Discussion:
do you know when MeeGo 1.2.0.2 will be available for ivi
Rich, Glenn
2011-08-10 08:52:26 UTC
Permalink
Thanks
Glenn
---------------------------------------------------------------------
Intel Corporation (UK) Limited
Registered No. 1134945 (England)
Registered Office: Pipers Way, Swindon SN3 1RJ
VAT No: 860 2173 47

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
Clark, Joel
2011-08-10 16:32:23 UTC
Permalink
It should be 6 weeks after the last update (July 15). Updates are supposed to happen every 6 weeks. You can get a pre-release copy at <http://repo.meego.com/MeeGo/builds/1.2.0/latest> repo.meego.com/MeeGo/builds/1.2.0/latest<http://repo.meego.com/MeeGo/builds/1.2.0/latest>


Regards
Joel

On Aug 10, 2011, at 1:54 AM, "Rich, Glenn" <glenn.rich-***@public.gmane.org<mailto:glenn.rich-***@public.gmane.org>> wrote:

Thanks
Glenn
---------------------------------------------------------------------
Intel Corporation (UK) Limited
Registered No. 1134945 (England)
Registered Office: Pipers Way, Swindon SN3 1RJ
VAT No: 860 2173 47

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
Nasa
2011-08-12 10:49:31 UTC
Permalink
Hi gang,

Having been watching bug reports and updates to packages in
the different repos, I was wondering what activity has been
going on for IVI(outside the API discussions on the list)?
What fixes (if any) are planned for 1.2.0 update 2? Is
there going to be a 1.2.1 or will that go directly to 1.3?
I haven't seen any updates related specfically to IVI for
the 1.2.0.90 branch (although some of the media apps have
changed in ways that would be useful to IVI)... Did I miss something?

The WIKI (roadmap && main page) info doesn't address these questions,
and the road has come to an end as far as the roadmap is concerned.

Nasa

BTW: if I come across harsh, that's not my intent (just wondering what
the plans are)

----- Original Message -----
Post by Clark, Joel
It should be 6 weeks after the last update (July 15). Updates are
supposed to happen every 6 weeks. You can get a pre-release copy at
<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
repo.meego.com/MeeGo/builds/1.2.0/latest<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
Regards
Joel
On Aug 10, 2011, at 1:54 AM, "Rich, Glenn"
Thanks
Glenn
---------------------------------------------------------------------
Intel Corporation (UK) Limited
Registered No. 1134945 (England)
Registered Office: Pipers Way, Swindon SN3 1RJ
VAT No: 860 2173 47
This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Clark, Joel
2011-08-15 15:54:28 UTC
Permalink
Hi gang,

Sorry I've been on vacation.

There are a couple of IVI specific changes in 1.2.0 Update2. There is a fix in the gnome-packagekit, to fix some flickering we saw with QML, GLES, and the EMGD graphics driver for IA PowerVR graphics. But for the most part we aren't working on a lot of update changes. Of course we will happily add any fixes submitted by the community.

We are also trying to get an updated version of the EMGD graphics driver, version 1.8 added to Update2, but it may miss the cutoff date. Updates are scheduled every 6 weeks, so we can always add whatever fixes come along.

We are not doing any IVI specific work in 1.2.0.90. Branch 1.2.0.90 is the pre-release version of 1.2.1. This branch was established for some platforms that wanted to continue to do UX specific development based on 1.2 core after the 1.2 release. Since the UX that boots by default in MeeGo IVI is only a "sample" and we expect folks will generally provide their own UX, we are not trying to track the ongoing platform specific UX work being done in 1.2.0.90.

So the next major release for MeeGo IVI is being worked on in Trunk.

One change I want to make for the next major MeeGo IVI release is to revert to a default boot to developer screen (xterm). This will enable a wider variety of platforms by removing dependencies in the current sample UX on specific boards and display configurations to successfully boot. We will provide some guidance or scripts to show how to enable the sample UX and would like to encourage alternate UX development. Perhaps we can include more than 1 sample UX that can be turned on after first boot.

I'll work on an update to the wiki to capture this information this week.

regards
Joel Clark
MeeGo IVI PM.


-----Original Message-----
From: Nasa [mailto:nasa01-***@public.gmane.org]
Sent: Friday, August 12, 2011 3:50 AM
To: Clark, Joel
Cc: meego-ivi-***@public.gmane.org; Rich, Glenn
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available for ivi

Hi gang,

Having been watching bug reports and updates to packages in
the different repos, I was wondering what activity has been
going on for IVI(outside the API discussions on the list)?
What fixes (if any) are planned for 1.2.0 update 2? Is
there going to be a 1.2.1 or will that go directly to 1.3?
I haven't seen any updates related specfically to IVI for
the 1.2.0.90 branch (although some of the media apps have
changed in ways that would be useful to IVI)... Did I miss something?

The WIKI (roadmap && main page) info doesn't address these questions,
and the road has come to an end as far as the roadmap is concerned.

Nasa

BTW: if I come across harsh, that's not my intent (just wondering what
the plans are)

----- Original Message -----
Post by Clark, Joel
It should be 6 weeks after the last update (July 15). Updates are
supposed to happen every 6 weeks. You can get a pre-release copy at
<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
repo.meego.com/MeeGo/builds/1.2.0/latest<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
Regards
Joel
On Aug 10, 2011, at 1:54 AM, "Rich, Glenn"
Thanks
Glenn
---------------------------------------------------------------------
Intel Corporation (UK) Limited
Registered No. 1134945 (England)
Registered Office: Pipers Way, Swindon SN3 1RJ
VAT No: 860 2173 47
This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Nasa
2011-08-15 16:31:36 UTC
Permalink
----- Original Message -----
Post by Nasa
Hi gang,
Sorry I've been on vacation.
There are a couple of IVI specific changes in 1.2.0 Update2. There is
a fix in the gnome-packagekit, to fix some flickering we saw with QML,
GLES, and the EMGD graphics driver for IA PowerVR graphics. But for
the most part we aren't working on a lot of update changes. Of course
we will happily add any fixes submitted by the community.
We are also trying to get an updated version of the EMGD graphics
driver, version 1.8 added to Update2, but it may miss the cutoff date.
Updates are scheduled every 6 weeks, so we can always add whatever
fixes come along.
We are not doing any IVI specific work in 1.2.0.90. Branch 1.2.0.90 is
the pre-release version of 1.2.1. This branch was established for some
platforms that wanted to continue to do UX specific development based
on 1.2 core after the 1.2 release. Since the UX that boots by default
in MeeGo IVI is only a "sample" and we expect folks will generally
provide their own UX, we are not trying to track the ongoing platform
specific UX work being done in 1.2.0.90.
So the next major release for MeeGo IVI is being worked on in Trunk.
One change I want to make for the next major MeeGo IVI release is to
revert to a default boot to developer screen (xterm). This will enable
a wider variety of platforms by removing dependencies in the current
sample UX on specific boards and display configurations to
successfully boot. We will provide some guidance or scripts to show
how to enable the sample UX and would like to encourage alternate UX
development. Perhaps we can include more than 1 sample UX that can be
turned on after first boot.
Joel,

I think this is a great idea (had been thinking about suggesting this
for a while now). Just have to make sure to document what the baseline
actually is...

