| NVNC.de | German Railway | Historical Data | Documentation | | Deutsch |
2.1. Basics on the Represenation of the Network
2.2. Lines and Segments
2.3. Routes
2.4. Operation Posts
2.5. Companies
This chapter introduces terminology and concepts used to describe the network. The terms used are rooted in the tradition of German railways and thus may differ from those used in Britain or North America. Where no simple translation was possible (or at least none sprang to the author's mind) the German term was kept. This seems especially useful since key words in the data are German.
Please note that some of the German terms are used with a slightly different meaning than usual. This stems from the fact that the project aims to describe a rather long time span. To keep things as simple as possible some concepts are used in a broader sense.
If one has to represent railway infrastructure using a digital apparatus such as a computer one cannot but use math. In the case of railway networks the research area is called Graph Theory and, despite its name, is rather practical.
A graph is a being of nodes and edges where each edge connects exactly two nodes but more than one edge may end at each node. That's pretty close to a railway network already if you assume the nodes to be junctions between lines and the edges to be the lines themselves. The only difference is that a line is usually longer than just from one junction to the next. Thus a line is a sequence of edges with the end node of one edge and the starting point of the next one in the sequence being the same node. This is called a path through the graph by mathematicians.
Furthermore, edges are not merely connections between nodes but represent real tracks layed on Earth's surface. Thus the edges can be described as a geometric curve. Every point on the curve has a unique distance from the starting point. Therefore, we can use this distance to identify every point on the curve. If the curve (which is an edge) is part of a line which is a sequence of edges, every point on the curve can be identified by its distance from the starting point of the line. It may be a little tedious to move along the entire line just to find a certain point but it is possible. This principle is used in the real world by the mile post system.
A curious thing is that not only points on the edges can be identified using this system but also the nodes. Technically, nodes are nothing but points on edges, i.e., those at the very beginning and end of the edges. Therefore we can identify a node by stating a line that travels through the node and its distance from the starting point of this line. Of course, there may be more than one way to identify a node: One for every line that travels through the node. This may look bad but it actually allows to define a new node whenever we happen to need one by merely stating the line and where it should be positioned on this line.
Finally, a note on negative distances. In real life they sometimes appear in result of some changes at the line's starting point. If this happens we actually would have to change all definitions for points on the line because their distance to the starting point changed. Instead of doing this, we call the starting point origin and position it in such a way that all other points keep their distances. With this, of course, some points may lay be behind the origin and thus we have negative distances.
Now, what edges out of the many possible make for a particular line? Since we do not devise a new network but try to represent an existing one we just have to have a look at this existing network. As it happens the railway allready uses a distance based system to find points on their network. We will exploit this system for our own purpose. At every junction we check which of the other edges at the node continues with the correct distance and add this to the line. Of course, this process is not straightforward and requires some inside into the history and present of the network. Whatever the choices are there is always room for controversy.
The lines created from this process are then assigned numbers. This means that a line is identified by a sequence of digits (this will become important later). The actual relation between lines and numbers is defined somewhere else. Note, however, that railway lines (as opposed to tramways or industrial railways) are always assigned 5-digit numbers.
The concept of lines is not exact enough, though. Many lines have not been opened at once but step by step over a period of time. Of course, we want to state openings of parts of lines as well. In order to do so, we need to be able to identify parts of lines uniquely. But first of all, 'part of a line' is a rather unhandy term so we will call it a segment instead.
To define a segment we can now use the concept of distance from an origin as introduced above. Every segment has a starting point and an ending point both of which have a distance from a lines origin. Taking these we can easily identify the segment. Further below we will encounter a second way to describe a point by using the key name of an operation post (a term which will be defined below as well).
If a segment is often used it is quite tedious to always state both the start and end points using their distance or a key name. Instead a named segment can be declared. The name of such a segment must consist of letters only. Using this convention a named segment can easily refered to by concatenating the line number and the segment name. If we have, for instance, a segment named Nd on line number 92100 references to the segment are simply given by 92100Nd.
Until now the model is simple and beautiful. Of course, it will not remain so. The problem is that in reality railway lines are sometimes moved to a slightly different route. This wouldn't be too difficult if the old route would disappear instantly and the new route would appear at the very same instant. This isn't the case. First permissions have to be aquired for the new route, then construction starts and some time later it is opened. The old route is sometimes used as siding or shunting track later on. It should be possible to describe all this properly.
What follows is that we have some edges that are not assigned to a line but somehow are part of it nonetheless. This is where routes (German: Trasse, pl. Trassen) come into play. They are used if there are multiple alternative edges between two nodes but only if the edges belong to the same line. Dedicated urban and long distance tracks usually form lines of their own.
In order to be able to define a route we need a segment of a certain line. We will use named segments for convenience's sake. Routes on such a named segment are numbered. There is a route number 1, a route number 2, and so on. To assign an event to a certain route we concatenate the route number to the segment's name. For instance:
92100Nd1
referes to route number 1 in segement Nd of line 92100.
At every time exactly one route of the possible routes is part of the line. So we can determine the route meant if only a distance from the line's origin and a time are given. All other routes may be refered to by using the full route name only.
Route numbers have to be ordered chronologically. The oldest route always has number 1. This way we can assume that when the line was first opened all routes 1 where part of the line without having to state this explicitely.
We have not yet reached the end of the route concept. If the railways are very mean the re-route part of a route. We want to be able to cover this, too. To do so, we simply use recursion by using a route name like a line number. You can define a named segment on a route by concatenating a segment name to the route name. On this new segment you can define routes again. And those routes again can be re-routed, ad infinitum. This is way line and route names may only contain digits while segment names are to be made of letters: Given a name as
92100Nd2Gc1Af4
we can easily determine the line, segments, and routes involved and find out that some railway with to many construction workers at their disposal has re-routed some part of a line for three times.[footnote: Their indeed is at least one line in Germany that has been re-routed that much, the Merseburg--Querfurt line near Halle (Saale) which crosses an area of extensive surface mining.]
Puting tracks into the countryside alone isn't very useful even if you can re-route them to excess. What you need is places where operation is controlled from. In German such a place is called a Betriebsstelle which rather clumsyly translates to operation post. There are many kinds of such operation posts but still only relatively few terms are necessary to describe them.
Every operation post needs to be identified. In theory that should be possible using its position in the network, i.e., stating a distance on a certain line. Unfortunately, posts are occasionally moved around without becoming a different post. Because if this the key name is used to identify a post. No two posts can have the same key name.
The key name closeley relates to the actual name of the post. In fact it is derived from the name assigned to the operational post as of 01 Jan 2000 or---if it was lifted before this date---its last name (exceptions may occur, or rather: are frequent). To turn it into a proper identifier some character have to be replaced. First of all, no spaces may occur so these are replaced by hyphens. German umlauts are replaced by the form with the normal vocal plus an e, e.g., ä becomes ae. The ß becomes ss.
Place names alone are not unique (think of names like Neustadt) and a number of addtions are used to make them unique, e.g., Neustadt (Weinstraße). We will not use the parantheses but append the specifier with a forward slash, i.e., we will write Neustadt/Weinstrasse. This way we can easily see what is part of the place name and what is a specifier. Sometimes such specifiers have been introduced without apparent reason or the specifier isn't necessary anymore in the smaller Germany of today. We will ignore the specifier in such cases to make the key name as short as possible.
Some block posts and junction posts only have an abreviation as their name, e.g. 'At'. Here we need to introduce specifiers of our own to make them unique. The key name of juction At will thus be 'At/Halle' but not 'Halle-At'.
Note also that a key name has to start with a capital letter but may contain small letters and digits later on.
In order to keep things simple, more specifically to avoid introducing another type of object besides lines, routes, and operation posts, the concept of operation posts is used a little wider than by the railways. We will also declare signal boxes and even points to be operation posts. But since they are not really, we call them false operation posts.
That we include signal boxes appears understandable. But why points? The reason is that the key name of an operation post may be used instead of a distance when stating positions on lines. In some cases it may be useful to do this for a point where a certain line emerges from another line, for instance, if the origin of the emerging line is exactly at this point.
There is reason to believe that operation posts are always situated at a railway line. Usually they have a certain length which makes it impossible to give their position by a single point. But we does anyway; the single point is given in relation to a certain well defined point. The position of passenger stations, for instances, is usually given relatively to the middle of the station building. Where a obvious reference point is missing a comment needs to be added to state the reference.
It has been noted in the introduction to this section that there are plenty of kinds of operation posts. Even posts of the same kind tend to be different. Both these facts are represented by the rank and properties of an operation post. The kind of post---or its type---is given by the rank. A post may be a station or a junction or a signal box. If a post serves more then one purpose it may have two or more ranks. A typical example is a junction that is combined with a stop.
Properties are additional features that a post may have or not. They may be operation related, e.g., that a post has no local staff, service related, e.g., that a post is not for freight, or equipment related, e.g., the manufacturer and type of a signal box. Not all properties are possible for all operation posts. The type of a signal box hardly makes sense for a stop.
A railway line is owned by someone and it is operated by someone. Due to the size of the operation this is usually done by companies rather then single humans. For simplicities sake we will also refer to government agencies as companies.
Just as operation posts companies have a key name to uniquely identify them. The key name never changes even if the actual name of the company does.
Large companies are usually split into smaller deparments. We will have to be able to cover this as well.
A concept for departments is currently being developed. In its final version it will be able to not only describe the various departments and what they approximately are doing, but also other offices and posts that are not directly related to railway operation.
| previous | Index | next |
Last updated on 23 Jun 2003.
Please mail your comments, suggestions, and complaints to
Martin Hoffmann <hn@nvnc.de>.