Make your own free website on Tripod.com

Information or Knowledge Warfare?

Copyright by Josip Pajk, December 1997.

A widely accepted conception of today is that we are faced with a Third Wave Warfare (1). Information nowadays is not only the objective, but also the main fighting instrument. After the Agrarian and the Industrial waves of warfare , in which unsteady energetic and rigid material structures has had the leading role in conducting wars, in the Informational wave of warfare the main performer is the adaptable informational system's structure. According to another classification (2), in the Age of Tools the driving power for simple warfare material were muscles (humanís and animalís) and other natural, raw energy structures like wind. The Age of the Machine perhaps initiates when the more and more sophisticated warfare material of that time commences to move autonomously by the aid of new synthetic energy structures discovered by thermodynamics. About the same time, due to the developments in the communications field (railroads, telegraph), begins the Age of Systems, in which the new machines, now capable even to fly and dive, could be mass engaged and coordinated on a much larger scale and territory. Todayís Age of Automation made its appearance when the evolution of such systems (in terms of numbers and sophistication) came to the point that they could no more be controlled in a traditional manner (by direct human influence). Automation was first applied to the lowest, regulation system's control level, cybernetics extended it to the control (operational) level, and today, with the dawn of artificial intelligence, we are faced with the first signs of automatization even on the systemís guidance (decision making) level.

[table.gif]

To understand the essence of these historical Waves or Ages, regardless of how they are named or classified, the first necessary step is to give a look at the basic structures that make up every dynamic system (natural or artificial) including the military ones. It would be rather wrong to state that, for example, information was of no importance in the Agrarian or Tools age. Even in Sun-Tzu's The Art of Warfare the "perfect victory" is that gained in a war that never put in action the material and energetic part of its warfare machinery. The basic goals of warfare art or science have not been changed throughout the centuries, so that even today all serious military analysts are more than ever aware of the inevitable threats and losses connected with battlefield operations. The information reliant modern systems give the possibility of using information as a "non-lethal weapon" in an all Information Warfare, but other, more "aggressive" analysts consider such a doctrine, with good reason, as a "paper tiger" if not supported by lethal, material and energetic structures. But, can things be such simple?

Structures and systems in history

In any system it is possible to identify three basic structural types that makes it:

The energetic structure (see me) , composed of consumable resource elements (energy) that can be transformed or recharged from the system's environment. They are the "life power" needed for the system survival and operation. Elements of this structure are unsteady, changeable categories very difficult to control;

The material structure (touch me), composed of visible, tangible and predictable three-dimensional physical system components, making it consistent and defining its boundaries. Consequently, all those stable, rigid components which alteration is very difficult;

The informational structure (feel me), composed by invisible elements like are knowledge, skill, competence, or goals, objectives, aims, which are not rigid as the material elements, nor unsteady like those energetic, but flexible and adaptable. They are influenced by, but also dictate the shape of the other two structures.

The elements that form the informational systemís structure are in fact elements that forms the model of a real system or process, and could be found only in another (dynamic) system, involved in some kind of process (operational or analytical) with the first system. For the storage (canning) and directioning (canalization) of the ever-changing energetic structures, the system needs rigid material structures. Moreover, for any modification made in the material structure it is necessary to spend some amount of the system's energetic structure (resources). How will these modifications take place, with what results, when and where, is dictated by the systemís informational structure.

Energy

[labile.gif] In the Agrarian Wave of Warfare the attention evidently had to be focused on the energetic structure of warfare systems, i.e. quantity. Season warfare between two harvests was prevalent at that time. Wealthier the state, the better equipped was itís army. This was a prerequisite for battlefield supremacy. The structure of such systems composed mainly of nonprofessional warriors (peasants) was unstable. The informational structure was also fluid, based on diffuse channels. A risen flag was enough to put in motion thousands of people, without practically any further possibility to influence their future behaviour on the battlefield. The elements of such systems are kept together by only one common goal (appertainment). There are no complex operational goals that could be fulfilled with this kind of structure. The opponent system thus can not be completely defeated. It is exhaustively engaged on a large scale and territory, by low-intensity, unstructuralized stochastic attacks very similar to corrosion. Agility, unpredictable behaviour, and deception are in the basis of this warfare doctrine known as guerrilla or partisan war.

Matter

[rigid.gif] The necessity of small nations (states) to withstand the corrosion imposed by such a doctrine, caused the design and production of rigid material structures that could protect their weak energetic structures and increase their power and operational range in a greater measure than the simple connection of man-animal muscle power. So, they start to construct material structures as body-armour, fortifications and mechanical systems like ships and catapults. Such rigid structures could protect their limited energetic structures and coordinately aim them towards the fulfillment of their particular goals. It is not by accident that the Industrial Revolution (3) started in Europe where small nations like Greece and Rome (in the Agrarian age) and then England, Spain, The Netherlands and France became world colonization superpowers. Rigid communication structural elements were built like roads, and then (by the invention of the steam engine) railroads. Feudal, vassal (unstable) social organization was replaced by monarchy (rigid hierarchical structure). The military structure was professionalized and hierarchically organized. In this structures deception was not honorable. Knighthood rules were applied in all fightings, from duels to battles, so even the informational warfare structure was rigid and predictable. The culmination of such a rigid system structure was in the absurd of the First World War trench warfare.

The described rigid structures can be schematically represented as well-connected standardized and specialized elements which form a rigid hierarchical structure with directed (point-to-point) communication channels. This is a very predictable (deterministic) and sluggish massive structure, but more powerful because of the timely coordinated operation of its elements. Control is applied by aid of commands, and all control elements, except those on the highest levels, are set to the regulatory level in the trivial role of command signals executors.

Information

[flexi.gif] During the First World War, the fact already known in the Agrarian Wave was recognized again: agility and manouvre are the necessary components of military operations at all levels. The rigid mechanical structures (machines) needs mobility and the possibility of communication if they had to be effectively controlled. The increase of distance between the more and more sophisticated system components, made possible by new telecommunication technology, resulted in the increased amount and speed of change of the data needed for controling those systems. This demands was far above the human capacity. Machines, for the first time able even to fly and dive, unlike the direct control of manpower, was already requiring indirect control. The control data had to be translated from machine to human understandable code and vice-versa. So, it was reasonable to leave the all-machine code on the lower control levels (regulation), and rise the code translation to higher, symbolic, control levels. On higher control levels the demand for a timed (real-time) human reaction is lower. The process is monitored in an "extended real-time" needed for the prediction of the system's future behaviour. Some kind of modelling of the real system's and processes' informational structure had to be made with elements of lower material and energetic potential. For these purposes it was necessary to have enough knoweledge about the system, which was not a demanding task for the machines (mechanisms), because their rigid informational structure was already known in the process of their design. However, the other two dynamic system types included in this complex systems (organisms and organizations) also had to be modelled, so cybernetics (4) established a common analytical model for the identification of the informational (controllable) structure of all dynamic systems, regardless of their particular material and energetic characteristics.

The structures of systems made from mechanisms (machines) and organisms (humans) established in an adequate organizational system (social, military, etc.) became more flexible, capable to respond to a variety of new process demands in a number of different ways not constrained with their natural (inherent) limits. Only in such structures control can be optimized because there is a possibility of choice (5). The same elements, which in one moment are part of a rigid structure, in another situation are able to operate autonomously. Alternative communication channels in communication networks ensure the consistency of the system in almost any situation. This structure is at the same time, and to the same extent, predictable (deterministic) and stochastic. Each element of this structure is completely aware and competent (has the information and knowledge) for the operation in its environment in accordance with the goals set from a higher level and its frame of competence.

Knowledge

Thus, we come to the notion of knowledge, recognition (identification), namely, the continual process of making new, and adjusting old informational structures i.e. their modelling. Today's virtual organizational systems exists on a global scale. Their material structure is spreaded all around the world (offices, individuals), the energetic structure (capital, labor and other resources) is founded either on the local level or circulates in a form of non-material transactions, and their informational structure is, although being complex, transparent, flexible and adaptable. Military systems are not immune to such changes. Rigid military organizational structures are disintegrating, leaving the place to another, more flexible way of mutual connection, especially in the informational domain. It does not mean that we are on the way of building an all-informational society. It must be emphasized that no informational structure can exist without an appropriate energetic (resources) and material (technology) structure. Technology even today makes possible to conduct a war like video game are played. But, unlike video games, in a future real war every mouse click will have far more serious consequences.

