[ros-kernel] ReactOS Roadmap
vizzini at plasmic.com
Wed Oct 15 00:35:59 CEST 2003
On Mon, 2003-10-13 at 14:26, Jason Filby wrote:
> Please suggest:
> major milestones (6, 12 and 18 months away)
> minor milestones (minimum granularity of a month)
For the next 6 months, I'd like to see both code and non-code
improvements. Some non-code improvements first:
1) Documentation of all of the tests, including how well they should
work on latest builds. This is the very beginning of a regression test
suite for us.
2) Adoption of the bugzilla isuse-tracking system that is installed on
reactos.com. This will greatly help coordinate releases and keep
quality of the OS on the rise.
3) Release of a ReactOS Development Kit, as described on ros-general
recently, in order to ease the process of bringing new developers into
4) Support of some method of kernel debugging. We have partial support
of GDB in a cuople of contexts and no support of microsoft debuggers.
I'd like to see either or both of those change in order to facilitate
5) Clean-up of existing processes, including the testing and release
processes, so that releasing doesn't saturate the available cycles of
key engineers any more than necessary, and equally so that important
things don't get missed.
6) DOCUMENTATION - #1 requirement here is aimed at developers, with
gradual progression backwards towards admins and eventually users. In
particular, I'd like to make it as easy as possible for new developers
to join the project, and documentation of processes, architecture,
standards, etc., are all important.
Given those things, we should be able to both improve existing
developers' productivity and maximize the possibility of bringing in new
developers to the project.
Coding milestones over the next 6 months:
1) I'd like to see NDIS be able to load a large number of ethernet
drivers, both virtual and physical.
2) I'd like to finish TCP to the point that I can run network servers on
ReactOS. I think this will be the bulk of my coding time.
3) I'd like to see us get the header situation that I alluded to in
another post figured out. Currently there are 4 different sets of
headers that we support to varying levels (MS PSDK, w32api, wine, and
the ms psdk).
4) There have been various observations and requests related to our cvs
structure and internal header structure; I'd like to have a discussion
in the next few weeks/months and make some decisions.
5) Along the lines of general stability and robustness of the codebase,
I would like to see the adoption of (more) assumption checking with
ASSERT() and PAGED_CODE() in the kernel, and possibly see some
discussion about general coding standards (aside from formatting...)
6) I'd like to see SEH added to the places that need it in the kernel
7) Apps: I'd like to see the basic accessories from Windows running -
notepad, calc, etc.
8) Begin the conversion to a fully PnP-driven kernel, including PnP
support for NDIS, scsiport, and videoport
My 12-month goals for ReactOS are primarily application-oriented. I'd
like to see things such as Mozilla and OpenOffice start to come to
life. We're also going to start needing serious work on third-party
device support in this timeframe, as well as meaningful support of
Windows Networking (i.e. cifs/smb shares, etc). We'll need to round out
support for the security components of the OS. USB and other system
buses will need to be developed within the framework of the PnP work.
My 18-month goal for ReactOS is simple: I want to see someone, anyone,
ship it. It can be as modest as the LindowsOS stuff Wal-Mart did in the
US, but I'd like to see actual end-users start to see benefit from
ReactOS. This is, after all, the only real goal of the project, and 18
months seems (at this point, anyway), a reasonable time frame for this.
I'm going to update my NDIS roadmap tonight; it can be seen at
http://plasmic.com/~vizzini/rosnet.html for the curious.
More information about the Ros-kernel