Nasa
Post by Nasa
I'll work on an update to the wiki to capture this information this week.
regards
Joel Clark
MeeGo IVI PM.
-----Original Message-----
Sent: Friday, August 12, 2011 3:50 AM
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available for ivi
Hi gang,
Having been watching bug reports and updates to packages in
the different repos, I was wondering what activity has been
going on for IVI(outside the API discussions on the list)?
What fixes (if any) are planned for 1.2.0 update 2? Is
there going to be a 1.2.1 or will that go directly to 1.3?
I haven't seen any updates related specfically to IVI for
the 1.2.0.90 branch (although some of the media apps have
changed in ways that would be useful to IVI)... Did I miss something?
The WIKI (roadmap && main page) info doesn't address these questions,
and the road has come to an end as far as the roadmap is concerned.
Nasa
BTW: if I come across harsh, that's not my intent (just wondering what
the plans are)
----- Original Message -----
Post by Clark, Joel
It should be 6 weeks after the last update (July 15). Updates are
supposed to happen every 6 weeks. You can get a pre-release copy at
<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
repo.meego.com/MeeGo/builds/1.2.0/latest<http://repo.meego.com/MeeGo/builds/1.2.0/latest>
Regards
Joel
On Aug 10, 2011, at 1:54 AM, "Rich, Glenn"
Thanks
Glenn
---------------------------------------------------------------------
Intel Corporation (UK) Limited
Registered No. 1134945 (England)
Registered Office: Pipers Way, Swindon SN3 1RJ
VAT No: 860 2173 47
This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or
distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Jeremiah Foster
2011-08-16 09:16:04 UTC
Permalink
Post by Clark, Joel
So the next major release for MeeGo IVI is being worked on in Trunk.
Are you planning on adding things like DLT, or NFS?
Post by Clark, Joel
One change I want to make for the next major MeeGo IVI release is to revert to a default boot to developer screen (xterm).  This will enable a wider variety of platforms by removing dependencies in the current sample UX on specific boards and display configurations to successfully boot.  We will provide some guidance or scripts to show how to enable the sample UX and would like to encourage alternate UX development.  Perhaps we can include more than 1 sample UX that can be turned on after first boot.
I think it might be useful if the UX where included as a package in
the repos but not shipped with the base image. Or one could perhaps
deliver two images, one with UXes and one without but then one can add
a particular UX later via the package manager. A small embedded image
would be particularly useful for integrators who'd like to test and
add their software to MeeGo IVI without having to go through the MeeGo
OBS route.

Regards,

Jeremiah
Clark, Joel
2011-08-16 14:35:41 UTC
Permalink
Post by Jeremiah Foster
Are you planning on adding things like DLT, or NFS?
DLT and nfs-utils are already in Trunk.
Post by Jeremiah Foster
I think it might be useful if the UX where included as a package in
the repos but not shipped with the base image. Or one could perhaps
deliver two images, one with UXes and one without but then one can add
a particular UX later via the package manager. A small embedded image
would be particularly useful for integrators who'd like to test and
add their software to MeeGo IVI without having to go through the MeeGo
OBS route.
Regards,
Jeremiah
I like the idea of installable UX(s) and smaller image. We wanted to go the 2 image (small and large) in the 1.2 timeframe, but never seemed to have time to make it happen

regards
Joel
Jeremiah Foster
2011-08-16 15:47:58 UTC
Permalink
Post by Clark, Joel
Post by Jeremiah Foster
Are you planning on adding things like DLT, or NFS?
DLT and nfs-utils are already in Trunk.
I'm wondering if they'll be included in the upcoming release.

Regards,

Jeremiah
Clark, Joel
2011-08-16 15:51:39 UTC
Permalink
Post by Jeremiah Foster
I'm wondering if they'll be included in the upcoming release.
Regards,
Jeremiah
Yes of course. Once it's in Trunk specific steps have to be taken to remove it from the release, so I always think of them as same/same.

regards
Joel
Alison Chaiken
2011-08-16 14:23:01 UTC
Permalink
Post by Clark, Joel
There is a fix in the gnome-packagekit, to fix some flickering we saw with QML, GLES, and the EMGD graphics driver for IA PowerVR graphics.
Is the new EMGD driver available now, and is it different than the one
in trunk/latest? Will it work with all the different Atom variants
(Pinetrail, Cedartrail, Medfield . . .) ?

Thanks,
Alison
--
Alison Chaiken
(650) 279-5600  (cell)
         http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
Clark, Joel
2011-08-16 14:55:11 UTC
Permalink
Post by Alison Chaiken
Post by Clark, Joel
There is a fix in the gnome-packagekit, to fix some flickering we saw
with QML, GLES, and the EMGD graphics driver for IA PowerVR graphics.
Is the new EMGD driver available now, and is it different than the one
in trunk/latest? Will it work with all the different Atom variants
(Pinetrail, Cedartrail, Medfield . . .) ?
Thanks,
Alison
The new EMGD driver (version 1.8) is available now from http://edc.intel.com/Software/Downloads/EMGD/#download
It is not in Trunk yet. It supports the same chipsets as the emgd-bin driver in Trunk:non-oss now. It supports US15W (Poulsbo/eMenlow) and ES6xx series (Tunnel Creek) chipsets.

It does not support Pinetrail, CedarTrail or Medfield. Pinetrail uses an entirely different graphics core, and while CedarTrail and Medfield have similar graphics cores, unfortunately there are enough differences in the hardware that they all have different graphics drivers. This comes from the fact they are all designed for different market segments and have different customer requirements.

regards
Joel
Alison Chaiken
2011-08-16 15:13:16 UTC
Permalink
 It supports the same chipsets as the emgd-bin driver in Trunk:non-oss
now. It supports US15W (Poulsbo/eMenlow) and ES6xx series (Tunnel
Creek) chipsets. It does not support Pinetrail, CedarTrail or Medfield.
I take it then that we should only use EMGD drivers that are labelled
as being appropriate for our particular chipset, as otherwise we met
get incompatible PowerVR kernel modules. Presumably then in using
mic2 to build images, we must exclude any EMGD components that don't
come from the $ARCH Support package group, e.g. "Oaktrail Support" .
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
Clark, Joel
2011-08-16 15:39:20 UTC
Permalink
Post by Alison Chaiken
I take it then that we should only use EMGD drivers that are labelled
as being appropriate for our particular chipset, as otherwise we met
get incompatible PowerVR kernel modules. Presumably then in using
mic2 to build images, we must exclude any EMGD components that don't
come from the $ARCH Support package group, e.g. "Oaktrail Support" .
When using mic2 to build Oaktrail images you should not include ANY EMGD components.
You should not mix up graphics driver components at all. Not even things like libGLESv2.so
They are not compatible.

regards
Joel
Alison Chaiken
2011-08-16 15:54:34 UTC
Permalink
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.

I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
Clark, Joel
2011-08-16 17:03:24 UTC
Permalink
Post by Alison Chaiken
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
The RPM dependencies in the netbook image should clearly identify to zypper what the correct packages to update are. This should work fine without a user/developer need to know the details.
Post by Alison Chaiken
It's already a bit confusing about which graphics
RPMs to use with which Atom processors.
I feel your pain too.
Post by Alison Chaiken
Perhaps the EMGD RPMs could have "sandybridge" or whatever in the package name.
Sandybridge is "GenX" and uses the Intel Integrated (in Linux upstream) Graphics driver. It does not use EMGD.
The package names would get just as confusing as the processor/graphics roadmap then.

regards
Joel
Grigory Abuladze
2011-08-17 08:57:50 UTC
Permalink
Alison

I have to say that 1.2 release has already dropped all of us who use ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later

I have no idea why this way was broken in 1.2 release, but it drops out a lot of people. They tried "new" release and it just doesn't work for them - so they simple walk away to do something for Android or iOS.
I remember list member NASA has provided some "workaround" and it was quite popular topic ). But not all of us can/want to spend time to try dancing around to "fix" drivers issue.

I would like to extend your proposal to keep it general and end user will decide which HW specific driver to install. And yes - proper naming will be definitely helpful here.

Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147

-----Original Message-----
From: Alison Chaiken [mailto:alchaiken-***@public.gmane.org]
Sent: 16 August 2011 16:55
To: Clark, Joel
Cc: meego-ivi-***@public.gmane.org
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi

Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.

I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
Nasa
2011-08-17 11:27:52 UTC
Permalink
Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-***@lists.meego.com/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Jeremiah Foster
2011-08-17 14:11:36 UTC
Permalink
Post by Nasa
Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user.
But isn't this reasonable since aren't the primary customers of an IVI
system going to be companies? Are individuals really going to hack
their IVI system? I would think they would be hesitant to burn a
warranty for a $30,000 car.
Post by Nasa
This statement
isn't meant as a knock on anyone, it's more of an observation.  Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.
There is actually a large amount of hardware that is similar to an IVI
production system available nowadays. You may have to consider other
architectures than just x86 however. ;)
Post by Nasa
The problem is that mixed signals were sent out about this at the begining of
the Meego project.  So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).
*sigh*

Regards,

Jeremiah
Andy Ross
2011-08-17 16:28:25 UTC
Permalink
Post by Jeremiah Foster
But isn't this reasonable since aren't the primary customers of an IVI
system going to be companies? Are individuals really going to hack
their IVI system? I would think they would be hesitant to burn a
warranty for a $30,000 car.
There is a huge market for aftermarket upgrades for cars. I'd say a
hackable IVI system would be a big part of this if it existed. But
yes: clearly Intel's goal here is sales to auto manufacturers, not
private garages and chop shops.

That said, there's nothing particularly unique about the existing IVI
images. Their kernels and the EMGD driver are specific to the Tunnel
Creek hardware platform (and really the Intel Crossville-OKI board
specifically), but everything above that will run on any MeeGo.

There's nothing stopping someone from cooking an "IVI" image based on
a pinetrail kernel and Xorg intel graphics drivers. Or for that
matter just hand-updating a netbook or tablet image with the IVI rpms.
I've done bits of the latter for testing purposes, though I haven't
gone as far as generating something installable.

Andy
Clark, Joel
2011-08-17 14:54:38 UTC
Permalink
I apologize for the pain. :-). Actually I've felt quite a bit if it too when it comes to trying to create a visually compelling experience on embedded HW systems. We have found considerable HW and driver specific dependencies. In general the idea was that each platform champion would provide the adaptations required for their platform. This would probably have produced multiple sample IVI boot images for different platforms. So far this has not happened. We have been struggling with the resources we have just to keep a couple of platforms adaptations working sufficiently to enable the graphically rich apps we borrowed from other devices, and voice recognition and text to speech, and touchscreen all working. I think all of these are essential to an IVI experience and should at least be able to be demonstrated on platforms. Same explanation for the Rotary Dial support, we just have not had enough folks contributing support for different I/O devices to enable more that the one sample dial.

But trying to preserve a consumer out of the box experience has proven to be a hindrance to enabling innovation on a wide variety of platforms, particularly since the IVI HMI/UX landscape is so varied. Which is why I suggested moving to a developer oriented out of the box experience.

I deeply appreciate the folks like Nasa and Alison who have not only preserved through the multiple variations and change to MeeGo direction that have made following MeeGo IVI difficult, but have continued to contribute their time and knowledge to help the rest of the Community despite the difficulties. You guys do have impact, perhaps more so than the large but silent corporate teams that prefer to stay one step removed from active participation. Any lack of change in response to your comments is primarily a result of lack of developers to change the bits. This is certainly the case in the absence of an IVI build for the S10-3T and Pandaboard etc which we have really wanted to be able to do, but just have not had folks to make it happen.

(with my stupid mail client, I also appreciate being able to top-post :-) )

regards
Joel


-----Original Message-----
From: Nasa [mailto:***@comcast.net]
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Cc: meego-***@lists.meego.com; Alison Chaiken; Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo

Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-***@lists.meego.com/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Alison Chaiken
2011-08-17 15:40:39 UTC
Permalink
Joel writes:


We have been struggling with the resources we have just to keep a couple of
platforms adaptations working sufficiently to enable the graphically rich
apps we borrowed from other devices, and voice recognition and text to
speech, and touchscreen all working. [ . . . ] This is certainly the
case in the absence of an IVI build for the S10-3T and Pandaboard etc which
we have really wanted to be able to do, but just have not had folks to make
it happen.

Joel, I appreciate your engagement with us. Intel's commitment to open
governance of MeeGo and open discussion is exemplary, as is the AppUp
Developer Program. Intel's generosity in event sponsorship and hardware
giveaways cannot be called into question.

