OSD development moves toward Qt
Michael Gao
This is to announce the upcoming major OSD development milestone, Qt on
OSD. Qt application framework and Qt based OSD applications will start
to show up on Neuros repository trunk at svn.neurostchnology.com on
12/01, 2007.
Compared to the existing Neux (Neuros application framework based on
Nano-X) system, the upcoming Qt support on OSD will move OSD development
to an exciting new level, with the much more comprehensive, flexible and
robust Qt platform. Qt on OSD will not only boost development
performance on OSD itself, it will also greatly improve the
cross-platform application support on OSD. With OSD Qt development, more
extensive OSD functionality and broader 3rd party application support on
OSD are on the way.
OSD Qt development will be based on Qtopia-core, system is structured
with a window manager server application, while major functionality
pieces coming as standalone applications. This gives system the most
flexibility and expandability. While existing OSD software already has
the capability to bring up 3rd party applications from any storage media
that OSD has access to, with Qt's strong cross-platform support, Qt on
OSD will make 3rd party application support much easier and enable OSD
to run lots of exiting Qt written applications.
Below are several important notes in regard to this milestone change,
1. Release plan
12/07, 2007: Daily developer release starts,
with limited multi-media functionality support
01/18, 2008: Alpha release with full multi-media support
02/08, 2008: Production software release
2. Upgrade path and prerequisite
Network based auto upgrade to this new platform based software is
transparent, nothing special is needed, while manual upgrade takes the
same download and pick to upgrade process with the requirement to start
with certain version (3.33-1.xx, version number TBD).
To run latest Qt based software on your existing OSD hardware, you'll
need to have a CF-card present with OSD as the program storage. While
next-gen OSD will eliminate this requirement, Neuros will provide the
CF-card as compliments to existing OSD users. We will post more details
on how you can obtain your CF-card in the near future at
http://www.neurostechnology.com/support.
3. Existing software development
Major enhancement and fixes will continue to apply to existing software
before new software goes to production, among which are 4GB 2.5hour
recording limit fix and video playback aspect ratio fix etc. Once the QT
version of releases are in production, around February 2008, all future
development effort will be focused on that.
With the above said, thanks a lot for all your interest and contiguous
support to Neuros and Neuros OSD, together we are looking forward to
making OSD an even more exciting open source device.
Regards,
/MG
greyback
this is welcome news, and I'm looking forward to this development!
One statement I wish to clarify is where you state "all future
development effort will be focused on [Qt]".
Am I correct in interpreting this as the codecs will be frozen in Feb
2008? If that is the case, there should be a real push right now to
gather videos which still have playback faults and get them fixed (if
possible). I've sent Johan clips of all my files that play back
incorrectly, (a small percentage of the all my videos) but there's
still work to be done in that regard.
I'd also like to ask on behalf of those people who use the CF card
slot for their media, as to if the OSD will cope if they remove the
permanent storage card and insert their own to play/record with?
Thanking you
-G
Fernando Cassia
http://www.gnome.org/mobile/
The question is ... did you consider GTK before making the choice or was QT selected just because?
http://www.linuxjournal.com/article/4870
I feel that the industry is going towards Gnome... from Nokia's Linux tablets to the OpenMoko platform... and I'd be saddened to see Trolltech's (and by extension, KDE's) life extended further by this move...
:-/
Are the hardware requirements of Qt so much lower than GTK?
Just thinking aloud...
FC
Daniel Díaz
I am also from the GTK+/Maemo camp, but am however very pleased to see
Neuros adopting an existing toolkit for future developments. This will
definitely aid in bringing in contributors (and already existing
contributions) for new applications to the system.
I subscribe to Fernando's words asking if other mobile environments
were considered and how the decision was made in favor of Qtopia.
Michael, you mentioned a permanent CF card for program storage. Can it
be an NFS server instead?, or at least an option. All of my media
files (including recordings) are accessed through NFS from a RAID
system, and would very much prefer that over a perishable CF card.
Greetings!
Daniel Díaz
yo...@danieldiaz.org
Ugo Riboni
> development effort will be focused on [Qt]".
>
> Am I correct in interpreting this as the codecs will be frozen in Feb
> 2008? If that is the case, there should be a real push right now to
> gather videos which still have playback faults and get them fixed (if
> possible). I've sent Johan clips of all my files that play back
> incorrectly, (a small percentage of the all my videos) but there's
> still work to be done in that regard.
No, it doesn't mean the work on the media system and the codecs will
be frozen or stopped.
All future applications developement will be on QT as opposed to the
nano-x/neux toolkit, which will be basically discontinued.
Everything else (codecs, nms, xmms2) will keep being supported as is
now, as it's part of the base system as is the OS itself.
> I'd also like to ask on behalf of those people who use the CF card
> slot for their media, as to if the OSD will cope if they remove the
> permanent storage card and insert their own to play/record with?
With the new system, CF card cannot be used anymore for media.
