Hello Jim and all,
I like the idea of having two releases of FreeDOS with different goals: a
FreeDOS 1.2 as an update of current FreeDOS 1.1, in order to have something
on a short term as an update of current distribution.
As for FreeDOS 2.0, I share my ideas here. I agree that it should be a big
leap forward on the development of FreeDOS, and as my biggest worries are
the disappearing of BIOS and 16-bit mode, they are about moving to 32-bit.
Three possibilities:
(1) My favourite would be to develop something similar to Microsoft's
VMM32.VXD: a 32-bit kernel that would allow you to run different
16-bit/32-bit VMs, with expansion by 32-bit "drivers" (the VXD) that
emulate BIOS and control the VM interoperativity.
With some help (multithreading), and some software to run EXE-PE's and
DLL's, FreeDOS could expand not only to running DOS-16bit apps, but also
legacy Win32 apps (e.g. console ones). With quite an effort, this scheme
may be extended to run Win16, POSIX (minGW?) or OS/2 apps.
Booting could start with 16-bit, and then "execute" the extender, or
alternatively, set a bootstrap sequence (set by SYS) to directly run the
extender.
The disadvantage is that this is quite a lot of work, I think Japeth has
moved on this way, so there's something where to start. The advantage is
that DOS would retain much of its flavour, and continue the tradition
to run legacy Microsoft software.
(2) A second option would be that, instead of a "new" 32-bit kernel, we use
Linux: you boot a Linux without anything extra, and that simply runs DOSEMU
(or different consoles with different DOSEMU's, in order to emulate to have
different VM's). My plan was to pick a Linux distribution and start
stripping it down to simply boot DOSEMU.
The advantage is that FreeDOS would live as long as Linux lives (and no
matter if it moves to 64-bit), that is, forever. The disadvantage is that
it would mean moving to a native POSIX world, it may loose some of the DOS
flavour, and everything that does not work on DOSEMU, would not run on
"FreeDOS 2.0".
(3) Go the FreeDOS-32 way that you mentioned. The advantage is the total
freedom and unconstrained start for a 32-bit kernel, based on what's there.
The disadvantage is that it may also require a lot of work, and that it may
be to start a work compatible with nothing that already exists.
I just thought I should mention them before entering a critical way that
requires a lot of work.
Cheers and happy 2015 to everyone.
Aitor
2014-12-31 23:51 GMT+01:00 Jim Hall <jh...@freedos.org>:
> Mike pointed out that the FreeDOS Road Map
> <http://www.freedos.org/wiki/index.php/FreeDOS_Road_Map> (wiki) is out of
> date and short on details and suggested a broad discussion on the road map,
> get consensus and have it updated.
>
> I figured we should start a separate discussion thread about that.
>
>
> First, a little history on the "road map":
>
>
> When I started FreeDOS in 1994, the goal was to create a free version of
> DOS, compatible with MS-DOS. (original PD-DOS announcement
> <https://groups.google.com/d/msg/comp.os.msdos.apps/oQmT4ETcSzU/O1HR8PE2u-EJ>)
> (original Free-DOS manifesto
> <https://groups.google.com/d/msg/comp.os.msdos.apps/W6MuhF__R9s/MgdzBrlanTwJ>
> )
>
> That aim remained essentially the same for a long time. And I believe we
> met that with version 1.0 a few years ago. FreeDOS 1.1 was an update to
> FreeDOS 1.0, so things didn't really change there.
>
> In 2009 <http://www.freedos.org/jhall/blog/?id=20090511-192954>, I
> briefly stepped away from FreeDOS to focus on a graduate program (I ended
> up changing jobs instead, and didn't do the M.S. program until a few years
> later). During that time, Pat Villani stepped in as project coordinator.
>
> Pat wanted to do something to spur development, so in 2010
> <http://www.freedos.org/wiki/index.php?title=FreeDOS_Road_Map&oldid=845>
> he wrote the first version of the FreeDOS Road Map (see old version
> <http://www.freedos.org/wiki/index.php?title=FreeDOS_Road_Map&oldid=845>).