Some authors recognize that fact, and are aware about the danger of such a simplification in the doctrine of warfare. The guerrilla way of fighting which characterized the agrarian or tools age, make its appearance in all other eras, and is not likely that it would not be applied in future wars either in a classical way, whith structures involving muscle power and hand-made tools (weapons), or in the form of cyber terrorists who, from their apartments anywhere in the world, would, by means of low-cost (handy) computers and modems successfully corrode economic or military superpowers.

The essence of information and its role in the system

At the very beginning of the modern telecommunication systems (information theory) evolution there were trends towards the cleaning of the newly discovered information from the noxious influence of energy and material needed for its handling. The reason for that endevours could be found in a fascinating fact that with a minimum of energetic resources enormous material and energetic potentials could be controlled. However, as the Holy Trinity is the union between the Father, the Son and the Holy Spirit, the informational structure could not be detached from the other two structures, material and energetic. An idea, or a belief, as the best example of an invisible informational structure element is causatively and consecutively connected with the person as a system and his/her material and energetic structure. In the same way, ethics, moral and religion are connected with the society and its hystory as the example of an organizational dynamic system. Another example could be the bondage between a doctrine and a military system. A computer, like any other mechanism, is nothing more than a number of well arranged components if not provided with an adequate software, or at least a instruction how to programme it.

Another wrong opinion that often makes confusion in information related debates is that the equipment which in fact serves only for information gathering is often presumed as a source of information. Radars are, for example, very often labeled as sources of information, although they are just one of the elements that forms the information handling channel. The radar is nothing more than an observer which transforms the state of the source of information which is actually in the real process (the presence of an aircraft) into the data that will only at the destination be transformed into information needed by the system. The information is generated by a structural change of state at the source in the real process, and it must generate a structural change of state at the destination in the system observing that process. The comprehension of a structure and its state exists exclusively in the system that handles the data related to that structure in order to form an informational structure (model) of the real process or system. This modelling process alters its informational structure (knowledge). What is thus the nature of information and what is the difference between information and other related notions (data, message, news)?

Communication

[chnl.gif] The presence of an aircraft is a fact which is described by data as are: its position relative to the observer, course, speed, etc. These data are arranged into words and sentences (statements) according to the syntax rules agreed in the information handling channel in order to allow the understanding of the message at the destination. The message is a statement prepared for transmission through a channel and it includes all the elements characteristic for the chosen communication channel. The message is a statement in an envelope. The envelope could consist of the destination and sender addresses, encryption protection elements, i.e. all standard elements that are not connected with the fact detected in the real process. The difference between news and message is only in the type of channel used for the transmission. The message from the source is transmitted to a known destination (point-to-point), and the news are transmitted by a diffuse channel from the source to unknown destinations (point-to-multipoint). Finally, information is that part of the message unknown at the destination until the moment of its reception, thus information must increase the ammount of knowledge at the destination. When this structure is applied, for example, on a known IEEE (ISO-OSI) 802.3 standard of communication, it is evident that information exists only at the highest (user) of the seven communication layers, regardless of what structures and inner processes are applied to transport it from the source to the destination.

Process observation

If the system is used exclusively for surveillance (monitoring) purposes, the information path is temporarily broken at the inner destination (i.e. command post). The fact (change of state) in the real process (source) resulted with the change of state in the situation picture (model) of the process at the destination. This fact does not necessarily has to generate the response of the system. But, dynamic systems can be identified with regards to their response to a stimulus from their environment (traced process). In this case both the source and the (final) destination of information are located in the real process. The identified fact in the real process generated the response (effect) of the system to the process state, which is again information for the system. This continues until the real process is brought by the influence of the system to the desired (stationary) state (without changes). In order to be able of controlling the system, some special types of data (commands) are needed. Their amount is defined in advance, and they are used to form sentences by means of which the system is brought into the appropriate state. It is important to emphasize that there are no other possibilities of influencing the process apart from those resulting from the change of the systems structural states.

Evidently processes (inner and outer) could not be separated from the system. Processes are generated by the system's material and energy structures according to the aims that are included in their informational structure. Only a system with clearly defined objectives is capable of tracking a process. Tracking a process means identifying the organizational degree of the system (structure) which supports it, namely its informational structure. Combat processes, as well as those in courthouse, are generated through the conflict of two systems, with the goal of disintegrating the structure of each other (especially those informational). To fulfill that goal it is indispensable to identify the weak points of the opponent structure and to have a knowledge how to attack them in order to achieve the best effect, that is, to have information and knowledge.

