Rimas Kudelis | 1 Dec 2011 10:23
Picon
Gravatar

FreeType patents are expired

Hi,

I'm wondering why subpixel hinting is not enabled by default in Haiku. 
If i'm reading it correctly[1], the algorithm used by FreeType for that 
purpose is not patented, and all other FreeType-related patents are 
expired, so why would Haiku be more afraid than Debian or any other 
Linux distro?

Rimas

[1] http://www.freetype.org/patents.html

Jerry Babione | 1 Dec 2011 13:42
Picon

Re: FreeType patents are expired

2011/12/1 Rimas Kudelis <rq-CPSw0JvRt6Q@public.gmane.org>
Hi,

I'm wondering why subpixel hinting is not enabled by default in Haiku. If i'm reading it correctly[1], the algorithm used by FreeType for that purpose is not patented, and all other FreeType-related patents are expired, so why would Haiku be more afraid than Debian or any other Linux distro?

Rimas

[1] http://www.freetype.org/patents.html


Rimas,

I'm sure that there are reasons we can't even think of to use the Free-Type algorithms. I can think of several not too.  For me, as a developer, I would prefer for the techniques, algorithms, and everything else be fresh and new.  I realize we are a re-implementation at Haiku.  There is also much that is new.  The point is, when others come at Haiku with any infringement type questions; I would wish that there be NO validity in the claims.  I know what a tall order that is.  So do you.

--
Jerry Babione
Founder-Just Plain Folks Org. Inc.
Stephan Aßmus | 1 Dec 2011 14:36
Picon
Picon
Gravatar

Re: FreeType patents are expired

Hi,

On 01.12.2011 10:23, Rimas Kudelis wrote:
> I'm wondering why subpixel hinting is not enabled by default in Haiku.
> If i'm reading it correctly[1], the algorithm used by FreeType for that
> purpose is not patented, and all other FreeType-related patents are
> expired, so why would Haiku be more afraid than Debian or any other
> Linux distro?
>
> [1] http://www.freetype.org/patents.html

I am afraid you misunderstood what patents the above refers to. The 
bytecode hinting patents have expired, and AFAIA aware, we subsequently 
enabled the bytecode hinter in our releases. However, the subpixel 
filtering is an entirely different matter. Different patents.

Best regards,
-Stephan

Rimas Kudelis | 1 Dec 2011 14:50
Picon
Gravatar

Re: FreeType patents are expired

2011.12.01 15:36, Stephan Aßmus rašė:
> Hi,
>
> On 01.12.2011 10:23, Rimas Kudelis wrote:
>> I'm wondering why subpixel hinting is not enabled by default in Haiku.
>> If i'm reading it correctly[1], the algorithm used by FreeType for that
>> purpose is not patented, and all other FreeType-related patents are
>> expired, so why would Haiku be more afraid than Debian or any other
>> Linux distro?
>>
>> [1] http://www.freetype.org/patents.html
>
> I am afraid you misunderstood what patents the above refers to. The 
> bytecode hinting patents have expired, and AFAIA aware, we 
> subsequently enabled the bytecode hinter in our releases. However, the 
> subpixel filtering is an entirely different matter. Different patents.

Yeah, thanks. I did some more reading.[1] Apparently, ClearType patents 
won't expire for another 8 years. Although it seems unclear whether or 
not Freetype infringes them and whether or not it's possible to 
circumvent those patents by using e.g. SubLCD. What a shame!

Regards,
Rimas

[1] http://david.freetype.org/cleartype-patents.html

Rimas Kudelis | 1 Dec 2011 14:50
Picon
Gravatar

Re: FreeType patents are expired

2011.12.01 14:42, Jerry Babione rašė:
> 2011/12/1 Rimas Kudelis <rq@... <mailto:rq@...>>
>
>     Hi,
>
>     I'm wondering why subpixel hinting is not enabled by default in
>     Haiku. If i'm reading it correctly[1], the algorithm used by
>     FreeType for that purpose is not patented, and all other
>     FreeType-related patents are expired, so why would Haiku be more
>     afraid than Debian or any other Linux distro?
>
>     Rimas
>
>     [1] http://www.freetype.org/patents.html
>
>
> Rimas,
>
> I'm sure that there are reasons we can't even think of to use the 
> Free-Type algorithms. I can think of several not too. For me, as a 
> developer, I would prefer for the techniques, algorithms, and 
> everything else be fresh and new. I realize we are a re-implementation 
> at Haiku. There is also much that is new. The point is, when others 
> come at Haiku with any infringement type questions; I would wish that 
> there be NO validity in the claims. I know what a tall order that is. 
> So do you.

Hi Jerry,
I'm not sure I'm following you, but AFAIK, validity of infringement 
claims is justified by how you did something, not by whether or not you 
did that completely on your own. All in all, I don't think your email 
answered my question. Stephan's did however.