Once you load one of the new firmwares, the card should stay
permanently glued to the OSD slot.
Removing it will prevent the system from booting, and likely prevent
it from running too.
The free cards we send out to people requesting them will probably
have a red sticker with skull and bones and menacing warnings on them,
to ensure people don't ever take these cards off of their slots. I
also considered suggesting sending out a tube of superglue with the
cards, but it looked like a bit too much ;)
In the future we may or may not relax this limitation, or allow you to
use the SD slot instead, but surely at the beginning using CF for
media will not be possible.
Unfortunately that is a tradeoff we needed to make for the current OSD units.
There should be an hardware revision coming up soon which will add an
internal NAND storage where the code will go, thus freeing again the
CF for media. But the current units cannot be retrofitted in any way
except by using the CF card slot, so instead of leaving current users
out in the cold, we opted for this solution.
It looked like the best compromise.
Cheers,
--
nero
Ugo Riboni
> were considered and how the decision was made in favor of Qtopia.
More on this for both of you later in the following days.
> Michael, you mentioned a permanent CF card for program storage. Can it
> be an NFS server instead?, or at least an option. All of my media
> files (including recordings) are accessed through NFS from a RAID
> system, and would very much prefer that over a perishable CF card.
You can always netboot, like you can in the current system.
We do all our developement with the OSD fetching the kernel via TFTP
and mounting the rootfs via NFS.
You can do this too, with a little bit of setup.
But if you were asking to pick up via NFS only the code that would
otherwise reside in the CF card, this will not be available in the
early releases.
We can consider adding this option later on, it doesn't seem
technically hard to do.
And thinking of it, you can probably already hack something like this
using a custom init script and unionfs (not 100% sure, just kind of
thinking aloud).
Cheers,
--
nero
Michael Gao
>
> > I subscribe to Fernando's words asking if other mobile environments
> > were considered and how the decision was made in favor of Qtopia.
>
> More on this for both of you later in the following days.
development too. The reason Qt was chosen is because we internally have
Qt resources and the knowhow to get it started, in addition to the fact
that crweb (now full-time with Neuros :)) had already ported the Qt base
system over to OSD about 6-8 months ago.
Regards,
/MG
Fernando Cassia
Sounds interesting. Did you buy a batch of ancient 32MB CF cards or something?
I really love using a 8GB Seagate microdrive in the OSD's CF slot. :(
If given the option, I'd really would love being able to select which slot acts as the "non-removable" storage...
Another solution would be inserting a CF/SD USB multi-reader into the SD slot... but I'm not sure it'd be supported given the USB port driver problems....
FC
Dan Fandrich
> With the new system, CF card cannot be used anymore for media.
> Once you load one of the new firmwares, the card should stay
> permanently glued to the OSD slot.
> Removing it will prevent the system from booting, and likely prevent
> it from running too.
But I would expect much less than 64 MB of the CF would actually be needed
for the OSD code, so it shouldn't take much work to partition the card
and use any remaining space in a separate partition to be used for
media storage.
>>> Dan
--
http://www.MoveAnnouncer.com The web change of address service
Let webmasters know that your web site has moved
Michael Gao
> I really love using a 8GB Seagate microdrive in the OSD's CF slot. :(
We do not recommend, but you still can record to and play from your CF
card even when part of it is used as program storage. OSD actually will
reformat CF-card to be ext-3 file system, thus it may no longer be
compatible with your portable device or Windose PC.
> If given the option, I'd really would love being able to select which
> slot acts as the "non-removable" storage...
Yes, this is a possible enhancement.
Thanks,
/MG
Vladimir Pantelic
> I can't say I'm very pleased, I was hoping for GTK.
>
> http://www.gnome.org/mobile/
>
> The question is ... did you consider GTK before making the choice or was
> QT selected just because?
>
> http://www.linuxjournal.com/article/4870
Did you add this link as a "pro" or "contra" GTK argument?
> I feel that the industry is going towards Gnome... from Nokia's Linux
> tablets to the OpenMoko platform... and I'd be saddened to see
> Trolltech's (and by extension, KDE's) life extended further by this move...
KDE barely needs "life support"...
Dragon Wisard
> I feel that the industry is going towards Gnome... from Nokia's LinuxKDE barely needs "life support"...
> tablets to the OpenMoko platform... and I'd be saddened to see
> Trolltech's (and by extension, KDE's) life extended further by this move...
I can't speak for the embedded scene, but it seems like some big names (at the least, Linus Torvalds and Patrick Volkerding) have picked KDE over Gnome as the desktop environment of choice for Desktop users. I think Gnome has recently gained temporary notoriety and attention with the Ubuntu explosion, but prior to the South African distro it was being dropped left and right by respectable distributions. Besides, even Torvalds will testify that in our industry (and life in general) choices are infrequently made on technical merit alone.