Identification of material (rigid) structures is done with active (energetic) methods. It is necessary to spend some amount of the system's energetic resources to induce the reaction of the opponent system (radar or laser systems). On the other hand, the energetic structure of the opponent system can be identified by passive methods, i.e. by rigid (material) structures like are IR, radar, or laser detectors. That is made possible due to the inevitable changes in the environmental state made by the presence and operation of the observed system.

Control of dynamic systems

ctl.gif

The relation between a dynamic system and a process can be represented in a simple form as shown besides. It is visible from the figure that monitoring (surveillance) is a constantly applied process, and control is composed of actions that are applied when the process has to be put back within the desired limits. Often, even in cybernetics, the control and the controlled part of the system are separately analyzed. It is maybe legitimate to apply such a method in rigid system structures where the elements on lower control levels are only the executors of commands from the upper levels, but in flexible systems it seems more appropriate to analyze the system as one (holistic) structure in which three control levels can be identified: regulational, control and guidance (see next figure).

The control levels defined in this way can be directly applied to military systems, namely to their tactical, operational and strategic levels. Complex and flexible organizational systems are established from relatively independent dynamic subsystems in which these three control levels can be also identified. This means that the elements of strategic behaviour can be found even at the lowest organizational levels. The platoon commander, for example, even if executing the commands given from a higher level, is responsible for maintaining the capability of the system in his competence of performing the given tasks, consequently he has to plan and predict the evolution of state of his subsystems and the systemís environment.

Regulators from the tactical level are rigid structures, what is understandable concerning that they are in the middle of the real process together with the process instrumentation (sensors and weapons) which they are regulating. Their main task is to keep the process (the number of process parameters they are tracking) in a narrow band of tolerance. The main characteristic of the elements on this level is reliability, namely the security that they would perform their task in a specified manner and on time.

The working parameters of the regulators, their place in the system and tolerances are set from the control level which has to posses a complete knowledge of the attributes (constrains) of all the regulators under its control, and their current state. It is important to underline that all information on the control level must come through the regulatory level, and that the control level must not have any possibility to control directly the process instrumentation. Every regulator tracks the process only in one part of the complete vector of process parameters. A better (more reliable) and wider picture of the process state can be only obtained on the control level with the classification and correlation of data concerning the current state of each regulator and the data concerning the part of the process they are tracking. The optimization (6) of the control processes implies the use of regulators in such a way that they are relatively equally loaded, deployed and exploited with the best possible effect on the process.

The process (or better, the system) is directed by means of objectives which are arriving on the control level from the strategic, guidance (governing) level. The control level is responsible for supplying the governing level with patterns of the process alteration in time, which are generated by tracking a great number of process parameters through an extended period of time. The obtained information are used at the governing level for the prediction of the systemís (process) future states, and generation of objectives that must be achieved by the system in order to give an effective response to the supposed future requirements.

Any "bypassing" of any control level has disasterous repercussions for the system. If the regulators are controlled directly from the governing level the whole system is made either unstable or rigid. A system without a governing level is reactive, hence it is not pro-active. The bias of building weapon systems of the "detect and destroy" or "sensor-to-shooter" type are manifestations of such convictions. For the sake of a faster system response, one entire control level is truncated. With such crippled structures it is, for example, impossible to gain a better effect with a coordinated timed attack at the same target with a number of different subsystems.

Thus, if the system has to be flexible, all three decision making levels are needed, in all subsystems, and on all organizational levels. It is in some way the delegation of responsibilities to subordinates. However, even today, the technology that made possible such way of organization is used as a mechanism for making systems more inflexible. The image from a TV camera, for example, can be transferred in a "live" or "frozen" form by such technology without any problem from the battlefield to a remote command post. This possibility is used in a manner that expensive imaging systems are built "to make possible the remote assessment to the commander". The question is: Who can better assess the situation on the battlefield other than a trained professional soldier which is already on the battlefield in a direct contact with the ever-changing process? What information is of greater importance to the commander at the remote command post; a picture that has to be processed in order to extract its limited and obsolete information content, or a flow of fresh quantitative (numerical) data gained from a complete assessment made by a subordinate commander on the battlefield, who has a better view of the situation and the knowledge for making the assessment? But, for these purposes there should be a professional soldier deployed on the battlefield, not a camera operator. Because of the more and more sophisticated electronic equipment to be used on the battlefield, the considerations that the soldiers of the future would be more computer operators than classical soldiers have became part of nearly every futuristic discussion on military affairs. If we make a digression to another everyday field on which a battle for profit is fought: a good broker on a stock market is not a computer expert, or an operator virtuous, but the one who has the best knowledge to find among the large amount of data on his display the information necessary for the fulfillment of his systemís goals. In other words, the information has to be recognized among the available data, and this is impossible without the right knowledge how it can be used in the most effective manner.