>
> In 2011 <http://www.freedos.org/jhall/blog/?id=20110502-164858>, Pat's
> health was getting worse, so I came back as project coordinator. Nothing
> had been done on the road map, so I replaced it with a copy/paste from a
> blog post I had written earlier (see blog post
> <http://www.freedos.org/jhall/blog/?id=20110502-164858>) until I could go
> back and update it "for real" later on.
>
> A few days ago, Harold (aka Mercury Thirteen) mentioned the road map on
> the wiki. I realized I never updated the road map, so I tacked a "THIS PAGE
> IS OUT OF DATE" notice at the top of the wiki entry.
>
> More recently, Chelson announced on our Facebook group
> <https://www.facebook.com/groups/freedos/> his kickstarter project to
> fund development of FreeDOS-32, with the hope that this kernel could be
> part of "FreeDOS 2.0". I shared his announcement on the website and on
> freedos-devel this morning.
>
>
>
> I think that brings us to this discussion. :-)
>
>
> *My thoughts on FreeDOS 1.2 and FreeDOS 2.0:*
>
> I think the next distribution will be FreeDOS 1.2 because I want to
> reserve "2.0" for a major shift in how FreeDOS software is organized and
> what we include. I'm currently looking at the packages we list on the
> Software List, so maybe we'll get to "2.0" soon - but I think it's safe to
> assume the next FreeDOS distribution will be 1.2 and the one after that
> would be 2.0.
>
>
> *My thoughts on FreeDOS 1.2:*
>
> FreeDOS 1.2 is basically an update from FreeDOS 1.1. The biggest change is
> probably the installer: It should be very simple, very straightforward.
> FreeDOS isn't very big or complex, so it doesn't make sense to have a lot
> of install options.
>
> The current install process (FreeDOS 1.1) has a lot of steps to it. I
> think we could simplify this a bit. I'm not sure about the full steps for
> the install process, but to brainstorm something, here's a sketched out
> plan:
>
> 1. *Boot the FreeDOS Install CDROM.* This is basically a "live" FreeDOS,
> which happens to boot into an automated install process.
> - If you Exit the process at any time, you go back to a DOS prompt. (This
> is useful for people who just want to run FreeDOS from CD without
> installing it.)
>
> 2. *Does the C: drive exist?*
> - If not, prompt the user to run FDISK. Reboot to re-read the partition
> table.
>
> 3. *Is the C: drive usable?*
> - If not, prompt the user to run FORMAT.
>
> 4. *Start the INSTALL program.*
>
> 5. *Run SYS to make the C: drive bootable.*
>
> 6. *Do any follow-up steps* (such as creating a default CONFIG.SYS and
> AUTOEXEC.BAT, set language, etc).
>
> 7. *Done*
>
>
>
>
> *My thoughts on FreeDOS 2.0:*
>
> I'd like to see a modernized version of DOS. This might be as ambitious
> as Marc Perkel mentioned in his 1991 letter to his Novell bosses about a
> modern DOS, which he called "NovOS" (read NovOS letter). Or it could be
> something less ambitious, basically a modernization of the FreeDOS
> userspace (utilities, etc) while keeping the original DOS kernel.
>
> I haven't used FreeDOS-32 but if it supports classic DOS programs on
> modern systems, while adding new and useful features, I would support using
> that kernel in FreeDOS 2.0. (I said this on Facebook, too
> <https://www.facebook.com/groups/freedos/permalink/10152599941377887/?comment_id=10152600011617887&offset=0&total_comments=27>.)
> Above all, application compatibility is 100% important. FreeDOS 2.0 needs
> to run applications written for MS-DOS.
>
>
> Thoughts?
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming! The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net
> _______________________________________________
> Freedos-devel mailing list
> Freedos-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/freedos-devel
>
>
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel
