N9, N950: Two Nokia MeeGo devices. Thoughts before going to bed.

The N9 was announced in Asia while I was sleeping here in Europe. Now I am about to go to bed again, but as I can´t do any serious blogging (videos, images) being online over an EDGE network which feels more like GPRS in terms of speed right now, I want to share some thoughts on the N9 and the N950.

The N950, which I didn´t really cover here yet on LINMOB, is the device that we saw on leaked images aeons ago (~ 1 year), a device much looking like the Nokia E7: An 4” HWVGA aluminum slab with a slide out, full 4 row QWERTY keyboard. Don´t get overly excited about this device if you aren´t a developer, as it will be tough to get one then. Read more on this over at thehandheldblog.com, and if you are interested in the actual differences between these two new Nokia Handsets, let me point to this post at meegonews.com

For the non Qt / Linux developing rest of us, there will be the keyboard less, polycarbonate N9, which looks different than the existing Nokia phones, it doesn´t resemble the N8 e.g. – I like it … I think I should stop repeating myself, so nothing on the specs here which I haven´t posted yet, they are decent, not breathtaking.

Pricing and availability. Not much info on that yet. There are, as I noticed earlier today, indicators that the N9 will only sell in a few countries. (->TheHandheldBlog) Considering the general excitement about this new Nokia product this would be a very sad thing, but Nokia has (from my view as a mobile linux lover) made tons of sad decisions since Steve Elop came aboard.

I already linked you to thehandheldblog.com twice, and I will do it a third time in this post, simply because I totally agree to their comment “With The N9, Nokia Shows The World Its Still Got It”.

Last but not least you should watch this video of the presentation (by netbooknews) – I can´t right now (EDGE):

FSO: Qt based Aurora for Palm Pre, Pre Plus, Pre²)

The world of mobile linux is like that today: There´s a huge buzz about Android, less, but still some about HP webOS and MeeGo and very little about LiMo. But what about all those projects that started with Openmoko back in the days? Well, some of them are still alive and as that´s why I share this announcement by mickey and morphis, which was posted to the SHR mailing list a two days ago:

Dear FOSS-Telephony lovers,

today we want to announce something that has been brewing in our minds
for quite a while and will change the way we develop the
freesmartphone.org middleware.

In the past, FSO has been too much developed without considering how the
features will actually be used by the API consumers.
Apart from the great work our friends from SHR did, there has only been
a handful of special purpose FSO clients, such as
the Emacs client, Zhone, and Zhone2. Zhone (and its successor Zhone2) is
currently an oversimplified approach based on a
non-maintainable Edje file. We have therefore decided to develop a new
testing/demonstrator for FSO named Aurora that
is supposed to be the driving force for further development.

AURORA

The aim of Aurora is to replace zhone and zhone2 as development UIs for
FSO. From the viewpoint of a middleware architect,
it’s essential to have clients available that use the various features
of the FSO services. On the other hand though, this
time we want to create something that is also suitable for day to day
use. Aurora is supposed to be something we call
a “featurephone client” ? featurephones being those things we used for
telephony before smartphones were invented.

Aurora being a featurephone client does not necessarily mean it will
never get the “smartphone features” Android or iOS
are popular for, it rather describes our approach as being
as-simple-as-possible. So for now you will not be able to
install additional apps or features. Everything (you need) is part of
the Aurora client.

DEVELOPMENT PROCESS

At the top of every application stack is the user. Pleasing him or her
is the topmost priority. Technology should not stand in the
 way, but rather support the user. Hence, Aurora releases will be done
as user milestones. For every user milestone, we will
pick a number of user stories to be implemented. We will then split a
user story into tasks and distribute among the contributors.

SUPPORTED DEVICES

We decided to only support the Palm Pre devices (Pre/Pre Plus/Pre 2) for
the first to-be-released version of Aurora. More
supported devices will join after the 0.1 release. This decision has
been forced by the fact that we are only very few people
working both on FSO and Aurora (and also on OpenEmbedded). Later on, we
expect to see the OpenEZX family of devices,
the Openmoko devices, the Nokia N900, and possibly also a bunch of HTC
smartphones supported.

TECHNOLOGY


Some words about the technology choices we have made for Aurora. The UI
components of Aurora will be based on Qt’s QML
(Qt Markup Language) and will have parts written in C++ and Vala
(although we’re going to use Python for prototyping as well).
We will support both Qt/X11 and Qt/Embedded, the latter being useful on
smaller systems, such as the OpenEZX family of devices
(48MB RAM, no GFX acceleration, etc.)
For the first release we will only provide Qt/Embedded based images for
the Palm Pre devices;
those flashable images will be based on OpenEmbedded, however we’d
welcome people taking care of creating releases based on Debian, Gentoo,
etc.

THE CODE

At the moment, there is not much to look at, but feel free
to download the current status via git.freesmartphone.org -> aurora.

HOW TO HELP

Speaking about welcoming people, the major aim of this announcement is
to find people who want to share this vision
and give us a bit of a hand. We are especially lacking artists and folks
who can improve our user interaction.
Apart from the technical reasons, we chose QML to have a very low
barrier of entry. QML is easy to understand
and it also comes with a GUI design tool.

If you are interested and share our vision, please feel free to contact
us. We can then see how you could help us to get to the end goal (see
roadmap) even faster.

There are two possibilities to make us aware of you:

– IRC: irc.freenode.net; channel: #openmoko-cdevel
– Mail: 

Thanks for your attention,

Mickey & Morphis.

Some may ask: Why should I use my smartphone as a featurephone? Well, you don´t have to. But if you are interested in free software, this is a way to free your phone. Just think of it as a cool idea: My phone runs on free software. Pretty cool, huh?

Comment: Delayed AOSP release of Honeycomb

Recently, Google´s Andy Rubin stated that it would take some more time until they would release the Sourcecode of Android Honeycomb. This is not a new phenomenon, it often took Google longer to release the Sourcecode than the SDK or the first device running the same Android release. This time it´s going to take a lot longer – but, honestly, this is not surprising.

First of all, one hears of random application / service crashes in Honeycomb, secondly screenshots of Honeycomb on smartphones have indicated, that this Honeycomb platform could have used some more time in the lab – I believe that the early (in terms of code finalisation) release of Honeycomb was actually a business decision: Get people exited with Honeycomb tablets before the iPad 2 is out must have been the idea, and actually, this seems to have worked out – while many tech journalists say, that the iPad 2 is the best tablet on the market right now, most of them mention that Honeycomb nontheless is a strong platform that just needs the right hardware (Samsungs new Galaxy Tabs are going to be great, the Motorola XOOM is often refered to as early, thick and heavy) to be a really strong competitor.

Releasing a software really fast always means that there are some ugly hacks in it, and it makes perfect sense that Google doesn´t want to release  source codes that contain ugly hacks and doesn´t run well on every hardware platform – and they can do so due to choosing the Apache license, which isn´t as demanding in that sense as the GPL (especially GPL v3) is – they may be sued to release the GPL licensed software parts though (kernel and a few libraries).

Still, openess is one of the mayor selling points of Android, and so I am very positive, that Google will release the source code of Honeycomb as soon as they have fixed the issues in it – they could have made Android closed  source in the first place if they had wanted too – you don´t need a Linux kernel to run Dalvik – a story the Blackberry Playbook tells us (most likely, they might actually run a virtualized Linux kernel – the TI OMAP 4 SoC is definitely powerful enough to do so).

On the other hand, as someone watching community projects and being in contact with open source people for some time, this is the time to mention again that Android isn´t as open as other projects – new releases are always created behind closed doors, while some bits others changed on the previous release are pulled in, this is not open development. It´s just a opening to allow modifications and additions, but then, again, this does it for many and is in a way more than what you are able to do on other platforms that rely on more GPLed code (think of webOS and the no ogg, no flac media player) – the platforms efforts that are entirely open source are rarely supported / pushed forward by companies and thus remain enthusiast free time projects (G(PE)² is an example here, as it was backed by a unnamed company for a rather short time).

Android is just the most open / free mayor platform out there, but that´s something, isn´t it?

LiMo R4 announced, devices to come?

