A GENERAL VIEW - Hard Real-Time Computing Systems: Predictable Scheduling Algorithms and Applications - page 2

military systems,
space missions,
consumer electronic devices,
multimedia systems,
smart toys, and
virtual reality.
In many cases, the real-time computer running the application is embedded into the
system to be controlled. Embedded systems span from small portable devices (e.g.,
cellular phones, cameras, navigators, ECG Holter devices, smart toys) to larger sys-
tems (e.g., industrial robots, cars, aircrafts).
Despite this large application domain, many researchers, developers, and technical
managers have serious misconceptions about real-time computing [Sta88], and most
of today's real-time control systems are still designed using ad hoc techniques and
heuristic approaches. Very often, control applications with stringent time constraints
are implemented by writing large portions of code in assembly language, program-
ming timers, writing low-level drivers for device handling, and manipulating task and
interrupt priorities. Although the code produced by these techniques can be optimized
to run very efficiently, this approach has the following disadvantages:
Tedious programming . The implementation of large and complex applications
in assembly language is much more difficult and time consuming than high-level
programming. Moreover, the efficiency of the code strongly depends on the pro-
grammer's ability.
Difficult code understanding . Except for the programmers who develop the
application, very few people can fully understand the functionality of the software
produced. Clever hand-coding introduces additional complexity and makes a
program more difficult to comprehend.
Difficult software maintainability . As the complexity of the application soft-
ware increases, the modification of large assembly programs becomes difficult
even for the original programmer.
Difficult verification of time constraints . Without the support of specific tools
and methodologies for code and schedulability analysis, the verification of timing
constraints becomes practically impossible.