Please feel free to share your thoughts on how the forecast can be improved.
For an up to date and actually usable overview of the upcoming Gnome updates i'd add a link to: http://live.gnome.org/TwoPointNineteen/ReleaseNotes/Draft which shows the current releasenotes for 2.19.
In GNOME 2.18, the Gnopernicus screen reader was replaced by Orca from Sun Microsystems. This next generation screen reader is scriptable to enable accessible applications to be more useable. In addition, multi-user accessibility was added to GNOME 2.20 so that sudo applications are accessible.
In GNOME 2.20, the at-poke object debugger will be replaced by Accerciser. This tool will also enable record and playback of Gtk applications. In addition, the Preferred Applications configuration panel will debut an accessibility tab for the end-user to select a preferred Assistive Technology.
Would be nice to hear about drivers and other stuff that is outside the kernel.org zone and probably will remain that way; zaptel drivers, l4linux and so on.
Disappointed there's no mention of Squashfs. Squashfs is used by a significant number of Linux projects worldwide, and so I'm surprised it gets very little coverage. However, Jonathan Corbet had never covered Squashfs on LWN either, and so I was expecting its absence.
The Evolution mail/calendar/address book bundle is a very important package to many. I know it is pretty much always running on my Linux box. Trying to keep up with the woes and patches for getting Palm PDA syncing to work with it is a daunting task. I'd be happy to see an Evolution section added to the forecast.
I remember just a few years ago that IPv4 was doomed to run out of addresses in a matter of months. Seems that IPv6 has made its way into the kernel, but my ISP never did ask me if I wanted an IPv4 line or an IPv6 line. My main recent awareness of IPv6 was that I learned Firefox runs much quicker on my Ubuntu Dapper box once I disabled IPv6. So, I'd like the forecast to take a moment to look backward at such "recent" features and report what the realistic expectations are for those features at this time.
Perhaps I'm just spotlighting the extent of my ignorance here, but IPsec is another area where I feel I would be helped by a compass that points to true North. I think its included in Linux, but I have no idea how thin the ice might be if I venture into it. Heck, I'm even confused by versions of BIND.
I'm a bit surprised that there's no section of the forecast to talk about programming languages. I sure could use an expert to fill me in on what to expect of C/C++, Java and Python and TCL/TK and Perl, oh my.
I never have fully understood why the kernel cares as much as it does about exactly which version of C it is built with, but it does. I gather there are major tectonic shifts happening with Java and "Open Source", but I haven't had time to keep up with those. And I hardly know what Perl 6 is going to be, to say nothing of "when". Is Python 3000 something I should keep on my radar?
ZFS can be used/tested on Linux already. There are two ZFS ports on the way. One of them uses the FUSE framework and the other one is the integration into Lustre on Linux, helped by SUN. Parted Magic and SystemRescueCD includes the former betas. It's interesting that the code which apparently should be in the kernel for high performance was released under GPL2 (e.g. the ZFS checksumming, compression support what GRUB uses). Some interesting links:
Since there are a few requests for tracking userspace stuff: perhaps delete KDE/Gnome from the current weather forecast, rename it to kernel weather forecast and restart a true platform forecast, covering userspace stuff such as KDE, Gnome, Perl6, Evo, GCC, ...
(Time permitting, obviously.)
(OTOH each of those deserve a single forecast bulletins if info should be worth as much as the kernel info is. And then you can start covering Distros, too. And ... And ... And ... go insane, in the end, no doubt.)
There seem to be some great efforts going into solving the memory deadlock issue preventing the ability to reliably swap over the network. This is increasingly important for a lot of organizations who use Linux and want to move to a more efficient deployment model such as diskless blades. It would be great to see these efforts given some visibility in the forecast!