Learn, hack!

URL

http://events.ccc.de/congress/2006/Fahrplan/attachments/1230-23c3_sw_reliable.odp

File name

1230-23c3_sw_reliable.odp

File size

1.7 MB

MD5

eb78ea4dc72d777220cce772a6465cb2

SHA1

e9b365f228a67747d22e75c97ff5d3b4a66b8dc3

The challenge of designing reliable is managed different in every industry. This lecture will give an overview how safety critical and reliable software is designed and produced in the area of aerospace industry and why this could also be interesting for other applications (like web-design) "The greatest of faults, I should say, is to be conscious of none. " - Thomas Carlyle It is the humans most valuable gift to be unprescise – as many new things would not be discovered otherwise. But for reliable (software) development this evolutionary gift turns to be a big challenge as the final product shall not contain any faults. This lecture is about preventing development faults on the example of the aerospace and automotive industry. Part one of this lecture is a basic introduction why reliability is an issue at all. It is obvious that a failure in the primary flight control will seriously endanger the live of people. But buried under functionality of a e.g. Web-application the criticality of these application cannot easily recognized. Often these application are not directly live threatening, but the loss of the income source also endangers lives (at least the quality). Part two will introduce the DO-178B standard with which the aerospace industry tries to handle the reliability challenge. The basic idea of this quality standard is very simple: “SW shall contain only functionality it has to”. To achieve this postulate the DO-178B specifies a series of processes and documents (which can be seen as artefacts of these processes). Of course some examples will illustrate this part. The last part will summarize the lecture and gives some ideas on the (re-)usage of the described methods for “classic” (non-safety critical) application.