|
|
![]() |
- Compatibility with a Windows and UNIX dominated world is not a
primary concern. It should be born in mind that desktop machines
make up a relatively small minority of all computers: embedded
systems are far more common. Most embedded systems do not suffer
from such backwards-compatibility concerns. However, there is no
reason why binary support for (say) Linux applications cannot be
added (possibly without even modifying Go!)
- Go! is definitely not a solution to all problems. As mentioned
in Overview,
Go! is component-based, and so is not necessarily suited to
applications which aren't object-oriented.
- Go! will not be portable across different processors. L4 showed
that the OS itself is a poor place to support portability due to
performance considerations. Recent developments such as CORBA allow
applications to run without modifications to their source code on
different OSs, without constraining the OS by demanding it be
portable
- To limit the scope of the project, multiprocessor issues will
be of secondary concern, and there are no plans to produce a
multiprocessor version of Go!
- Distribution is of secondary concern. This is mainly to limit
the scope of the project, but also because it is envisaged
middle-ware such as CORBA could better play this role.
- For all the mention of CORBA, Go! will not be CORBA compliant.
This too is to limit the scope of the project. A CORBA
implementation of the Go! ORB would be an interesting project
however.
| |