Editorial 3Q99
Poor quality RTOS
In our market study, we discovered that many people who are using
an RTOS are not happy with the product they choose. On average, 50% of
the people want to change from the RTOS they use today to another version.
Some RTOSs are doing better, but others are doing worse.
One of the reasons is the poor quality of some products. It seems that
some RTOS vendors do not apply/implement good quality assurance techniques
at all. This has been discovered in our evaluation program. An example
of this can be found in the executive summary we publish in this issue.
The version tested was pSOS
2.6.1. for Intel platforms. We discovered different non-acceptable
issues in a product, which has already been in the market place for quite
some time. This also correlates with a higher degree of “want to change”
attitude for this product revealed by our market study.
The issue is: how many tests do vendors carry out, before they release
the product? We have the impression that only beta testing on the customer’s
site is used for getting most of the “bugs” out. We don’t understand why
you, as a customer, should spend time and money “debugging” a beta version
of an RTOS.
The reasons are obvious. When you test an RTOS in the context of an
application, not all features of the RTOS are used. Also, some erroneous
behaviour is difficult to detect because it is masked by the application.
Our test suites seem to be most efficient. We discovered, for example,
“behaviour” bugs in pSOS and we were
able to point out to the manufacturer what error he had to look for
in the source code. The importance of such an organised testing environment
is crucial. Indeed, the need for RTOS is growing very fast and the number
of systems people use in their daily lives with such RTOS on board is rising
just as quickly.
Not all RTOS are in life critical systems, but who wants blank pages
using a copy machine because the RTOS is not doing the job well enough?
One of the signs that the RTOS topic is very hot is simply the number
of papers we received for this magazine.More than 40! We therefore decided
(just like 2 years ago) to print 2 magazines instead of one. The next one
will be available mid October.
Another sign is that in a very diluted market comprising more than
100 listed commercial RTOS products, some newcomers have been showing up
again. It is really a mystery why people think there is a market for yet
another RTOS.
This and the following issue will provide you with an update on the
current state of the art in the RTOS. We deal with different topics ranging
from OSEK, RT-Linux to JAVA support. I hope you will enjoy this fresh information.
One subject will not be covered in this magazine: CE 3.0. Its availability
is continuously being delayed. It will be probably the beginning of the
year 2000 before we will know if, this time, CE 3.0 can really pose a threat
to its competitors.