What's missing in my opinion is *one* long-term-supported official IVI
Developer Supported Hardware Platform analogous to the TI Pandaboard or the
Samsung Origen or the Congatech board. If Intel devoted just one employee
to maintaining a repo and a wiki for an officially supported Atom developer
board, that would make all the difference to those developers and small
businesses who want to create IVI-specific apps, not struggle with
touchscreen drivers. For someone new to MeeGo IVI, there is considerable
effort involved in figuring out which HW is most compatible, ordering the
board and accessories, finding the best repos, etc. This is a shame even
from the Intel marketing point of view, as the customers I speak to are
extremely interested in MeeGo on Atom for IVI applications, and they need a
lot of help and patience to get started.

Intel must have spent a lot of money giving away Ideapads and ExoPCs. The
developer outreach program from the IVI point of view appears to be at 95%
of its goal, with a shortfall in the 5% of support. There is still time
to fix the problem, but given the long lifecycle for automotive products,
I'm concerned that MeeGo IVI may miss the boat for the next generation.
--
Alison Chaiken
(650) 279-5600 (cell)
http://www.exerciseforthereader.org/
"What you do for a living is not be creative, what you do is ship." -- Seth
Godin
Jeremiah Foster
2011-08-18 15:22:09 UTC
Permalink
Post by Clark, Joel
We have been struggling with the resources we have just to keep a couple of
platforms adaptations working sufficiently to enable the graphically rich
apps we borrowed from other devices, and voice recognition and text to
speech, and touchscreen all working.  [  .  .   . ]  This is certainly the
case in the absence of an IVI build for the S10-3T and Pandaboard etc which
we have really wanted to be able to do, but just have not had folks to make
it happen.
Joel, I appreciate your engagement with us.    Intel's commitment to open
governance of MeeGo and open discussion is exemplary,
Sorry, I must take exception to this. The governance has been sorely
lacking in my opinion. The TSG has disappeared amongst other problems.
I don't see how you consider this "exemplary" especially when you
compare to projects like SPI or Debian which have real open
governance.

Regards,

Jeremiah

Grigory Abuladze
2011-08-17 17:10:48 UTC
Permalink
Let me try to explain it from this point:
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is spend on optimization and/or what will be feasible in near future we would like to use Sandy Bridge based tablet.

This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video driver is embedded into release image!

Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by embedding EMGD and how many pain and problems it brings?

I don't see problem in validating IVI only on specific HW and in same time releasing generic image(s). It will left door slightly open for everyone else who don't have this HW.

Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.


-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

I apologize for the pain. :-). Actually I've felt quite a bit if it too when it comes to trying to create a visually compelling experience on embedded HW systems. We have found considerable HW and driver specific dependencies. In general the idea was that each platform champion would provide the adaptations required for their platform. This would probably have produced multiple sample IVI boot images for different platforms. So far this has not happened. We have been struggling with the resources we have just to keep a couple of platforms adaptations working sufficiently to enable the graphically rich apps we borrowed from other devices, and voice recognition and text to speech, and touchscreen all working. I think all of these are essential to an IVI experience and should at least be able to be demonstrated on platforms. Same explanation for the Rotary Dial support, we just have not had enough folks contributing support for different I/O devices to enable more that the one sample dial.

But trying to preserve a consumer out of the box experience has proven to be a hindrance to enabling innovation on a wide variety of platforms, particularly since the IVI HMI/UX landscape is so varied. Which is why I suggested moving to a developer oriented out of the box experience.

I deeply appreciate the folks like Nasa and Alison who have not only preserved through the multiple variations and change to MeeGo direction that have made following MeeGo IVI difficult, but have continued to contribute their time and knowledge to help the rest of the Community despite the difficulties. You guys do have impact, perhaps more so than the large but silent corporate teams that prefer to stay one step removed from active participation. Any lack of change in response to your comments is primarily a result of lack of developers to change the bits. This is certainly the case in the absence of an IVI build for the S10-3T and Pandaboard etc which we have really wanted to be able to do, but just have not had folks to make it happen.

(with my stupid mail client, I also appreciate being able to top-post :-) )

regards
Joel


-----Original Message-----
From: Nasa [mailto:nasa01-***@public.gmane.org]
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken; Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo

Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-ivi-***@public.gmane.org/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Clark, Joel
2011-08-17 17:23:05 UTC
Permalink
We did it because
1) Sandybridge was never identified as a target platform for the meego-ivi-ia32 bootable image.
2) Nobody proposed, planned, resourced or contributed to creating a bootable IVI image for Sandybridge.
3) the IVI platforms identified as target the meego-ivi-ia32 bootable image all are supported by the "wrong" video driver (EMGD)
4) the QML based applications developed for MeeGo 1.2 ran dog slow without graphics HW acceleration.

If anyone wants to propose, plan, resource or contribute to creating a bootable IVI image for Sandybridge, I would be very happy to provide my support for their efforts.
The key is to have a kickstart file to build the image, someone to test the results and fix any issues reported.

regards
Joel


-----Original Message-----
From: Grigory Abuladze [mailto:gabuladze-2by2kkC8J4GX/***@public.gmane.org]
Sent: Wednesday, August 17, 2011 10:11 AM
To: Clark, Joel; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available


Let me try to explain it from this point:
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is spend on optimization and/or what will be feasible in near future we would like to use Sandy Bridge based tablet.

This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video driver is embedded into release image!

Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by embedding EMGD and how many pain and problems it brings?

I don't see problem in validating IVI only on specific HW and in same time releasing generic image(s). It will left door slightly open for everyone else who don't have this HW.

Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.


-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

I apologize for the pain. :-). Actually I've felt quite a bit if it too when it comes to trying to create a visually compelling experience on embedded HW systems. We have found considerable HW and driver specific dependencies. In general the idea was that each platform champion would provide the adaptations required for their platform. This would probably have produced multiple sample IVI boot images for different platforms. So far this has not happened. We have been struggling with the resources we have just to keep a couple of platforms adaptations working sufficiently to enable the graphically rich apps we borrowed from other devices, and voice recognition and text to speech, and touchscreen all working. I think all of these are essential to an IVI experience and should at least be able to be demonstrated on platforms. Same explanation for the Rotary Dial support, we just have not had enough folks contributing support for different I/O devices to enable more that the one sample dial.

But trying to preserve a consumer out of the box experience has proven to be a hindrance to enabling innovation on a wide variety of platforms, particularly since the IVI HMI/UX landscape is so varied. Which is why I suggested moving to a developer oriented out of the box experience.

I deeply appreciate the folks like Nasa and Alison who have not only preserved through the multiple variations and change to MeeGo direction that have made following MeeGo IVI difficult, but have continued to contribute their time and knowledge to help the rest of the Community despite the difficulties. You guys do have impact, perhaps more so than the large but silent corporate teams that prefer to stay one step removed from active participation. Any lack of change in response to your comments is primarily a result of lack of developers to change the bits. This is certainly the case in the absence of an IVI build for the S10-3T and Pandaboard etc which we have really wanted to be able to do, but just have not had folks to make it happen.

(with my stupid mail client, I also appreciate being able to top-post :-) )

regards
Joel


-----Original Message-----
From: Nasa [mailto:nasa01-***@public.gmane.org]
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken; Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo

Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-ivi-***@public.gmane.org/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Grigory Abuladze
2011-08-18 08:45:29 UTC
Permalink
Joel

It looks like we have misunderstanding here - what I am trying to say is:
- 1.1 was usable on wide range of HW
- 1.2 suddenly become usable on _VERY_ limited HW
And the answer to question "why?" is somewhat strange.
That's it - as simple as it looks.

And proposal was simple as well - please keep video driver more generic by default.

It would be great to have SandyBridge support, but it isn't main intention of discussion.

Regards
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.

-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 18:23
To: Grigory Abuladze; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

We did it because
1) Sandybridge was never identified as a target platform for the meego-ivi-ia32 bootable image.
2) Nobody proposed, planned, resourced or contributed to creating a bootable IVI image for Sandybridge.
3) the IVI platforms identified as target the meego-ivi-ia32 bootable image all are supported by the "wrong" video driver (EMGD)
4) the QML based applications developed for MeeGo 1.2 ran dog slow without graphics HW acceleration.

If anyone wants to propose, plan, resource or contribute to creating a bootable IVI image for Sandybridge, I would be very happy to provide my support for their efforts.
The key is to have a kickstart file to build the image, someone to test the results and fix any issues reported.

regards
Joel


-----Original Message-----
From: Grigory Abuladze [mailto:gabuladze-2by2kkC8J4GX/***@public.gmane.org]
Sent: Wednesday, August 17, 2011 10:11 AM
To: Clark, Joel; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available


Let me try to explain it from this point:
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is spend on optimization and/or what will be feasible in near future we would like to use Sandy Bridge based tablet.

This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video driver is embedded into release image!

Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by embedding EMGD and how many pain and problems it brings?

I don't see problem in validating IVI only on specific HW and in same time releasing generic image(s). It will left door slightly open for everyone else who don't have this HW.

Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.


-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

I apologize for the pain. :-). Actually I've felt quite a bit if it too when it comes to trying to create a visually compelling experience on embedded HW systems. We have found considerable HW and driver specific dependencies. In general the idea was that each platform champion would provide the adaptations required for their platform. This would probably have produced multiple sample IVI boot images for different platforms. So far this has not happened. We have been struggling with the resources we have just to keep a couple of platforms adaptations working sufficiently to enable the graphically rich apps we borrowed from other devices, and voice recognition and text to speech, and touchscreen all working. I think all of these are essential to an IVI experience and should at least be able to be demonstrated on platforms. Same explanation for the Rotary Dial support, we just have not had enough folks contributing support for different I/O devices to enable more that the one sample dial.

But trying to preserve a consumer out of the box experience has proven to be a hindrance to enabling innovation on a wide variety of platforms, particularly since the IVI HMI/UX landscape is so varied. Which is why I suggested moving to a developer oriented out of the box experience.

I deeply appreciate the folks like Nasa and Alison who have not only preserved through the multiple variations and change to MeeGo direction that have made following MeeGo IVI difficult, but have continued to contribute their time and knowledge to help the rest of the Community despite the difficulties. You guys do have impact, perhaps more so than the large but silent corporate teams that prefer to stay one step removed from active participation. Any lack of change in response to your comments is primarily a result of lack of developers to change the bits. This is certainly the case in the absence of an IVI build for the S10-3T and Pandaboard etc which we have really wanted to be able to do, but just have not had folks to make it happen.

(with my stupid mail client, I also appreciate being able to top-post :-) )

regards
Joel


-----Original Message-----
From: Nasa [mailto:nasa01-***@public.gmane.org]
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken; Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo

Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-ivi-***@public.gmane.org/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Nasa
2011-08-18 10:44:25 UTC
Permalink
----- Original Message -----
Post by Grigory Abuladze
Joel
- 1.1 was usable on wide range of HW
- 1.2 suddenly become usable on _VERY_ limited HW
And the answer to question "why?" is somewhat strange.
That's it - as simple as it looks.
And proposal was simple as well - please keep video driver more generic by default.
While I feel you...

I don't believe (someone will correct me if I am wrong) that there is
a *generic* driver solution. Given that the present UX is heavely based
on OpenGL, drivers have to support that. I don't know of a generic
driver that supports dri -- so implementing that solution would break the
UX. More specifically, the IVI target is emdedded devices which lack
the power to do use the full OpenGL specification and thus don't have
drivers for that. Instead they use OpenGL ES, which is a stripped
down version of the OpenGL spec. Now you need drivers to support that...

That said, this is another reason I like Joel's suggestion of seperating
the base IVI from the UX. This would allow interested parties to get
a fully bootable system and then add drivers/UX via zypper as thier
hardware requirements demand.

