Re: Purism Librem Phone
 

Off topic but I feel like I need to point out the common miscomprehension about that 4GB limit.

There is indeed a 4GB limit on 32-bit systems but it is per process. There is nothing stopping you designing a 32-bit machine with more that 4GB of RAM and a clever memory management system that can handle it. Each process would still be limited to 4GB but not the whole system.

Not that any of that matters (yet) to mobile systems. It's not like we have phones with more than 4GB sold in six-packs.

Re: Purism Librem Phone
 

this.....
What's the point of a smartpone that is as useful as some candy bar crap. If you can't do things in browser than you can't do them at all in most cases.
My jolla screen was busted so I'm using AOSP ATM and while I do miss sailfish as OS and gestures, but from pure app POV it is way better even just for the basic stuff. Bunch of free navigation apps that just work, file managers with simple but full network acces, working telegram, VIA browser that is 721478 light years ahed of the one in sailfish....

Re: Purism Librem Phone
 

I do like my sailfish phone, and librem phone seems interesting but I am worried that it might be a linux desktop distro placed on a phone compared to a proper thought out OS made to run on a mobile device. BUT time will tell and we shall see what happens.
Some people do want their favourite distro on a phone

Re: Purism Librem Phone
 

Well, there is not really anything preventing "regular" distros from carrying mobile optimized apps. They already usually support at least two pretty disparate use cases - big headless server/server clusters and single user workstations/desktop computers. So supporting another use case/form factor should not be *that* hard.

Even Sailfish OS isn't (thankfully!) that different from a "normal" Linux distro under the hood - just some additional mobile middleware and little optimizations/hacks here and there.

So I think running normal Linux distros on mobile devices should be genrally fine and could help Sailfish OS by more people working on mobile Linux distros (not embedded hacks).

But that doesn't mean it won't take time to actually get to regular distro (say Fedora/Debian/Arch/etc.) work on a mobile device in a usable form - you need to package and stabilize all the middleware, create touch interfaces for applications, etc.

Doable ? Certainly!
Fast ? Likely not.

Re: Purism Librem Phone
 

Reaching the $1.5 million milestone weeks ahead of schedule enables Purism to accelerate the production of the physical product. The company plans to move into hardware production as soon as possible to assemble a developer kit as well as initiate building the base software platform, which will be publicly available and open to the developer community.


https://puri.sm/posts/librem-5-phone...0percent-mark/

Also found this "new" video:

https://www.youtube.com/watch?v=emb3ImTFyIY

Re: Purism Librem Phone
 

AFAIK PureOS itself is GTK :/ So I guess n900/neo fans would like it.

But personally think they should use Qt/QML instead of GTK.

Why? It simple:

Porting SFOS apps would be easier if they used Qt/QML. We also have Plasma mobile and Nemo mobile which also uses Qt/QML.

So I don't get why on earth use GTK nowdays on a mobile OS. It is easier write apps in QML/C++ than using C and some bindings(gtkmm/pygtk) since no one want to use lowlevel C this days when write apps. Also coders are lazy and want an IDE and QtCreator already in place. GTK has no "pro" IDE AFAIK. Many still prefer Vim(me also) but still if you want top make mobile apps fast you need an IDE with possibility to debug and so on.

I personally prefer QtCreator over bloatead Android studio. Jolla/Qt team has done a great job with QtCreator + mer tooolchain. So why invent the wheel and use GTK and some unknown IDE? I really don't get it.

Well let's see what happens...

Edit: it seems from the video above that they will provide some kind of framework for both GTK and Qt. Nice :)

Re: Purism Librem Phone
 

There was a post about a month ago on GNOME's desktop-devel-list looking for people interested in writing the mobile UI and asking if GNOME as a project was interested in participating. So, my guess is that they currently have nothing, not even the stack chosen. And that makes their funding campaign look really odd, because it implies they have a mobile OS. I get the impression what they actually have is a distro running stock GNOME 3 on ARM.

https://mail.gnome.org/archives/desk.../msg00011.html

Re: Purism Librem Phone
 

Check the video above it may give some more info. It seems they want to work near both Gnome and KDE community and make some "framework" for both...

Re: Purism Librem Phone
 

Indeed, but there are not only architecture changes :
only the higher range of iMx6 features the 4 cores, most only have one or two, whereas this mid range iMx8M have 4 cores. I can't find where I read that the iMx6 they where targeting had only 2 cores, I may have dreamed that one...
imx6 are limited to 1.2GHz whereas imx8 go to 1.5GHz.
And the A53 should be more power efficient.

But yeah, as said before, even the iMx8 is weaker than the octo-cores and the like found in Android flagships.

By the way, the higher range of iMx8 exist with 4xA53+2xA72, but is not targeted here. It probably costs a lot more.

Re: Purism Librem Phone
 

They are considering to use the i.MX 8M Quad instead of the i.MX 6 Quad, i.e. this is just going from four low-end 32bit cores to four-low end 64bit cores, see Wikipedia.

Furthermore the Vivante GC7000 GPU used in all i.MX 8 versions currently has no FOSS driver support at all; I believe that is withholding them to do this change, as that work can hardly be achieved within the given timeframe of 15 to 24 months.
BTW, in contrast to that the FOSS driver and MESA support for the recent generation Qualcomm Adreno 5xx GPUs is mostly done, and done for Adreno 3xx and 4xx (what cannot be said for the older Vivante GPU generations either: still a lot of work to do for them).

The perceived performance for interactive uses comes primarily from single-thread performance, so that core-craze is hardly worth much beyond two cores and largely irrelevant beyond four cores.

It is the effective average IPC (instructions per clock) and clock frequency of a core what matters (multiplied with each other): Thus modern high-end ARM cores have more than a threefold performance advantage compared to low-end ones, for each core.
Multiplying that with the number of cores effectively being used for your workload gives a rough estimate of the perceived performance difference!