ponents which permits their identification.
In all these cases, the Identifier-word must be considered as data, it can not be considered as address due to the possible discontinuity in the ascendent order of the code.
Advantages of alternatives 2., 3. and 4. are:
- They need a lower additional number of bits per component in the computer memory to define all data which identify it, because they are themselves data-words.
- They could permit us, at first glance, to have directly some information either about the component’s type, or about the diagram to which it belongs, or about the posi
tion of its symbolic representation in the drawing, etc.
167 -depending on the code used.
Disadvantages of these alternatives are:
- The Identifier-word requires more bits, due to the
prac-T D
tical impossibility of using exactly all the 2 , or 2 , or 2N *, etc. different codes in each case, for "T" different types of components, "D" different drawings, "N>" different components in the drawing, etc. This is a consequence of the variable-length code required in these cases compli
cating considerably the processing in the computer, the algorithms required, etc.
- Determination of the number of spare bits for possible future expansions in the system is difficult, since it may be necessary to change the whole system /lists, programs, etc./ if a relatively large expansion takes place.
- They make it more difficult to implement the whole system due to the existence sometimes of useless bits in the Identifier-word•
- They require relatively much computer time for fetching in a given Table stored in the computer memory, all other data corresponding to the component which sends the Iden
tifier-word. This time will be different for each component depending on its position within the Table.
- In case one component belongs to more than one drawing, it would then be necessary to refer it somehow to two or more different Component Vector in a Table, or send two or more Identifier-words to the computer when the component in
forms it about its state. These solutions will complicate considerably the processing and the format of the Table.
- The verification of the validity of the Identifier-word
168
-received is more difficult, there must be a List in the computer equally ordered than the arriving order of Iden
tifier-words at the computer system.
Prom the previous analysis, we conclude that from these given alternatives the most advantageous is the first one, that is, that which consider the Identifier-word arising from the
component as a fixed-length addressing-word for the computer.
Nevertheless, we do not affirm categorically that this one is the most convenient solution in every case, although it seems to be the most commendable in large systems.
Using this alternative, there follows a form by which all data referring to all drawings can be stored in the computer memory, in order to process the "state” information arising from the components together with it for updating the pic
ture files. At the same time, we will continue the analysis of each computer task enumerated in Sect. 3.6.2.1, in order to give consistency to our criteria.
3.6.2.1.2 Component Vector
About each supervised and already identified component of the process it is necessary to know:
- Drawing to which it belongs
- Type of format of the arriving state-word in order to interpret it
- Coordinates of the cell wherein the symbol representing it appears in the drawing
- Type of Element with which the computer will deal.
These data compose the Component Vector of each component in the process.
In case some component be a part of more than one drawing, these data logically must be specified for each drawing
- 169
-independently. They preferably must be stored one after the other in the computer memory, thus requiring an easier
algorithm for updating. This eventual possibility can be in
dicated with only one bit in the last word of each Component Vector.
The number of computer words required to specify wholly the corresponding Component Vector, is determined by the size of the system and the characteristics of the components which compose it. They are addressed by indirect addressing using the respective Identifier-word received from the components.
This is schematically illustrated in Pig. 3.4
Fig З А
170
-In Pig. 3.4, bit "S" has a value "0" if the subsequent data- word group is related to another component; it has value "l"
if the succeding data-word group is related to the same component /that previously addressed/. Therefore, the last data-word group related to the same component in the sequence /with an arbitrary number of them/ must have the bit "S" with value "0".
All these data are previously stored in the computer memory according to the specific drawings prepared for the system and to the conclusions of the further analyses in this paper.
Addresses of all the first words of all the Component Vectors of the Vector List /having or not several successive data- groups/ related to one component, are stored successively in a Pointer List, either orderly or without order. To each
Component Vector there corresponds one address in the Pointer List on a one to one basis, although a component may have an arbitrary number of data-word groups within the Component Vector assigned to it. Words of this List are addressed by the Identifier-word arising from the components in the process.
Data included in the Component Vector
Each data-group of the Component Vector related to each component of the whole system must include the following necessary data: