should be improved/updated, if we want to present such a graph!) . For all processes and products, information needs to be provided in the guidance system.

When there is a clear 1 – 1 relationship between process and product (eg DALY process – DALY product), we will only ask for description of either process or product, in order to avoid confusion. In most cases, the process is the important object that needs to be described in the method guidebook. When there is no such direct 1 – 1 relationship, we will ask for separate description of process and product. An example of this is the process meta analysis, and the product exposure response function (ERF). Even though the process of meta analysis can lead to an estimation of the ERF, the meta analysis can also lead to an estimation of another product (e.g. severity weight), and an ERF (the product) can also be derived from another process (e.g. expert judgement).

How should an article about disability-adjusted life years be structured? The question is not at all obvious, so the different options were tested in a previous version of this page. Only the conclusions are presented here.

In this ontology, the scope can be seen as a research question. The answer to this question is the result, and the definition tells how this result can be or was achieved.

Approach	Scope	Definition	Result
Method (process description object)	DALY estimation is a process for measuring summarised burden of disease adding up increased mortality (years of life lost) and years lived with disability. Purpose: DALY translates the impacts of diseases into life years based on their severity and duration, so that different diseases can be measured using a single currency, the life year. DALYs are based on disease-specific weights. (In contrast, QALYs evaluate the quality of life in a certain health state, not disease.) The research question: What is a good way to estimate DALYs?	The definition contains the reasoning and motivation for calculating DALYs. It compares and discusses different alternatives. It contains links to methodology articles. It is also open for comments and further developments.	The result describes the state-of-the-art method and formulas. Formula: DALY = life-years lost YLL (due to mortality) + life-years with disease YLD = YLL + number of disease cases * severity weight of the disease * the duration of the disease. The severity weights for diseases come from the variable Variable:Disability-adjusted weights for diseases The result is motivated by the content of the definition. Sub-attributes of the result include input (the upstream variables), procedure (the calculations), and output (format of the process output).

Conclusions:

The process description (method) seems to be the most suitable object for DALY. The result attribute of the DALY process description describes the state-of-the-art procedure for estimating DALYs. If there is not a single best method, several methods can be described. The definition would then discuss the good and bad properties of each method, and their limitations. Usually a process is a general one, and therefore it will produce a large number of products for specific purposes in several risk assessments. There is no need to describe a "generalized product", as all general information is already described in the process description.
The purpose of all process descriptions is to describe a good process for achieving the outcome described in the scope. Thus, scope belongs to the purpose of a particular process description object.
The definition of a process description describes the information you need to understand whether some procedure is suitable and good for the purpose. The definition gives rationale for the procedure selected.
To be precise, a process description is a product object that describes how to actually do the work. The doing itself is the process object, but that is something that vanishes as soon as the work is done and the product has been produced. We try to be precise when talking about the process or process description. However, we do not emphasize their differential nature (process vs. product, respectively), because the difference between doing and talking about doing is probably clear to the reader until someone tries to define that difference using metaphysical terminology.

Example: products and processes in the Intarese guidance system