
|
[openbeos]
||
[Date Prev]
[08-2001 Date Index]
[Date Next]
||
[Thread Prev]
[08-2001 Thread Index]
[Thread Next]
[openbeos] Re: Thoughts on rewriting BeOS, if it is necessary (LONG)
- From: "Michael Phipps" <mphipps1@xxxxxxxxxxxxxxxx>
- To: openbeos@xxxxxxxxxxxxx
- Date: Sun, 19 Aug 2001 06:59:12 -0400
>On Sun, 19 Aug 2001 05:54:13 PM, DarkWyrm <bpmagic@xxxxxxxxxxxxxxx> wrote:
>
>First, I would like to mention that my native language is German, so if you
>notice
>all the grammar and spelling errors in my posts, this doesn't mean I'm stupid
>:)
>
>>Assuming that Palm will not open source anything BeOS and will not develop it
>>either (a definite possibility), I have been looking into the possibility of
>It is very much possible. And even if Palm now says that they will continue
>BeOS
>support, it is possible that they will do very little development, neary no
>advertising,
>a bad distribution, whatever. You never now what those people in companies
>decide.
>They are also free to drop BeOS support anytime.
Sigh. This is probably the best arguement that I know of for "open source
everything".
I doubt that it can be worse than Be Post Focus Shift.
>>using the 2.4.0 Linux kernel along with a bare-bones version of XFree86 as a
>>base and building the various servers (app server, net server, etc) and going
>>from there. Guillaume Maillard started me on this and I think it's at least
>This depends on what we really want to archive.
>Of couse rewriting a kernel from scratch is hard (and I'm not the person who
>could do this),
>but what do we really want. Do we want to stay compatible with BeOS?
>Or do we wnat to merly have a normal Linux kernel and then add MediaServer,
>Interface Kit and so on.
I personally want to stay 100% BeOS compatible for a first release. The future
can include
improvements (as Be did). I have no interest in improving Linux.
>The first problem is the BFS (BeOS file system) We really need to stay
>compatible
>with this, as ALL BeOS user use it.
>If you rewrite a BeOS compatible kernel, you could probably simply use the bfs
>driver
>from BeOS R5 PE and load it (the interface it uses to comunicate with the
>kernel IS defined,
>and the rest can be seen in dosfs sample code),
We have a significant piece of this, already. I think that BFS is too critical
a piece to not keep.
[snip]
>You need to decide if you want to do
>
>* create just another open source os
>* add Media Kit, Interface Kits and other kits to Linux
>* or if you want to recreate the BeOS, keeping binary compatibility.
#3.
>If you have a kernel consisting of a open source license, you can start
>porting all those GPLed drivers the exist with Linux.
>This is not possible with the current BeOS because of legal issues.
License was never really the issue with those. I would gladly face anyone
in court who would make the (foolish) arguement that the linking with a
kernel that occurs with a driver would require the whole kernel to be open
sourced. Drivers are something end users can produce and load. That wouldn't
force Be to open source.
|

|