Nasa
Post by Grigory Abuladze
It would be great to have SandyBridge support, but it isn't main intention of discussion.
Regards
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
-----Original Message-----
Sent: 17 August 2011 18:23
To: Grigory Abuladze; Nasa
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
We did it because
1) Sandybridge was never identified as a target platform for the
meego-ivi-ia32 bootable image.
2) Nobody proposed, planned, resourced or contributed to creating a
bootable IVI image for Sandybridge.
3) the IVI platforms identified as target the meego-ivi-ia32 bootable
image all are supported by the "wrong" video driver (EMGD)
4) the QML based applications developed for MeeGo 1.2 ran dog slow
without graphics HW acceleration.
If anyone wants to propose, plan, resource or contribute to creating a
bootable IVI image for Sandybridge, I would be very happy to provide
my support for their efforts.
The key is to have a kickstart file to build the image, someone to
test the results and fix any issues reported.
regards
Joel
-----Original Message-----
Sent: Wednesday, August 17, 2011 10:11 AM
To: Clark, Joel; Nasa
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is
spend on optimization and/or what will be feasible in near future we
would like to use Sandy Bridge based tablet.
This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video
driver is embedded into release image!
Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by
embedding EMGD and how many pain and problems it brings?
I don't see problem in validating IVI only on specific HW and in same
time releasing generic image(s). It will left door slightly open for
everyone else who don't have this HW.
Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
-----Original Message-----
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
I apologize for the pain. :-). Actually I've felt quite a bit if it
too when it comes to trying to create a visually compelling experience
on embedded HW systems. We have found considerable HW and driver
specific dependencies. In general the idea was that each platform
champion would provide the adaptations required for their platform.
This would probably have produced multiple sample IVI boot images for
different platforms. So far this has not happened. We have been
struggling with the resources we have just to keep a couple of
platforms adaptations working sufficiently to enable the graphically
rich apps we borrowed from other devices, and voice recognition and
text to speech, and touchscreen all working. I think all of these are
essential to an IVI experience and should at least be able to be
demonstrated on platforms. Same explanation for the Rotary Dial
support, we just have not had enough folks contributing support for
different I/O devices to enable more that the one sample dial.
But trying to preserve a consumer out of the box experience has proven
to be a hindrance to enabling innovation on a wide variety of
platforms, particularly since the IVI HMI/UX landscape is so varied.
Which is why I suggested moving to a developer oriented out of the box
experience.
I deeply appreciate the folks like Nasa and Alison who have not only
preserved through the multiple variations and change to MeeGo
direction that have made following MeeGo IVI difficult, but have
continued to contribute their time and knowledge to help the rest of
the Community despite the difficulties. You guys do have impact,
perhaps more so than the large but silent corporate teams that prefer
to stay one step removed from active participation. Any lack of change
in response to your comments is primarily a result of lack of
developers to change the bits. This is certainly the case in the
absence of an IVI build for the S10-3T and Pandaboard etc which we
have really wanted to be able to do, but just have not had folks to
make it happen.
(with my stupid mail client, I also appreciate being able to top-post :-) )
regards
Joel
-----Original Message-----
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo
Grigory,
I feel your pain.
There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
(some who have followed this list will remember me bring that
thread up not so long a go).
Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of
that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.
The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).
I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel
has
been doing a better job of communicating all of this.
Nasa
BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...
----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Alison Chaiken
2011-08-18 14:33:35 UTC
Permalink
Nasa <nasa01-***@public.gmane.org> wrote:

allow interested parties to get a fully bootable system and then add
drivers/UX via zypper as thier hardware requirements demand.


Agree: what was annoying me this week was ruining a working set of drivers
by zypp-updating from IVI repositories. My whiny complaint is not about
lack of driver support for my hardware, which I *had* working, but the fact
that my /usr/lib was overwritten by drivers for a different Atom processor
when I updated.

Having said that, in an ideal world we'd have specific repos for IVI for
widely used and relevant HW platforms. This week I've been trying to
build a liveusb IVI image for ExoPC using mic2 and a mix of packages from
1.2.0.90, 1.1.99 and Arfoll's own special brew. The set of packages is
one that I have actually working on the hardware, where I have patched their
dependencies up with a script, but mic2 knows I'm cheating and fails with
messages like, "meego-ux-daemon requires libGLESv2.so.2" and "libqtgui4
requires libEGL.so.1". In fact the set of packages contains those
libraries, but not where mic2 has been told to look for them. I guess I
need to figure out how to rebuild the packages whose dependencies are
failing against the libraries I'm actually using or even better, so that no
assumption is made, as described in previous paragraph.
--
Alison Chaiken
(650) 279-5600 (cell)
http://www.exerciseforthereader.org/
"What you do for a living is not be creative, what you do is ship." -- Seth
Godin, vimeo.com/27067169
Nasa
2011-08-18 14:59:23 UTC
Permalink
----- Original Message -----
Post by Alison Chaiken
allow interested parties to get a fully bootable system and then add
drivers/UX via zypper as thier  hardware requirements demand.
Agree: what was annoying me this week was ruining a working set of
drivers by zypp-updating from IVI repositories.   My whiny complaint
is not about lack of driver support for my hardware, which I *had*
working, but the fact that my /usr/lib was overwritten by drivers for
a different Atom processor when I updated.
Having said that, in an ideal world we'd have specific repos for IVI
for widely used and relevant HW platforms.    This week I've been
trying to build a  liveusb IVI image for ExoPC using mic2 and a mix of
packages from 1.2.0.90, 1.1.99 and Arfoll's own special brew.    The
set of packages is one that I have actually working on the hardware,
where I have patched their dependencies up with a script, but mic2
knows I'm cheating and fails with messages like, "meego-ux-daemon
requires libGLESv2.so.2" and "libqtgui4 requires libEGL.so.1".  
I think (note the work *think*) I know what your issue is...

From my forum post here: http://forum.meego.com/showthread.php?t=3446
I noted one needed to install some Mesa packages that provided those
libraries (I'm not at home to be able to tell you for sure). The default
install doesn't install these as they come with the emgd-bin file. So
in your patterns (installed via package-groups) look for where that is called
and replace with the files from my forum post, see how far that gets you.

BTW: I am inferring this from reading a little about MIC2, I haven't had
a chance to try it myself. And don't forget to include in this change your
own video subsystem driver.


Nasa


In
Post by Alison Chaiken
fact the set of packages contains those libraries, but not where mic2
has been told to look for them.     I guess I need to figure out how
to rebuild the packages whose dependencies are failing against the
libraries I'm actually using or even better, so that no assumption is
made, as described in previous paragraph.
--
Alison Chaiken
(650) 279-5600  (cell)                                                
             http://www.exerciseforthereader.org/
"What you do for a living is not be creative, what you do is ship." --
Seth Godin, vimeo.com/27067169
Clark, Joel
2011-08-18 14:06:24 UTC
Permalink
Well the answer is pretty simple as well. Platforms get targeted and supported based on having someone step up and agree to support them. The meego-ivi-ia32 images target and support specific platforms based on contribution from folks that considered the graphics driver for those platforms to be a key element. Admittedly these folks are trying to sell hardware by demonstrating their HW specific value to large vendors.

My experience with MeeGo has been that we can't do IVI specific development without managing HW specific dependencies. There are simply too many different input and output components with differing driver dependencies and for embedded devices, little Plug and Play capability. I think these dependencies will be increasing in the future as more IVI specific features are added. A common generic default image would probably be limited to VESA graphics and keyboard/mouse input, a 2D PC image. This would remove the ability to include any MeeGo applications or applications with rich user interfaces, including scrollwheels, touchscreen, dual display, animation, automotive sensors etc.

I consider having a small minimal image that runs on cheap widely available development systems to be a very high priority. I'm hoping someone is willing to volunteer to maintain a build image in the MeeGo build system for such a platform. I'd be happy to see any number of images targeting folks favorite platforms maintained in the build system. The irony is that before folks can do exciting innovative applications, someone has to do the grunt work of getting the platform HW adaptations to work.

regards
Joel


-----Original Message-----
From: Grigory Abuladze [mailto:gabuladze-2by2kkC8J4GX/***@public.gmane.org]
Sent: Thursday, August 18, 2011 1:45 AM
To: Clark, Joel; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

Joel

It looks like we have misunderstanding here - what I am trying to say is:
- 1.1 was usable on wide range of HW
- 1.2 suddenly become usable on _VERY_ limited HW
And the answer to question "why?" is somewhat strange.
That's it - as simple as it looks.

And proposal was simple as well - please keep video driver more generic by default.

It would be great to have SandyBridge support, but it isn't main intention of discussion.

Regards
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.

-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 18:23
To: Grigory Abuladze; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

We did it because
1) Sandybridge was never identified as a target platform for the meego-ivi-ia32 bootable image.
2) Nobody proposed, planned, resourced or contributed to creating a bootable IVI image for Sandybridge.
3) the IVI platforms identified as target the meego-ivi-ia32 bootable image all are supported by the "wrong" video driver (EMGD)
4) the QML based applications developed for MeeGo 1.2 ran dog slow without graphics HW acceleration.

If anyone wants to propose, plan, resource or contribute to creating a bootable IVI image for Sandybridge, I would be very happy to provide my support for their efforts.
The key is to have a kickstart file to build the image, someone to test the results and fix any issues reported.

regards
Joel


-----Original Message-----
From: Grigory Abuladze [mailto:gabuladze-2by2kkC8J4GX/***@public.gmane.org]
Sent: Wednesday, August 17, 2011 10:11 AM
To: Clark, Joel; Nasa
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available


Let me try to explain it from this point:
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is spend on optimization and/or what will be feasible in near future we would like to use Sandy Bridge based tablet.

This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video driver is embedded into release image!

Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by embedding EMGD and how many pain and problems it brings?

I don't see problem in validating IVI only on specific HW and in same time releasing generic image(s). It will left door slightly open for everyone else who don't have this HW.

Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.


-----Original Message-----
From: Clark, Joel [mailto:joel.clark-***@public.gmane.org]
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available

I apologize for the pain. :-). Actually I've felt quite a bit if it too when it comes to trying to create a visually compelling experience on embedded HW systems. We have found considerable HW and driver specific dependencies. In general the idea was that each platform champion would provide the adaptations required for their platform. This would probably have produced multiple sample IVI boot images for different platforms. So far this has not happened. We have been struggling with the resources we have just to keep a couple of platforms adaptations working sufficiently to enable the graphically rich apps we borrowed from other devices, and voice recognition and text to speech, and touchscreen all working. I think all of these are essential to an IVI experience and should at least be able to be demonstrated on platforms. Same explanation for the Rotary Dial support, we just have not had enough folks contributing support for different I/O devices to enable more that the one sample dial.

But trying to preserve a consumer out of the box experience has proven to be a hindrance to enabling innovation on a wide variety of platforms, particularly since the IVI HMI/UX landscape is so varied. Which is why I suggested moving to a developer oriented out of the box experience.

I deeply appreciate the folks like Nasa and Alison who have not only preserved through the multiple variations and change to MeeGo direction that have made following MeeGo IVI difficult, but have continued to contribute their time and knowledge to help the rest of the Community despite the difficulties. You guys do have impact, perhaps more so than the large but silent corporate teams that prefer to stay one step removed from active participation. Any lack of change in response to your comments is primarily a result of lack of developers to change the bits. This is certainly the case in the absence of an IVI build for the S10-3T and Pandaboard etc which we have really wanted to be able to do, but just have not had folks to make it happen.

(with my stupid mail client, I also appreciate being able to top-post :-) )

regards
Joel


-----Original Message-----
From: Nasa [mailto:nasa01-***@public.gmane.org]
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Cc: meego-ivi-***@public.gmane.org; Alison Chaiken; Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo

Grigory,

I feel your pain.

There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
here: http://www.mail-archive.com/meego-ivi-***@public.gmane.org/msg00297.html
(some who have followed this list will remember me bring that
thread up not so long a go).

Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.

The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).


I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel has
been doing a better job of communicating all of this.

Nasa

BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...


----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Nasa
2011-08-18 14:43:44 UTC
Permalink
----- Original Message -----
Well the answer is pretty simple as well. Platforms get targeted and
supported based on having someone step up and agree to support them.
The meego-ivi-ia32 images target and support specific platforms based
on contribution from folks that considered the graphics driver for
those platforms to be a key element. Admittedly these folks are trying
to sell hardware by demonstrating their HW specific value to large
vendors.
My experience with MeeGo has been that we can't do IVI specific
development without managing HW specific dependencies. There are
simply too many different input and output components with differing
driver dependencies and for embedded devices, little Plug and Play
capability. I think these dependencies will be increasing in the
future as more IVI specific features are added. A common generic
default image would probably be limited to VESA graphics and
keyboard/mouse input, a 2D PC image. This would remove the ability to
include any MeeGo applications or applications with rich user
interfaces, including scrollwheels, touchscreen, dual display,
animation, automotive sensors etc.
Joel,

You missed my previous email on this :}

I maybe off a little here, but the only driver dependancy that
would cause a system to unuseable are video cards drives. Other
drivers, if started, would complain a lot - but you could still
use the box (then you could go in and stop those drivers from
running if you so desired). I bring this up, as you could include/support
a range of drivers for a range of other types of hardware and leave
out the video subsystem in a default IVI release.
I consider having a small minimal image that runs on cheap widely
available development systems to be a very high priority. I'm hoping
someone is willing to volunteer to maintain a build image in the MeeGo
build system for such a platform. I'd be happy to see any number of
images targeting folks favorite platforms maintained in the build
system. The irony is that before folks can do exciting innovative
applications, someone has to do the grunt work of getting the platform
HW adaptations to work.
I am interesting in getting this working on my own carpc... I finally
figured out how to get MIC2 installed on my home machine. Now to learn
how to configure .ks files and hopefully I can have some images working
for systems using video cards supported by the i915 driver (which is already
in the repo).