Knowledge in dynamic systems

Unlike systems and subsystems who communicate with each other with input-output variables, namely, with their manifestation (operation) in the environment, structures that form any particular relatively independent system communicate with state variables, i.e. are much more firmly interconnected and mutually dependent. The state of the material and(or) energetic structure of the system has a direct influence to its informational structure and vice-versa. The changes in one structure certainly are followed by changes in the other two structures. The structural states and inner system processes are very difficult to identify from the outside of the system. The identification is commonly done by analyzing the system behavior, i.e. its response to known input signals (time dependent variables). The system in these cases is represented as a "black box".

The model used in the Theory of Systems for dynamic or causal systems, i.e. systems with memory, fits very good in this discussion. All dynamic systems (mechanisms, organisms, organizations) can be pictured and explained with such an informational structure.

The system is under the action of a multidimensional input signal vector U(t) from the environment, on the basis of which the system generates a response or output signals Y(t). For simple, rigid and static systems it is all. The same input signals will always produce the same response of the system. But, in dynamic systems (with memory) U(t) with the time passing also produces a change in the (structural) state of the system, X(t), which changes, how the functions (inner processes) of response generation, so the functions for the production of the variables V(t), conditionally here named as learning, which affects the processes of changing the state of the system (memory or knowledge). It means that the response of a causal system is not solely affected by the current state of the input signals, but also by the way in which they have been changing in the past. In other words, it means that a dynamic system will not necessary respond in the same way to the same input variables because its state is changing over the time. What does it mean in the example of a military organizational system and its informational structure?

In reality, a military system, as a dynamic organizational system, in all the previously mentioned historical periods, could be represented with such a structure. The key process in the identification of the informational structure of any dynamic system is the identification of the two internal vectors (V and X) connected with its state. It is important to mention at this point that the input vector (U) consists, either from directed signals from the environment and other systems (direct threats), and undirected signals (noise) which also affect the system. It means that the system "feels" the state in the environment only through the vector U(t) in which are included also all the possible disturbances from out of the system. All the disturbances and noise caused by the system malfunction without any visible outer influence, are of other nature and are included in the vectors V and X. The input vector U could, in addition, have some signals which does not affect the system in any way, because the system can not recognize it if there are no sensors able to detect those signals. The target for a system does not exist if it is not in a range of its sensors.

In the same way the content of the output vector Y are not only the material and energetic manifestations of the system in combat operation, but also all other (informational) elements that are on its disposal for the operation (action) towards the environment and other systems.

Unlike the vectors U and Y which are connected to outer processes' states, the content of vectors V and X are data connected with inner system processes. V(t) is composed by the information variables needed to prepare the new system state. Those information are generated in a process of "learning". The environmental state U(t) is here analyzed and information are gathered according to the present state X(t) of the system. It is visible that the vectors V and X can change even if the input variables are constant. In practice that means that a dynamic system in iterations adapts its state with the state in the environment, it is self organized. In the vector X(t) are included also the short-term and long-term goals of the system. If there are no new information V the state of the system could not be changed, its behavior is static. The memory part of the system structure is different from other elements because it has only one input and one output (it could not be parameter controlled). What does it mean? The change of V(t) may not necessary immediately have a consequence in the change of the state X. Information could be accumulated (tracked) for a given period for a better definition of an appropriate future change of state. In fact the systemís state could be changed even if at that moment there were no changes of the vector V. The memory part of the system generates the new state by a long term tracking of the information gathered by the learning part of the system to prepare the system for an adequate response to the expected signals U that have not been yet applied to the system.

