Vorige                       Titel                            Volgende
______________________________________________________________________


The following things can be done with an UFI:

- The write task can write appending to the end of the UFI. (The data
  that the write task writes is put at the end of the datablock
  already in the UFI.
- The read task can read destructively from the beginning of the UFI.
- The read task can set a pointer, the 'current' pointer anywhere
  in the UFI and read from it. The read task is the only one wich can
  affect this pointer.
- Any task can read any byte from the UFI provided it can find the UFI.
  This way of access is relatively slow when the UFI is large.

The size of an UFI is only limited by the size of the memory of the
computer.
To make a FIFO one task is made the write task of an UFI and another
task is made the readtask of that UFI. If the writetask writes into
the UFI and the readtask reads from it destructively the FIFO effect
has been achieved. This makes it possible to use FIFO's with no
sizelimit.
A task can get access to an UFI in three ways.
It can create the UFI itself.
It can get the UFI from it's parent on startup of this task.
It can find the UFI as a file in the directory.

An UFI also has an owner. The owner is the task wich determines
who is the readtask and who is the writetask. Also the owner can
put the UFI in the directory and make it a file. In this way other
tasks kan get access to the UFI even after the owner has been killed.

Directory
What we have been calling the directory until now is in fact a number
of directories. A directory is a set of directoryentries.
Any directoryentry can contain one of the following:
- An UFI. The UFI and the directoryentry together form a file.
- A program. This contains the code of a certain program.
- A directory.

There is one main directory called root. All other directories are
subdirectories of another directory (possibly root).

One UFI, program or directory can be represented by one or more
directoryentries in one or more directoryentries. These directory-
entries may have different levels in respect to the root. However, no
directory may contain itself in any way.
Every directoryentry has a name wich is unique in the directory con-
taining this entry. The name is used to find a particular entry in
a directory.

______________________________________________________________________

Vorige                       Titel                            Volgende