--
http://www.dragonwisard.net/
Pat Gunn
Volkerding is almost a non-player at this point (as Slackware is now a minor
distro), and Linus doesn't make distros. Not that I really want to jump
into this argument (as I don't use much of either environment), but I would
like to invite you to substantiate your claims, as your language suggests
that you both have strong opinions on the matter and may know little of
what you're talking about. What respectable distributions have been
dropping GNOME "left and right"? The following distros are considered
the most widely used:
*Ubuntu - GNOME, KDE variant available
*Debian - GNOME-centric, KDE available
*Fedora/RHEL - Defaults to GNOME, with KDE available
*SuSE/OpenSuSE - Defaults to KDE, with GNOME available (prior versions were KDE-only)
*Gentoo - Whatever the user installs
If the "respectable distros" you mention are just minor players, I suggest
that it doesn't matter what they do - none of the big distros seem to
be dropping GNOME (nor KDE, for that matter). Saying that it's "being dropped
left and right" and characterising the general interest in it as
"temporary notoriety" seems dishonest and manipulative to me. If you can
back up your claims, please do so.
---
Pat Gunn
mod: csna, bmcm, bmco, cooa, cona, clpd, coom
http://dachte.org
"Once the game is over, the King and the pawn go back in the same box"
--Italian Proverb
Vladimir Pantelic
>> On Nov 26, 2007 5:52 AM, Vladimir Pantelic <p...@nt.tu-darmstadt.de> wrote:
>>
>> > > I feel that the industry is going towards Gnome... from Nokia's Linux
>> > > tablets to the OpenMoko platform... and I'd be saddened to see
>> > > Trolltech's (and by extension, KDE's) life extended further by this
>> > move...
>> >
>> > KDE barely needs "life support"...
>> >
>>
>> I can't speak for the embedded scene, but it seems like some big names (at
>> the least, Linus Torvalds and Patrick Volkerding) have picked KDE over Gnome
>> as the desktop environment of choice for Desktop users. I think Gnome has
>> recently gained temporary notoriety and attention with the Ubuntu explosion,
>> but prior to the South African distro it was being dropped left and right by
>> respectable distributions. Besides, even Torvalds will testify that in our
>> industry (and life in general) choices are infrequently made on technical
>> merit alone.
>
> Volkerding is almost a non-player at this point (as Slackware is now a minor
> distro), and Linus doesn't make distros. Not that I really want to jump
> into this argument (as I don't use much of either environment), but I would
> like to invite you to substantiate your claims, as your language suggests
> that you both have strong opinions on the matter and may know little of
> what you're talking about.
All I wanted to say is that neither KDE nor GNOME are on "life support" or being
dropped - they are *both* active and healthy and therefore not the right thing
to judge whether to go QT or GTK
Regards,
Vladimir
Anthony Taylor
>
> All I wanted to say is that neither KDE nor GNOME are on "life support" or being
> dropped - they are *both* active and healthy and therefore not the right thing
> to judge whether to go QT or GTK
>
> Regards,
>
> Vladimir
IMNSHO, there's really only one technical discussion concerning the
choice of GTK+ or Qt: C++.
I personally dislike Qt simply because it is tied exclusively to C++. I
dislike C++ so intensely, I will not program using Qt. This is a
personal preference, obviously, but my opinion is based on technical
issues, and is not *entirely* personal.
However. I would never dream of driving the debate strictly on my
personal preferences, especially since I've not written anything of
interest for the OSD. (I do love it, though, and hope to have time to
write something interesting. Once I figure out what I want that it
doesn't already do.)
To me, since Qt has an open license, there's no other real debate. It's
all on whether or not you like C++.
(For the record: yes, I know there are bindings for Python, Java, Mono,
and so on. I like C. Pure and simple C. Objective-C if I'm in an OO
mood. Sometimes LISP. Smalltalk once in a while. That sort of thing. So
moving to Qt would probably discourage me from hacking the OSD. Not that
it matters, as I said-- I've not done anything interesting yet.)
Anyway, that's my opinion. And that's all it is: an opinion. I just
thought the debate on KDE vs. Gnome completely missed the technical point.
- Tony
Pat Gunn
> and so on. I like C. Pure and simple C. Objective-C if I'm in an OO
> mood. Sometimes LISP. Smalltalk once in a while. That sort of thing. So
> moving to Qt would probably discourage me from hacking the OSD. Not that
> it matters, as I said-- I've not done anything interesting yet.)
It looks like there are not Perl/Qt bindings for Qt4, so it looks like
this shift will mean I won't be doing dev stuff either if we lose the
ability to rely on X....
Ugo Riboni
> > and so on. I like C. Pure and simple C. Objective-C if I'm in an OO
> > mood. Sometimes LISP. Smalltalk once in a while. That sort of thing. So
> > moving to Qt would probably discourage me from hacking the OSD. Not that
> > it matters, as I said-- I've not done anything interesting yet.)
>
> It looks like there are not Perl/Qt bindings for Qt4, so it looks like
> this shift will mean I won't be doing dev stuff either if we lose the
> ability to rely on X....