Regards,
Rimas

Kevin Ar18 | 1 Dec 2011 20:02
Picon
Favicon

Re: Status of LLVM+Clang & Go Language?

Thanks for the replies.
> >> Is it possible to download it (pre-built) somewhere?
>
> I can build and put up the 3.0 release somewhere when its released,
> which will be today/tomorrow.
No need to, unless you think others want to use it.  I saw the parts about building Haiku with clang (and the problems)....  I was mostly asking from the perspective of using llvm+clang and go compiler in Haiku itself.  Anyways, it's no big deal, as I probably won't end up using it anyways.  Still, thanks for the info.
Jerry Babione | 2 Dec 2011 01:17
Picon

Re: FreeType patents are expired



On Thu, Dec 1, 2011 at 7:50 AM, Rimas Kudelis <rq-CPSw0JvRt6Q@public.gmane.org> wrote:
2011.12.01 14:42, Jerry Babione rašė:
2011/12/1 Rimas Kudelis <rq-CPSw0JvRt6Q@public.gmane.org <mailto:rq <at> akl.lt>>

   Hi,


   I'm wondering why subpixel hinting is not enabled by default in
   Haiku. If i'm reading it correctly[1], the algorithm used by
   FreeType for that purpose is not patented, and all other
   FreeType-related patents are expired, so why would Haiku be more
   afraid than Debian or any other Linux distro?

   Rimas

   [1] http://www.freetype.org/patents.html


Rimas,

I'm sure that there are reasons we can't even think of to use the Free-Type algorithms. I can think of several not too. For me, as a developer, I would prefer for the techniques, algorithms, and everything else be fresh and new. I realize we are a re-implementation at Haiku. There is also much that is new. The point is, when others come at Haiku with any infringement type questions; I would wish that there be NO validity in the claims. I know what a tall order that is. So do you.

Hi Jerry,
I'm not sure I'm following you, but AFAIK, validity of infringement claims is justified by how you did something, not by whether or not you did that completely on your own. All in all, I don't think your email answered my question. Stephan's did however.

Regards,
Rimas


Rimas,

I understand.  I've fought two such law suites in the 30+ years I've been writing software.  I learned that even independent development is not always safe.  What I know today, is not too study the methods that others used to create their software. Imitation can cause a trip to court. Sometimes re-inventing the wheel saves a lot of money,

-- 
Jerry Babione
Founder-Just Plain Folks Org. Inc.
helix84 | 3 Dec 2011 02:30
Picon
Favicon
Gravatar

running Haiku on real hardware

Hello Haiku developers,

today for the first time I left the comfort of virtualization and
found the time and hardware to try Haiku natively. I'd like to share
with you a few observations and experiences.

First, I dd'd (wow, that's a nice verb, first time I used it) the raw
image in an unformatted partition and added a chainloader entry to
GRUB. After reboot I got surprised by "superblock not found" from
GRUB. Luckily, search turned up this wonderful informational article:
https://www.haiku-os.org/blog/mmlr/2009-02-08/makebootable_what_and_why_and_how_do_it_manually
So after running makebootable from Linux, I was finally able to boot
Haiku. Later I was surprised again when I tried to boot the same raw
partition from VirtualBox via a vmdk disk descriptor like I used to.
But I quickly realized that the offset written by makebootable was
from the start of the physical disk, while the vmdk contains only this
one partition so the offset is from start of the partition.
Understadable, though a bit inconvenient if I want to switch between
bare metal and virtualized environment where all drivers work fine.

When I first started Haiku on bare metal (a Fujitsu-Siemens Lifebook
E8310), the keyboard didn't work. I tried to look for some
compatibility option in BIOS concerning keyboard, but there is none.
Then I tried recovery mode and keyboard worked. I was trying to
pinpoint exactly which recovery option helps, but it turned out none
was necessary! It's sufficient to press Shift to bring up the Haiku
boot menu and then proceed booting normally. Keyboard will then work
in Haiku itself. If someone can explain why this is so, maybe he can
also fix it.

A second problem I noticed is that when I restart Haiku, which invokes
warm boot, this notebook hangs at the BIOS screen after POST, shortly
before handing off control to MBR. When I press Ctrl-Alt-Del at this
stage, the same thing happens. I have to power it off, then it boots
normally. The system also boots normally after a Power off from Haiku
(as opposed to Restart).

I can provide any pci listings or whatever if you tell me how.

Next I'm going to investigate if Haiku supports WPA2/AES with hidden SSID.