When LiMo 3 was released, nobody was really excited – it was last year in february, and at that time Intel and Nokia got all the attention for their MeeGo announcement, and besides that, Android was becoming huge at that time. Now, with MeeGo being pretty much dead in the handheld / smartphone form factor, this may change. While there wasn´t much reaction on the announcement of  LiMo 4 during MWC this might change as soon as devices will be out (there is not a single LiMo 3 device available in Europe, btw).

When you look at the LiMo platform architecture overview, and you know what this looked like before, you notice that there is a new UI stack: EFL – Enlightenment Foundation Librarys. In fact, this is a contribution by Samsung, who hired Carsten Haitzler, BDFL of this project, a while ago.

At CeBIT 2011 I had the chance to talk to C. Haitzler for about half an hour, and he told me that “there will be something in the future” – of course he can´t talk about anything, but with Samsung investing in EFL to get a faster UI stack into LiMo, it´s likely that there will be Samsung LiMo based device later this year (he also told me that there is no EFL inside Bada), I guess that there will be an announcement in summer. I´ll not speculate just quote a few lines he wrote in his blog:

Samsung is putting real resources behind EFL and using it to make a production-ready OS. The OS not only is Linux based, It uses all the other infrastructure from Linux (DBus, Glibc, Xorg, and much much much more). It is also going to be Open Source (GPL, LGPL etc.) and with Opensource upstream gaining contributions back from Samsung.

This is a real effort and not just some research experiment. Stay tuned. Things will only improve from here. If you were hoping for a slew of MeeGo handsets, then maybe you should also keep an eye out for something from Samsung (actual product details not available yet – if it be a tablet, handset or TV or anything else for that matter).

Many of you will ask: Why LiMo? This question has to be asked, I totally agree. The great thing about LiMo is that it builds on “real GNU/Linux” – Android doesn´t, it just uses the kernel and some more small parts, but it has it´s own framebuffer UI stack (no X), its own C library “bionic” (no (e)glibc) and so on. Still, opensource people haven´t been exactly exited about LiMo – LiMo phones aren´t hackable (LiMo2 phones like the Samsung GT-i8320 (sold as Vodafone 360 H1 in Europe) have non writeable rootfs (cramfs) and locked bootloaders, and there were open alternatives like Openmoko. With Openmoko no more in the smartphone market, and it´s dated at release hardware aging even more, this issue is partly solved – while contribution and hacking on LiMo still will be difficult with LiMo being a middleware and an IP pool. Consumers didn´t really like LiMo phones, either – LiMo based deviceslacked an ecosystem in the past, there were few native applications and no free SDK. But these issues are addressable, and with EFL, next generation LiMo based phones will have a blazing fast UI which certainly will be fascinating, and feel more fluid than comparable Android devices. There might be a comeback.

GTA04: Speed up your FreeRunner / Neo1973

After having decided to reactivate linmob (rebranded to LINMOB) I asked myself what would be a good first article, and there could be only one answer: Write about something that isn´t sprayed on the walls of every other tech blog / tech news site, and it would be great if that would fit to past linmob articles.

GTA04 (Image by Golden Delicious Computer)

Luckily I then remembered the GTA04 project (gta04.org) which is building replacement boards for Openmoko´s FreeRunner / Neo1973 phones. The company behind this thing is well known in the german Openmoko community – Golden Delicious Computers / handheld-linux.com – they were not only official Openmoko distributor, but they later offered Buzz- and Bass-fixes.

So, what will we find on these boards?

The main two chips may not be top notch in 2011 (they were top notch in 2009) but definitely a lot better than the dusty chips in the ARMv4/ARM9 GPRS Freerunner:

  • TI OMAP3 3530 600MHz/720MHz ARMv7 / Cortex A8 SoC (almost similar to the OMAP 3430 found in the Palm Pre(Plus), Motorola Droid / Milestone, Nokia N900) with integrated PowerVR SGX graphics
  • Option GTM601W UMTS module

In addition to that you´ll find a fascinating set of features, which is on par with todays solutions:

  • Wi2Wi WLAN/Bluetooth module + antenna
  • GPS module + antenna switch
  • LIS302 (accelerometer)
  • LSM303 (compass and accelerometer)
  • ITG3200 (gyroscope)
  • BMP085 (barometric altimeter)
  • Si4721 (FM transceiver)

What remains yet unknown is the amount of RAM and Flash memory – I hope it will be were it is supposed to be in 2011 – it should be 512MB Ram and at least the same amount of ROM/Flash memory. It should be possible to add in a 1.3 MPixel camera.

