Messages

Note that manufacturers say they require the closed source firmware to comply with regulations.

That may be an excuse in part and part of the job is also done by the driver but it is something to take into account.

I wonder how much out of spec the cards could go with custom firmware.

How will you get certifications for such opensource hardware?

If it's really too much trouble, we might indeed better opt for a soldered in GPU. Then of course, we'll have to decide on a GPU. I don't think we'd have the resources to offer multiple configurations.

How about making a mainboard with a MXM slot and a default MXM card?

You can also test other MXM cards available to verify that the slot works as intended.

Then you get working device with a card of your choice and people are free to swap in a different card.

Or you can start shipping a different model with a different card later.

The engine might ported be but have you seen anybody building the browser even for ppc4le?

So do you want we add to the poll OpenPPCbooK ?

With the most popular name having 13 votes and the deadline for choosing the name being two days ago I guess there is no much point.

Well, as you may know, we do not have enough options for names.

More importantly we have lack of voters. When there are not enough people interested I guess it makes sense for the project leaders to decide .

Here are a few more better and more goodly names.
Power Bisquit
Gnew Sandwich Laptop

Reminds me of really old laptop which had an ISA slot called toaster. I have never seen it but I can imagine it was pretty much built around that slot.

Cheese and Fries in Disguise
It's-my-laptop-so-you-better-give-it-to-me-because-just-because-okay
What have these nerds come up with now?
Oh, yeah? But does it run linux?

There is also some sort of IBM JDK.

It is probably ppc64le only in recent versions, though.

I don't like flash either.

It was promising platform in the past and provided features that the web did not provide otherwise.

The web has caught up since so with very recent versions of 'major' web browsers you can get pretty much anything flash provides in html5 only.

Unfortunately, many websites have not caught up and still use flash for some major features meaning they will break without flash.

In particular youtube still requires flash to play back old low-res videos AFAICT.

I think the platform was killed by Macromedia not allowing alternative renderers. They tried to exert tight control over the platform which in the end bit them back. At certain point wintel+mac was no longer 99.999% of everything and an alternative renderer that was required for the platform to survive at least as a niche alternative was not there. Today you would not consider flash for new applications unless you use a framework that has it only as one of possible targets because you want to reach users. And most users are 'mobile' these days which means mostly flash-free.

Practically the Adobe blob or the Google build of the same is the only way to render flash content to date. So if you want flash consider x86 emulation.

The maintainer and developer of TenFourFox is apparently a fantastic person, I guess contacting him for clarification wouldn't be a bad thing.

But as far as I know, he does all of his TenFourFox work in XCode, and he leverages Mac OS-only APIs like Core Text and Core Image. I'm not sure how portable that code is.

His main working OS is to this day OS X 10.4.

You can just build the corresponding version of Darwin which is basically OS X without some proprietary apps and without Apple support.

Given it is conserved in the state it was back then providing optimization for powerpc/altivec it might be nice platform to try out.

How much effort it will be to get it going and update hardware drivers is difficult to tell. Surely the newer versions do support more current hardware but they focus on Intel. Meaning some obscure pieces required to run on the power notebook will need to be ported from elsewhere and there might be endian problems with drivers already present because nobody ever built them for bigendian.

b) PPC64 Big Endian Road - Kernel 4.x days - Its basically down to Fedora 20 Server
and Ubuntu 14 MATE. Here we have newer MATE/Cinnamon, and really fresh FF, Libre 5.x etc.
Also we need 64-bit for 4GB+ RAM

Do you have a recent FF that plays Youtube videos on ppc64 BE?

Has a name been decided already?

It seems the poll is still open but I do not see any particularly awesome name.

BTW powertop is a name of certain piece of software .. somewhat akin to iotop and nethogs.

PoPCorn is kind of inventive at least and features PPC in it so it somewhat relates to the project and its deliverable ;-)

m-ower does not display properly for me. I see ? in place of the first character. Is it Π-ower or π-ower or ..? Π indeed looks a lot like M if you limit yourself to Latin characters.

progressbook and openbook already exist so much like powerbook they are out of question. Although progressbook is used in relation to school so might be kind of funny to overload that term.

Names containing Libre and some English word sound like you cannot speak one language properly and have to mix two. Isn't libre French?

I don't particularly like the names with cube. I expect notebook to be flat ;-) And IIRC powercube was (unofficial?) name of a Mac computer model.

I totally support MeOW but I do not see any connection to the project.

I think something like OpenPPCbooK would be ok. It does not embed any well known existing trademark. "Open" reflects the choices of the project to consider NDA-free and accessible hardware and software where practical but including components with more restrictions to achieve reasonable performance and features where necessary. So not truly "free hardware".

What are the relevant events where you can get one?

Between power-hungry server chips and crippled Freescale chips the choice is difficult. IBM went little-endian which makes porting software that runs on Intel waaay easier and distros started switching to that but the Freescale chips don't support AltiVec in LE ..

Nothing stops you from making a RISC-Vook.

I think choosing PowerPC is kind of conservative - notebooks have been built around the platform in the past so it should work reasonably well .... ehm it used to with the IBM chips waaay back when they were reasonably powerful to give at least netbook class performance. I am not really that convinced about those Freescale chips. And given the thing is planned to have discrete graphics you can pretty much use any random core you can find that comes in a package with PCIe and some other reasonable peripherials .... like RISC-V.

PowerPC has the advantage of relatively mature port. It may be kind of dying out but you still have a lot of software that is already ported .. or was and can be refreshed easily.

Hello,

I find build services are crucial for building a distribution. There is only so much the developers can build on their PCs.

The most versatile publicly available build service I know of is OpenSUSE OBS https://build.opensuse.org/ It can obviously build for OpenSUSE but it has also options to build for Debian, Arch and other distributions.

OpenSUSE currently does not release for PPC but here is project for resurrecting PPC support https://en.opensuse.org/Portal:PowerPC
The PPC architecture is included in automated testing https://openqa.opensuse.org/group_overview/4

The only alternative build service I know is launchpad https://launchpad.net/ It builds only for Ubuntu.