Nasa
regards
Joel
-----Original Message-----
Sent: Thursday, August 18, 2011 1:45 AM
To: Clark, Joel; Nasa
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
Joel
- 1.1 was usable on wide range of HW
- 1.2 suddenly become usable on _VERY_ limited HW
And the answer to question "why?" is somewhat strange.
That's it - as simple as it looks.
And proposal was simple as well - please keep video driver more generic by default.
It would be great to have SandyBridge support, but it isn't main intention of discussion.
Regards
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
-----Original Message-----
Sent: 17 August 2011 18:23
To: Grigory Abuladze; Nasa
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
We did it because
1) Sandybridge was never identified as a target platform for the
meego-ivi-ia32 bootable image.
2) Nobody proposed, planned, resourced or contributed to creating a
bootable IVI image for Sandybridge.
3) the IVI platforms identified as target the meego-ivi-ia32 bootable
image all are supported by the "wrong" video driver (EMGD)
4) the QML based applications developed for MeeGo 1.2 ran dog slow
without graphics HW acceleration.
If anyone wants to propose, plan, resource or contribute to creating a
bootable IVI image for Sandybridge, I would be very happy to provide
my support for their efforts.
The key is to have a kickstart file to build the image, someone to
test the results and fix any issues reported.
regards
Joel
-----Original Message-----
Sent: Wednesday, August 17, 2011 10:11 AM
To: Clark, Joel; Nasa
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
- Head Unit Manufacturers (xxx xxx) are not yet convinced in IVI
- We are trying to demonstrate them what it will be and how it will look
- To do it we take ExoPC(WeTab) and provide them demo (sort of)
- Also to simulate what would be possible in case of more time is
spend on optimization and/or what will be feasible in near future we
would like to use Sandy Bridge based tablet.
This simple way did work up to 1.2 release of ivi...
And the only things why it is now broken - it's because "wrong" video
driver is embedded into release image!
Why? Why did you do it?
Typing one line in console after installation - is what is "saved" by
embedding EMGD and how many pain and problems it brings?
I don't see problem in validating IVI only on specific HW and in same
time releasing generic image(s). It will left door slightly open for
everyone else who don't have this HW.
Thansk
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
-----Original Message-----
Sent: 17 August 2011 15:55
To: Nasa; Grigory Abuladze
Subject: RE: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be available
I apologize for the pain. :-). Actually I've felt quite a bit if it
too when it comes to trying to create a visually compelling experience
on embedded HW systems. We have found considerable HW and driver
specific dependencies. In general the idea was that each platform
champion would provide the adaptations required for their platform.
This would probably have produced multiple sample IVI boot images for
different platforms. So far this has not happened. We have been
struggling with the resources we have just to keep a couple of
platforms adaptations working sufficiently to enable the graphically
rich apps we borrowed from other devices, and voice recognition and
text to speech, and touchscreen all working. I think all of these are
essential to an IVI experience and should at least be able to be
demonstrated on platforms. Same explanation for the Rotary Dial
support, we just have not had enough folks contributing support for
different I/O devices to enable more that the one sample dial.
But trying to preserve a consumer out of the box experience has proven
to be a hindrance to enabling innovation on a wide variety of
platforms, particularly since the IVI HMI/UX landscape is so varied.
Which is why I suggested moving to a developer oriented out of the box
experience.
I deeply appreciate the folks like Nasa and Alison who have not only
preserved through the multiple variations and change to MeeGo
direction that have made following MeeGo IVI difficult, but have
continued to contribute their time and knowledge to help the rest of
the Community despite the difficulties. You guys do have impact,
perhaps more so than the large but silent corporate teams that prefer
to stay one step removed from active participation. Any lack of change
in response to your comments is primarily a result of lack of
developers to change the bits. This is certainly the case in the
absence of an IVI build for the S10-3T and Pandaboard etc which we
have really wanted to be able to do, but just have not had folks to
make it happen.
(with my stupid mail client, I also appreciate being able to top-post :-) )
regards
Joel
-----Original Message-----
Sent: Wednesday, August 17, 2011 4:28 AM
To: Grigory Abuladze
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefo
Grigory,
I feel your pain.
There are some disconnects between what is being developed and
what is being expected. A good place to start reading up on this
is a thread I started back in March "Supported Hardware", see
(some who have followed this list will remember me bring that
thread up not so long a go).
Futher investigation reveals that the IVI team has a very specific and
limited view of what constitutes an IVI system, which is aimed more
at commerical entities than the casual potential user. This statement
isn't meant as a knock on anyone, it's more of an observation. Most of
that
hardware isn't readyly available for hobbiest, which is a small set as it is,
limiting thier involvement.
The problem is that mixed signals were sent out about this at the begining of
the Meego project. So the statement from Bug 17351
(directly related to this thread, read it), "Since we are discussing
hardware that this was never intended for, I am marking Resolved invalid."
does kinda of get under my skin (1st and only thing I will say on that
comment).
I think the idea of seperating the IVI from the UX is/would help clarify
this for anyone looking to jump in (where the none UX would be the *official*
item releases). As the direction this effort has gotten clearer, Joel
has
been doing a better job of communicating all of this.
Nasa
BTW: The integration of the new drivers isn't the first instance of this
disconnect -- the touchscreen issue (bug 12777 I believe) is a result of
the same thinking. As is the inclusions of a specific rotary dial...
----- Original Message -----
Post by Alison Chaiken
Alison
I have to say that 1.2 release has already dropped all of us who use
ExoPC and other non new EMGD compatible things.
As I can see 1.1 had sw " rasterizer" as default and thus
- anyone was able to install it on various HW
- and add preferable driver later
I have no idea why this way was broken in 1.2 release, but it drops
out a lot of people. They tried "new" release and it just doesn't work
for them - so they simple walk away to do something for Android or
iOS.
I remember list member NASA has provided some "workaround" and it was
quite popular topic ). But not all of us can/want to spend time to try
dancing around to "fix" drivers issue.
I would like to extend your proposal to keep it general and end user
will decide which HW specific driver to install. And yes - proper
naming will be definitely helpful here.
Grigory.
--
Grigory Abuladze
Application Solutions (Electronics and Vision) Limited.
Tel: +44 (0) 1273 409147
-----Original Message-----
Sent: 16 August 2011 16:55
To: Clark, Joel
Subject: Re: [MeeGo-ivi] do you know when MeeGo 1.2.0.2 will be availablefor ivi
Joel, what I'm really thinking about is the many Lenovo Ideapads and
ExoPCs that Intel has generously given to developers: thank you! I
would hope that as the repos are refreshed, that AppUp participants
whose expertise is in QML rather than Linux graphics are not going to
get their machines borked if up they "zypper update" from trunk.
I suggest that someone (Bob Spencer, Amar Kona . . . ) make it
clear what the best update path for developers using this older
hardware is. Otherwise when 1.3 drops, a lot of folks using older
HW may be unhappy. It's already a bit confusing about which graphics
RPMs to use with which Atom processors. Perhaps the EMGD RPMs could
have "sandybridge" or whatever in the package name.
--
Alison Chaiken
(650) 279-5600  (cell)
             http://www.exerciseforthereader.org/
INplanatory documentation: that which sucks comprehension right out of
the reader's mind and destroys it.
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
_______________________________________________
MeeGo-ivi mailing list
http://lists.meego.com/listinfo/meego-ivi
Loading...