Do you mean rely on nano-X ?
Because we never had a real X server on the OSD to begin with, nano-X
despite the name isn't really X.
Also, we don't have perl on the OSD either, last i checked ;)
Don't take my comment the wrong way, I am just trying to understand if
i'm missing something here.
Cheers,
--
nero
Pat Gunn
I've been playing with getting Perl working... the NXLib project
looked "close enough" to proving an X-like API that NanoX not really
being X11 shouldn't be a big concern.
Ugo Riboni
> > > this shift will mean I won't be doing dev stuff either if we lose the
> > > ability to rely on X....
> >
> > Do you mean rely on nano-X ?
> > Because we never had a real X server on the OSD to begin with, nano-X
> > despite the name isn't really X.
> > Also, we don't have perl on the OSD either, last i checked ;)
> >
> > Don't take my comment the wrong way, I am just trying to understand if
> > i'm missing something here.
> > Cheers,
>
> I've been playing with getting Perl working... the NXLib project
> looked "close enough" to proving an X-like API that NanoX not really
> being X11 shouldn't be a big concern.
Oh, that's pretty nice. How is performance, though ?
I've heard that NxLib wasn't really big on that side, but it's just
hearsay at this point.
Intersting info, though, didn't hear you were working on this before,
--
nero
Ugo Riboni
> all on whether or not you like C++.
Speaking to the people who don't like C++. I understand you. Everyone
has his preference in languages.
(hell, after all i would personally write my software in Ruby, if it
was realistically possible ;))
However in the OSD C++ will be used majorly for the GUI code an some
applications logic.
The NMS media server is still C, xmms2 is still C, we'll add C API to
the new scheduler.
If there's some system component that you will need to access from C
and you can't, we can add C APIs to it you request for a reasonable
purpose (i.e. not just "because it would be nice").
And most importantly, if you have a really cool idea for the OSD, you
still can make it into reality using C, assuming it's not a mostly-GUI
idea.
Make a great back-end in C that allows the OSD to do
cool-functionality-foo, with a great API, and if the function is
interesting a GUI can be put on top it by Neuros.
Or you can team up with some other hacker who is more C++ minded than
you. Collaboration is one of the cornerstones of open source, after
all.
And please notice that "you" in the sentences above is really meant as
"you all", I didn't mean it personally.
You see, another time when preference in language does matter ;)
> Anyway, that's my opinion. And that's all it is: an opinion. I just
> thought the debate on KDE vs. Gnome completely missed the technical point.
I think everyone is also missing another important thing by sliding
the debate on kde v/s gnome (aside of the fact that it's really qt v/s
gtk, we're not talking desktops here, we're talking toolkits. The
discussion about the distros in the other thread is really pointless,
in my opinion).
The thing is that we had to make a choice, and we knew that whatever
we ended up with was going to have tradeoffs, and people not liking
them for one reason or the other.
Everyone knew that choosing any of the estabilished toolkits was at
risk of holy wars, and flames, and all that "nice" stuff.
Hell, i was even in the mind of picking up again my early work on the
e17 (enlightenment) port to somehow sidestep the issue...
Frankly, to me what we should paying attention to is that we are
moving away from Nano-X and our homebrew toolkit Neux.
This has a lot of benefits to everyone, that far outweight the issue
of making some of out contributors unhappy.
- We are moving to an estabilished, reliable, well-tested tookit. Some
else wrote it all already, and is making it grow for us. We cut away a
lot of maintenance baggage, saving time that can be then used to
improve the rest of the system, which is what really matters.
- It has a well-documented, well-defined API to work with. Everyone
can pick up the code and help on things, which is a bit hard to do on
Neux, since it's pretty spaghetti and undocumented. Even if you don't
usually work with QT, i would bet that understanding what a piece of
QT code does is still easier than understanding what a piece of Neux
code does.
- Besides, no one of the new hackers looking at the OSD knows what
neux is, so they are not especially encouraged to get one to hack on
it. So we get neither the QT hackers, nor the GTK hackers by staying
with Neux. By moving to one of the two, we at least attract half of
them. And 0.5 > 0 in my opinion ;)
Hope this explains better some of the thinking behind the decision.
Just some more things for you to chew on,
--
nero
Dragon Wisard
1. Volkerding is NOT a non-player and Slackware isn't any less relevant today simply because it doesn't make daily headlines. Slackware doesn't aim to make headlines. That's not it's role. Slackware is a fast, stable, conservative distribution that is well regarded even outside the Linux community. Just ask any Solaris or BSD user which distro they'd rather put on their machine if they were forced to use Linux. I was recently browsing some Linux magazines and saw an interview in which developers from Gentoo (one of your "respectable distros") mentioned that they look upstream to distributions like Slackware for patches. If other distros are following Slackware's lead, Volkerding must be doing something right.
2. Torvalds doesn't make distributions, but neither does RMS. Torvalds DOES get headlined regularly (whether it's warranted or not). Also, Torvalds usually knows what he's talking about and has a fairly good track record at making accurate predictions with regards to community trends.
3. Debian prefers the GNOME project because it's a GNU project for historical reasons. Debian did not make the decision based on technical merit (which is not to say that GNOME is inferior, but technical merit was not the primary deciding factor). Slackware dropped GNOME in favor of KDE specifically because Volkerding felt that KDE had a technical advantage over GNOME at the time and supporting both didn't seem to coincide with Slackware's goals/best interests. Support for GNOME is still available for Slackware through 3rd parties.
4. Ubuntu picked Gnome for two very obvious reasons. A) Ubuntu is Debian, Debian uses Gnome. B) Gnome is known for it's (overly?) simplistic approach to user interfaces. It presents the end user with the fewest options that might confuse or scare away newcomers. It is a commonly held notion amongst the Linux community and journalists that Gnome is easier for new users than KDE.
5. It's difficult for anyone to measure which distros are the most common with any reasonable degree of certainty. While Ubuntu, Fedora/RHEL and SuSE are obvious choices, what about PCLinuxOS (the new favorite on DistorWatch.com) which installs KDE by default? Sabayon (which is basically a binary installer for Gentoo) also defaults to KDE and has been increasing in press and popularity recently. Mandrake/Mandriva, MEPIS, and Knoppix also use KDE.
Although I may have been too hostile towards Gnome initially, the point I'm trying to make is that KDE is still a well used and supported Desktop Environment and certainly has a user base that is at least as large as Gnome.
Also, if you're going to get picky about details, Slackware doesn't actually use KDE as it's default Desktop. Although KDE is officially included, Slackware defaults to text-mode. In fact, most Slackers I know don't use either KDE or Gnome.
I also don't use much of either environment. (Fluxbox on my laptops, wmii on my desktops)
Neuros has already made their decision and they have reasonably justified it. Supporting Qt because they already have the code, resources, and know-how is a practical decision. To make the political choice of Gnome because everyone else is using it, or because it's part of GNU, or because some people don't like KDE wouldn't make any sense for a company that wants to meet real-world deadlines and deliver a commercially viable product.
Pat Gunn, does this response satisfy your inquiry?
Fernando Cassia
On Nov 26, 2007 5:52 AM, Vladimir Pantelic <p...@nt.tu-darmstadt.de> wrote:> I feel that the industry is going towards Gnome... from Nokia's LinuxKDE barely needs "life support"...
> tablets to the OpenMoko platform... and I'd be saddened to see
> Trolltech's (and by extension, KDE's) life extended further by this move...
I can't speak for the embedded scene, but it seems like some big names (at the least, Linus Torvalds and Patrick Volkerding) have picked KDE over Gnome as the desktop environment of choice for Desktop users.
Yeah, like SUSE moving to Gnome as its default beginning with 10.1. :)
http://www.linux.com/articles/54547
" In just over half an hour, SUSE 10.1 was ready for business on the desktop machine. The default desktop. SUSE's default GNOME desktop shows a sky blue ...
10.2 offered a choice of the default desktop, and OpenSUSE 10.3 is offered as separate Gnome and KDE versions (apparently to appease the small but vocal community of KDE complainers).
http://www.osnews.com/story.php/18653/Sneak-Peeks-at-openSUSE-10.3-SUSE-Polished-GNOME-2.20
"OpenSUSE has been driving innovation on the Linux desktop, and in today's serial we'll be discovering just what has been happening on the GNOME front. Among other things, openSUSE 10.3 is set to contain, and be among the very first to have, the new GNOME 2.20. We'll see what new things you can expect from this version, what additional polish openSUSE brings to the desktop, and finally we'll be talking to JP Rosevear, an openSUSE and GNOME developer, to find out a little more."
Debian 4.0 arrives
http://www.desktoplinux.com/news/NS8547892609.html
"The default desktop environment is GNOME, but alternative CD images are now available for KDE and Xfce."
I don't care if Linus Torvalds, his mother or George W. Bush uses KDE... what matters is the default desktop used by distros... and right now the most popular use:
Ubuntu: Gnome
RedHat/Fedora: Gnome
SUSE SLED: Gnome
Debian: Gnome
Only the entry-level distros like Linspire and Mandriva continue with KDE...
And like I said... in the mobile space... a lot more companies have aligned behind the Mobile Gnome initiative than KDE...
Probably a result of this
Trolltech ports Qt to Windows CE and Windows Mobile
http://www.windowsfordevices.com/news/NS3088688399.html
The logic of the above announcement is beyond me... (Qt on WinCE? why oh why?). Yeah, as if anyone had any success in convincing Windows developers to use non-microsoft tools and toolkits.... ask Borland...
FC
Fernando Cassia
FC
Pat Gunn
were talking about GNOME/KDE, which arn't exactly the same thing
as GTK/Qt - presumably nobody would suggest putting GNOME or KDE
on the OSD (both of them being too large/memory-intensive). Still,
I think the desktops built around these are relevant because they
affect the interest drawn to the toolkits. For what it's worth,
KDE and GNOME apps both feel very nice to me, and GTK and Qt apps
both have a pleasant look (far better than, say, apps built with
venerable toolkits like Tk, Motif, or Athena). Having
a lot of attention drawn around the toolkit (even if its primarily
through its big brother, be that KDE or GNOME) is important - I
want to know that there are language bindings that see a lot of use
for plenty of languages. It's been stated that maybe we should write
the non-gui parts of our apps and let a Neuros person write the top
bit - this isn't helpful - any barrier to development of this nature
is a hassle and will cause people to look elsewhere. I know the OSD
is theoretically a media device, but if it is widely enough produced,
I think it might be useful for much more (essentially a more well-rounded
Gumstix), and I'd like to see it have such usage - that does mean
keeping it easy to develop for. GNOME and KDE's steering of interest
towards their respective toolkits figures into that to the extent that
it means wide usage of the toolkits and language support for them.
Given that GTK and Qt are both available under GPL-ish licenses nowadays,
thankfully we don't need to argue or worry about licensing too much.
> Admittedly I may have overstated the case. However, I disagree with several
> details of your rebuttal.
>
> 1. Volkerding is NOT a non-player and Slackware isn't any less relevant
> today simply because it doesn't make daily headlines. Slackware doesn't aim
> to make headlines. That's not it's role. Slackware is a fast, stable,
> conservative distribution that is well regarded even outside the Linux
> community. Just ask any Solaris or BSD user which distro they'd rather put
> on their machine if they were forced to use Linux. I was recently browsing
> some Linux magazines and saw an interview in which developers from Gentoo
> (one of your "respectable distros") mentioned that they look upstream to
> distributions like Slackware for patches. If other distros are following
> Slackware's lead, Volkerding must be doing something right.
It may be fast and stable, and it may have a lot of respect (I used it
ages ago as my second distro, SLS being my first, and I still regard it
with a certain quaint fondness) with nuances attached, but long ago it
lost most of its userbase to other distros. Like OS/2 during the later
years, it's no longer very relevant, even if unlike OS/2 or AmigaOS, it
continues to be developed.
It would certainly be an overstatement to say that Gentoo is following
Slackware's lead. It makes sense to take patches from everywhere you
can, whether that means OpenBSD or Fedora.
> 2. Torvalds doesn't make distributions, but neither does RMS. Torvalds DOES
> get headlined regularly (whether it's warranted or not). Also, Torvalds
> usually knows what he's talking about and has a fairly good track record at
> making accurate predictions with regards to community trends.
Did I mention RMS somewhere? Neither of them make a distro, and so they're
not really relevant to this matter.
> 3. Debian prefers the GNOME project because it's a GNU project for
> ...... (snip)
It doesn't matter why the distros picked KDE or GNOME. What matters is
how broadly they're used and what that means for support for programming
towards GTK/Qt. I look at GTK and I see very broad language support. I
don't see quite as much with Qt. Maybe other people care about this
for other reasons (e.g. broader allegiance to one camp), but my concerns
are mostly focused around this.
> 5. It's difficult for anyone to measure which distros are the most common
> with any reasonable degree of certainty. While Ubuntu, Fedora/RHEL and SuSE
> are obvious choices, what about PCLinuxOS (the new favorite on
> DistorWatch.com) which installs KDE by default? Sabayon (which is basically
> a binary installer for Gentoo) also defaults to KDE and has been increasing
> in press and popularity recently. Mandrake/Mandriva, MEPIS, and Knoppix also
> use KDE.
PCLinuxOS is a non-player at this point. Knoppix isn't very relevant
because it's more broadly used as a bootable-install than a development system.
Mandrake/Mandriva is relevant. I don't think it's that hard to measure
mindshare, especially among the development crowd (who I think we should
focus on for this metric as they're the people who push the actual toolkit's
features).
> Although I may have been too hostile towards Gnome initially, the point I'm
> trying to make is that KDE is still a well used and supported Desktop
> Environment and certainly has a user base that is at least as large as
> Gnome.
Hostility is not a problem, but the misinformation you provided (e.g. that
respectable distros are dropping GNOME left and right) wasn't cool. If you
had been the only person following these matters in the conversation,
people would have left with misinformation, and so people might need to
keep an eye on you :)
> Neuros has already made their decision and they have reasonably justified
> it. Supporting Qt because they already have the code, resources, and
> know-how is a practical decision. To make the political choice of Gnome
> because everyone else is using it, or because it's part of GNU, or because
> some people don't like KDE wouldn't make any sense for a company that wants
> to meet real-world deadlines and deliver a commercially viable product.
I don't think it's been reasonably justified, just presented as a fait
accompli. I think it's an unfortunate choice, and while I don't intend to
make a big fuss over it, I wanted to both toss in my two cents and reply
to your comments, which I think were either dishonest or misinformed.
There may people who are like you describe, but I'm not suggesting GTK
for any of those reasons - Trolltech's release of Qt under the GPL has
satisfied me, licensewise, and I don't particularly care about being
on the GNU bandwagon in itself. These matters may have some importance
purely for the kinds and numbers of eyes they bring to the toolkit itself, but
they're not important in themselves.
> Pat Gunn, does this response satisfy your inquiry?
What kind of a question is that? :)
Dragon Wisard
On Nov 26, 2007 12:08 PM, Dragon Wisard <dragon...@gmail.com> wrote:On Nov 26, 2007 5:52 AM, Vladimir Pantelic <p...@nt.tu-darmstadt.de> wrote:> I feel that the industry is going towards Gnome... from Nokia's LinuxKDE barely needs "life support"...
> tablets to the OpenMoko platform... and I'd be saddened to see
> Trolltech's (and by extension, KDE's) life extended further by this move...
I can't speak for the embedded scene, but it seems like some big names (at the least, Linus Torvalds and Patrick Volkerding) have picked KDE over Gnome as the desktop environment of choice for Desktop users.
Yeah, like SUSE moving to Gnome as its default beginning with 10.1. :)
Read my previous post.
I don't care if Linus Torvalds, his mother or George W. Bush uses KDE... what matters is the default desktop used by distros... and right now the most popular use:
Ubuntu: Gnome
RedHat/Fedora: Gnome
SUSE SLED: Gnome
Debian: Gnome
How did you determine "popularity" here? The limited accuracy of statistical information with regards to Linux usage is a very well documented problem.
Also, I'm glad you have such low regard for the person who maintains the kernel for the OS you're probably using right now. You can trust him build stable, secure kernel but consider him an idiot when voices his opinion on projects that are dependent on his work. (Granted neither KDE or Gnome are reliant on the kernel directly, but the distributions that propagate their use are directly and completely dependent on it, with the possible exception of Gentoo and Debian which run on Hurd and BSD.)
I am also glad (not sarcastically, this time) that you hold GWB in such low regard. I voted for the other guy.
And like I said... in the mobile space... a lot more companies have aligned behind the Mobile Gnome initiative than KDE...
Probably a result of this
Trolltech ports Qt to Windows CE and Windows Mobile
http://www.windowsfordevices.com/news/NS3088688399.html
The logic of the above announcement is beyond me... (Qt on WinCE? why oh why?). Yeah, as if anyone had any success in convincing Windows developers to use non-microsoft tools and toolkits.... ask Borland...
Umm... PORTABILITY!
It's not like Microsoft is a small player. There are many devices that currently run WinCE, having Qt running on them means that as a developer, your application will run on more devices with less porting effort. Also, there are a large number of Qt applications on mobile devices like cell phones. The OpenMoko project has barely gotten itself off the ground and a lot of the apps on Maemo don't seem relevant to what Neuros is doing (imho).
The traditional problem with using non-microsoft tools/toolkits is that users have to download "extra stuff" to run your apps. If you negotiate a packaging deal with device manufacturers to pre-install your libraries you eliminate that as a issue. Especially in the mobile world, manufacturers will WANT to do this because they want as much 3rd party software running on their devices as possible (it increases the usefulness and value of their device). Java is doing fine on Windows, ask Sun.
MMMokay, now that my nemesis on this list has stated that the decision to go Qt is set in stone, I'm willing to let this discussion die. I'm not going to start a pro-GTK flame war. :)
FC
--
http://www.dragonwisard.net/
Dragon Wisard
It may be fast and stable, and it may have a lot of respect (I used it
ages ago as my second distro, SLS being my first, and I still regard it
with a certain quaint fondness) with nuances attached, but long ago it
lost most of its userbase to other distros. Like OS/2 during the later
years, it's no longer very relevant, even if unlike OS/2 or AmigaOS, it
continues to be developed.
I still disagree with your surmisal that Slackware is a minor distro. I think this is a common misconception. However we are probably measuring by different standards considering the distributions you picked as "respectable" (not saying they aren't).
It would certainly be an overstatement to say that Gentoo is following
Slackware's lead. It makes sense to take patches from everywhere you
can, whether that means OpenBSD or Fedora.Did I mention RMS somewhere? Neither of them make a distro, and so they're
> 2. Torvalds doesn't make distributions, but neither does RMS. Torvalds DOES
> get headlined regularly (whether it's warranted or not). Also, Torvalds
> usually knows what he's talking about and has a fairly good track record at
> making accurate predictions with regards to community trends.
not really relevant to this matter.
> ...... (snip)
> 3. Debian prefers the GNOME project because it's a GNU project for
It doesn't matter why the distros picked KDE or GNOME. What matters is
how broadly they're used and what that means for support for programming
towards GTK/Qt. I look at GTK and I see very broad language support. I
don't see quite as much with Qt. Maybe other people care about this
for other reasons (e.g. broader allegiance to one camp), but my concerns
are mostly focused around this.
There are even more people using the Windows XP shell. Does that mean it's better? I haven't written any GUI's with either, but there are a comparable number of applications written in either GTK or QT, and a large number that are written in both. I'm not qualified to say which is the best from a technical perspective, but it looks like either technical merit has little influence on usage, or the two are comparably well written.
PCLinuxOS is a non-player at this point. Knoppix isn't very relevant
> 5. It's difficult for anyone to measure which distros are the most common
> with any reasonable degree of certainty. While Ubuntu, Fedora/RHEL and SuSE
> are obvious choices, what about PCLinuxOS (the new favorite on
> DistorWatch.com) which installs KDE by default? Sabayon (which is basically
> a binary installer for Gentoo) also defaults to KDE and has been increasing
> in press and popularity recently. Mandrake/Mandriva, MEPIS, and Knoppix also
> use KDE.
because it's more broadly used as a bootable-install than a development system.
Mandrake/Mandriva is relevant. I don't think it's that hard to measure
mindshare, especially among the development crowd (who I think we should
focus on for this metric as they're the people who push the actual toolkit's
features).
Ok, I mention PCLinuxOS because it's recently been grabbing headlines (probably because of it's coup over Ubuntu on the DistroWatch ranking). I disagree with your dismissal of Knoppix, however. It doesn't matter *how* a distribution is used, by your previous statements, use and public exposure is enough to merit popularity.
I don't think it's been reasonably justified, just presented as a fait
> Neuros has already made their decision and they have reasonably justified
> it. Supporting Qt because they already have the code, resources, and
> know-how is a practical decision. To make the political choice of Gnome
> because everyone else is using it, or because it's part of GNU, or because
> some people don't like KDE wouldn't make any sense for a company that wants
> to meet real-world deadlines and deliver a commercially viable product.
accompli. I think it's an unfortunate choice, and while I don't intend to
make a big fuss over it, I wanted to both toss in my two cents and reply
to your comments, which I think were either dishonest or misinformed.
There may people who are like you describe, but I'm not suggesting GTK
for any of those reasons - Trolltech's release of Qt under the GPL has
satisfied me, licensewise, and I don't particularly care about being
on the GNU bandwagon in itself. These matters may have some importance
purely for the kinds and numbers of eyes they bring to the toolkit itself, but
they're not important in themselves.
I personally feel that the justifications Neuros has provides are reasonable. I would like to know why, specifically, you feel GTK is a better choice (especially since you have discounted my earlier generalizations). Besides, Neuros is an open project. There is nothing to prevent anyone from porting GTK to the OSD. In fact, I believe Joe would probably applaud the effort.
What kind of a question is that? :)
> Pat Gunn, does this response satisfy your inquiry?
Let me state it differently this time. Did I sufficiently address all the concerns you raised, and what might I have gotten wrong?
P.S.: I'm the worst kind of troll. Sorry mate.
---
Pat Gunn
mod: csna, bmcm, bmco, cooa, cona, clpd, coom
http://dachte.org
"Once the game is over, the King and the pawn go back in the same box"
--Italian Proverb
--
http://www.dragonwisard.net/
luka
back
being mostly low-level C coder with no app/GUI experience I took part
in
Sharp Zaurus/Trolltech contest (BTW, might be a very good idea for
OSD
as well. Sharp's got a lot of attention and bunch of very good
applications
for the price of few PDAs). I haven't won anything :-(, yet was
rewarded by
getting experience of writing nice app (media catalogue - DVDs, tapes,
CDs...)
relatively fast due to a well designed APIs and excellent
documentation.
Another nice QT feature is QVFB (IIRC, QT virtual frame buffer?) which
allows
to run app on a development host w/o need to load it onto target
device.
That cut debugging time a lot.
This move will make my life much easier as otherwise hacking current
OSD
GUI as part of the feature I'm planning to implement could be a
significant
challenge for me.
lpotter
platform API out there.
Fernando Cassia
I think this is an excellent move! Qt is really the best cross
platform API out there.
Here, have a cookie. ;)
FC
Joe Born
>
> Sounds interesting. Did you buy a batch of ancient 32MB CF cards or
> something?
>
> I really love using a 8GB Seagate microdrive in the OSD's CF slot. :(
>
> If given the option, I'd really would love being able to select which
> slot acts as the "non-removable" storage...
>
> Another solution would be inserting a CF/SD USB multi-reader into the
> SD slot... but I'm not sure it'd be supported given the USB port
> driver problems....
There are a host of questions like this from users on the forum too
http://forums.neurostechnology.com/index.php?topic=8360.15 look at page
two. Can someone go on there and answer as best you can. I'm not shy,
but I'm not sure what the answers are myself.
Ugo Riboni
I added a reply to the topic.
Instead of wasting bandwidth to copy and paste it here, you can go read it at:
http://forums.neurostechnology.com/index.php?topic=8360.msg44456#msg44456
Cheers,
--
nero