When such a structure is applied in the analyze of systems, and when the content of each vector is correctly identified and quantified as a functional dependence between the variables, for the modeling or simulation of the system behavior there are powerful mathematical tools available. But the purpose of this writing was to emphasize the essential difference between information and knowledge in dynamic systems.

The state of outer processes U(t) acts on the system in two ways: induces a response Y(t) according to the state X(t) (7) and at the same time affects the state changes by the aid of new information V(t). The vector V(t), in fact, is composed by information, i.e. the way in which the system interpreted the state in its environment. V(t) is a picture of the real state in the environment. The way in which the system will interpret the real state of the outer processes depends of its inner state X(t), i.e. its knowledge. If the system has no knowledge that could enable it to correctly interpret the state of the real process, the system can not gather any new information V(t) which could change its state (the quantity and quality of its knowledge).

Now we can make a conclusion: information in a system is, not the fact that happened in the real process, but the picture that the system made by observing the process, which is in a tied connection with the state of the system or the knowledge it possess for a correct understanding of the event for the upgrade of its knowledge and an elaboration of a correct response. So, we came to the question made in the title:

Information or knowledge warfare?

The times when the commanding officer was able with one look from the nearby hill to be aware of the state on the battlefield, are far ahead. This privilege for the time being have only the commanders on the lowest commanding levels. Even they scarcely use their natural by natural evolution developed "sensors" for battlefield awareness. They make their decisions on the base of information gathered from the data given to them by the synthetic picture of the situation on the battlefield. In such a situation they "must" believe that the symbols they see on their tactical displays are real targets that have to be treated. As far as there is such a confidence the system works. But what if there is no more confidence, if on the display there are targets that do not really exist, or if a real threat is not identified and displayed in time? It's the time when knowledge appears on the scene. No matter how technologically advanced, the system is rigid if in it there is no possibility of some kind of selection. The use of technology increases the human capabilities far beyond those given to him by natural evolution. But, per definition, each technological achievement is a rigid (known) structure that can be very easily identified and, for that reason, its "weak points" founded to degrade or make impossible its operation. Flexibility means dynamics of the system, i.e. the ability of the system to change its structure according to the controled processes change. When concerning technology (rigid, material structure), flexibility can be achieved by diversity and quantity. It is very easy to jam one radar, but if a surveillance system relies on a great number of simple radar, optical and electrooptical systems working in different frequency bands, spread all around the territory, the picture of the real process generated by the data from such a wide and diverse base will be undoubtedly better than the one relying on just one technology.

The technology for itself is not enough warranty of system effectiveness. The believe that some system is better because of its advanced technological base is a mistake that very often have to be paid badly. History is full of such examples. The knowledge how to use the system in a most effective way according to the situation is, thus, essential. With the appropriate integration of systems that are on a lower technological level, in a regular base better effects are achieved. As a rule, the simplest the solutions the better they are. The known anecdote of a ship that was threaten by the getting near of an abandoned mine and a "silly" idea that all the crew have to blow on it until the ship passes by (8), give a good picture on how information and knowledge affect the solving of dubious unpredictable situation.

The knowledge how to use in the most effective way the technology that is on disposal at the moment is the element that gives supremacy. For those purposes competent professionals are needed on every organizational level. But not those professionals whose power is the power of technology, and will be paralyzed if the technology fails. Flexibility is a delicate game between lability and rigidity. Surprise, as an important element of warfare hardly can be achieved in a labile, and especially in a rigid organizational structure. For the advance to the "new frontiers" knowledge is crucial for recognizing and overcome the present ones.


Footnotes

(1) A. & H. Toffler

(2) M. van Creveld

(3) If the appearance of the first standard and uniform warfare material produced in specialized arsenals is picked out as the start of the Industrial Revolution, its beginning can be moved far back to the ancient Babylon and Egypt.

(4) N. Wiener applied his cybernetic methodology for the first time during the Second World War for the purpose of automatic tracking of aircraft, since their evolved dynamic characteristics made manual tracking and aiming the main source of errors in air -defence systems.

(5) Choice (selection) of the best solution, equipment or co-operator is the essence of control (management).

(6) Optimization is possible only on the control level. There are no such terms like "optimal regulation" or "optimal guidance".

(7) The reaction could be omitted if the systems' state dictate so.

(8) The ship was actually rescued by the aid of water jets from the shipís fire-hoses.