What remains is the same 2.8 inch 285dpi VGA screen with resistive touchscreen and the same case – to things that make this upgrade board a lot less interesting for the average user.

But let´s rather talk about the advantages, which is openess. In times where companies like Motorola or Sony Ericsson lock down the bootloaders of their smartphones which makes hacking them (even though they run Android and thus a Linux kernel (looking way back at Motorola EZX/MAGX phones you understand that this is not an Android issue – a Linux kernel does not imply that your phone is hackable), a phone with an open bootloader, embraced by the Open Source community can be a strong solution.

By the way: The GTA04 root is the Openmoko Beagle Hybrid and there are plans for even more advanced devices – of course development costs money – I doubt this will go on without a certain amount of sales.

For further information head over to the already mentioned gta04.org website, to the GTA04 article of the Openmoko.org wiki, to handheld-linux.com or check out the Openmoko Community Mailinglist. If you speak german (or can live with Google translation) make sure to have a look at freeyourphone.de.

Thoughts on usabilty on FreeRunner

Today I installed Debian onto my Neo FreeRunner, to be precise the relatively new, Debian based community distribution FYP, which basically combines FSO/Zhone with LXDE. As it is a very new distribution, this post can’t really critisize it, it’s just about writing up some things i noticed and some ideas regarding usability on the Neo Freerunner – rather regardless of the distributions. FYP is just the drop of water inspiring to write this, the first two paragraphs will be Debian related, so you can probably skip them if you’re in a hurry.

First of all, it is great to have an easy way to setup Debian on the FreeRunner, as I said before, Debian armel has great repos, compared to most other FR distribution, which make it easy to try out many applications on the FR without having to compile them oneselves.

But you have to keep in mind, that Debian isn’t focussed on mobile devices, so additional repositorys with software aimed at mobile devices are a necessity in my opinion. This is an observation I made while using Debian on the MDA Pro/HTC Universal some time ago, and I didn’t change my mind on that, examples i remember are the embedded version of Abiword or tools like xmonobut helping you with right clicks, which would be nice to have in such an additional repository, but which partially could lead to confusion in regular Debian.

But let’s come to real usability issues, anyway. In terms of usability there are two kinds of distributions on FreeRunner, those who aim to be finger-friendly, and those that just want to have basic features finger friendly, but require a stylus for the rest – FYP is one of the latter. I think that both usage ways have their point, i like finger friendly efforts a lot, but I like to use software I know from my desktop as well (BTW: I would almost always use a stylus, as long it is a small and handy one – as i hate finger prints, wear,tear and dirt on the display – even think about getting a capacitive stylus for G1 – nontheless: finger friendly is stylus friendly too, as you don’t have to be that exact), at least as long as their is no finger friendly alternative.

An usability issue, which is rather an aesthetic issue, i see on distributions like SHR is that they mix different toolkits, like etk and gtk – the applications look different, and while enlightenment related apps are close to be too beautiful performancwise, gtk apps often look old fashioned or sometimes even bad at all. I think it was Midori on SHR unstable (I know, just a framework demo, not aiming to be a real distribution, but nontheless used by some as one), which looked really bad, using the nice old orange OM2007.2 icon theme, but a plain grey gtk+ theme (i am sorry that i didn’t check, which one) – but the OM2007.2 GTK theme wouldn’t have resulted in a much more appealing UI anyway.

In my opinion these optical issues should be fixed by an “optical initiative”, which rolls out one theme, which makes GTK+, ETK.. etc. apps fit better optically, while being fast and not eyecandy in a way which messes up performance, as performance does really matter and anyway: I personally adore simplicity. This means: Don’t use too power hungry GTK+ engines (e.g. I was really surprised to see FYP using clearlooks… imho not a good idea) to match up with Enlightenment effects, rather remove some “bling” of Enlightenment.
I believe the style “hackable:1” inherited from OM2007.2, the ASU look and the style of new and upcoming “Paroli” are good examples of what I mean.