Regards,
~~helix84

François Revol | 3 Dec 2011 13:37
Picon
Favicon
Gravatar

Re: running Haiku on real hardware

Hi,

Le 03/12/2011 02:30, helix84 a écrit :
> First, I dd'd (wow, that's a nice verb, first time I used it) the raw
> image in an unformatted partition and added a chainloader entry to
> GRUB. After reboot I got surprised by "superblock not found" from
> GRUB. Luckily, search turned up this wonderful informational article:
> https://www.haiku-os.org/blog/mmlr/2009-02-08/makebootable_what_and_why_and_how_do_it_manually
> So after running makebootable from Linux, I was finally able to boot
> Haiku. Later I was surprised again when I tried to boot the same raw
> partition from VirtualBox via a vmdk disk descriptor like I used to.
> But I quickly realized that the offset written by makebootable was
> from the start of the physical disk, while the vmdk contains only this
> one partition so the offset is from start of the partition.
> Understadable, though a bit inconvenient if I want to switch between
> bare metal and virtualized environment where all drivers work fine.

IIRC you can set the vmdk to have the partition at the same offset in a
larger disk so it fits.

> When I first started Haiku on bare metal (a Fujitsu-Siemens Lifebook
> E8310), the keyboard didn't work. I tried to look for some
> compatibility option in BIOS concerning keyboard, but there is none.
> Then I tried recovery mode and keyboard worked. I was trying to
> pinpoint exactly which recovery option helps, but it turned out none
> was necessary! It's sufficient to press Shift to bring up the Haiku
> boot menu and then proceed booting normally. Keyboard will then work
> in Haiku itself. If someone can explain why this is so, maybe he can
> also fix it.

You'll want to check for existing tickets on http://dev.haiku-os.org/
or create one...

> 
> A second problem I noticed is that when I restart Haiku, which invokes
> warm boot, this notebook hangs at the BIOS screen after POST, shortly
> before handing off control to MBR. When I press Ctrl-Alt-Del at this
> stage, the same thing happens. I have to power it off, then it boots
> normally. The system also boots normally after a Power off from Haiku
> (as opposed to Restart).

Again, check for existing ticket.

> Next I'm going to investigate if Haiku supports WPA2/AES with hidden SSID.

WPA2 support is possible if you install wpa_supplicant but it's not
finished yet.

François.

Jerry Babione | 3 Dec 2011 13:51
Picon

Re: running Haiku on real hardware



On Sat, Dec 3, 2011 at 6:37 AM, François Revol <revol <at> free.fr> wrote:
Hi,

Le 03/12/2011 02:30, helix84 a écrit :
> First, I dd'd (wow, that's a nice verb, first time I used it) the raw
> image in an unformatted partition and added a chainloader entry to
> GRUB. After reboot I got surprised by "superblock not found" from
> GRUB. Luckily, search turned up this wonderful informational article:
> https://www.haiku-os.org/blog/mmlr/2009-02-08/makebootable_what_and_why_and_how_do_it_manually
> So after running makebootable from Linux, I was finally able to boot
> Haiku. Later I was surprised again when I tried to boot the same raw
> partition from VirtualBox via a vmdk disk descriptor like I used to.
> But I quickly realized that the offset written by makebootable was
> from the start of the physical disk, while the vmdk contains only this
> one partition so the offset is from start of the partition.
> Understadable, though a bit inconvenient if I want to switch between
> bare metal and virtualized environment where all drivers work fine.

IIRC you can set the vmdk to have the partition at the same offset in a
larger disk so it fits.

> When I first started Haiku on bare metal (a Fujitsu-Siemens Lifebook
> E8310), the keyboard didn't work. I tried to look for some
> compatibility option in BIOS concerning keyboard, but there is none.
> Then I tried recovery mode and keyboard worked. I was trying to
> pinpoint exactly which recovery option helps, but it turned out none
> was necessary! It's sufficient to press Shift to bring up the Haiku
> boot menu and then proceed booting normally. Keyboard will then work
> in Haiku itself. If someone can explain why this is so, maybe he can
> also fix it.

You'll want to check for existing tickets on http://dev.haiku-os.org/
or create one...

>
> A second problem I noticed is that when I restart Haiku, which invokes
> warm boot, this notebook hangs at the BIOS screen after POST, shortly
> before handing off control to MBR. When I press Ctrl-Alt-Del at this
> stage, the same thing happens. I have to power it off, then it boots
> normally. The system also boots normally after a Power off from Haiku
> (as opposed to Restart).

Again, check for existing ticket.

> Next I'm going to investigate if Haiku supports WPA2/AES with hidden SSID.

WPA2 support is possible if you install wpa_supplicant but it's not
finished yet.

François.


The boot issue may well be on a ticket.  I have the same issues when I attempt to run Haiku with the "Plug and Play" OS option turned on in the BIOS.

--
Jerry Babione
Founder-Just Plain Folks Org. Inc.

Gmane