Additionally I would propose to implement a simple way to enable the user to switch between GTK-iconsets and themes (or rather two configurations of one theme) like a button among illumes applications (or even better as a panel plugin), as there are basically two kinds of applications using the GTK+-Toolkit on OpenMoko devices: Those written for the Neo FreeRunner and familiar embedded devices (OM2007.2, again) and those written for the desktop but being lightweight enough to run on embedded devices like the FR (btw, it would be nice to have a probably xmonobut like easy way to perform right mouseclicks, too) .

I strongly believe that this could lead to a better usability or even workflow ;-) (and so to more popularity for the free platform of OpenMoko and the community) and would be happy if anybody would pick this ideas up soon – being pretty sure that i am not the only one worried about usability and aesthetics.

Plans for my last igeekmod release

Well, as I don’t look at my statistics, I don’t know how many people are reading this blog. I know that it isn’t that hard to find it using popular search engines, but that’s all. Years ago, I cared about all this SEO stuff, which I am just bored of today – when I comment on engadget I don’t link my blog, and so I do in many forums.

This means, that sometimes the posts I write are mostly memorandums to myself.
But this one isn’t, it is some kind of announcement. I feel ready with my feature list and this means, that i’ll submit it here.

First of all, thanks to our chinese friends, we have EDGE on igeekmod R2 – and of course, I will keep this feature. Ok, sorry. I should talk of the NEW stuff. Well, some additional librarys and programs: I am about to integrate less languages and therefore librarys like ncurses, I will try to integrate elinks, moon-buggy and maybe (not sure yet) irssi. The main problem about additional software is, that I am unable to modify ezxterm to a more comfortable environment with a char-chooser (or call it keycode chooser) to get rid of the need to have an ascii code table in your pocket (or brain) and to be able to easily enter keycodes like F1-F12, as some programs require such input.
Very recently, I had another idea: Changing the browsers identification. I put that on my list, being unsure wether it can be done, because I thought of changing it to androids string might change the look of some mobile services (especially those by Google, if you’ve tried the Android emulator, you might know that there is a difference). And even if the included Opera 8.5 isn’t the best mobile browser on earth, it is not the worst either, I wouldn’t even say it is one of the bad ones.
My list has to more points: VPNC (for my university, i will have to remember to remove my connection settings… ;) ) and little optimizations regarding presets (connections, email-accounts, SIP). That’s it.

And please don’t ask for a day it will be ready: It’s done, when it’s done.* ;)

* But I hope to have it done before OpenEZX/FSO/QT Extended is ready for A910.

Still one A910.

I bought another one, as you know, and it arrived in a good (but not perfect..) constituition. But I was stupid enough to kill my old one, to prove that you shouldn’t flash R58(A910i) firmwares to R57 (A910) devices. (If anyone wants to donate something for my igeekmod work, why not donate an A910 ;))

But as I posted on Motorolafans already, i “finished” two working flashfiles for A910 yesterday. And igeekmod R2 will take its time to make it really that good I want it to be. I hope that I can improve the phones speed, though I have no idea on how to do so (anybody?), as it is to slow as it is.

That was the reason for posting all these ANDROID videos, btw. Let’s hope that WyrM and ao2 will be successfully spending their time on their A910s – ANDROID would be really attractive on A910, as it seems to run really fast – even on relatively underpowered phones.

As time goes by

I have to admit that I haven’t made any progress on A910s firmware in this month, which is worth mentioning. That makes me kind of sad, but as there was plenty of other (somehow more important) stuff to so, it is ok. And there are still some problems.
My HTC Universal is still on Debian/Titchy Linux sometimes, but there is an issue with it: It becomes slower and slower, as time goes on, and I can’t really tell you why. Of course I use apt-get sometimes, but is that the reason for everything becoming a lot slower?
Would be sad, if it was like that.
Anyway, I’d prefer an OpenEmbedded/Ångstrøm based distribution for the HTC, as these are more optimzed size regarding architecture and size – but it’d be great if this could use Titchys daemon, which do a good job.
BTW: Android is said to be delayed. Or not?
And my E2831 lies in its box and I don’t know whether I should sell it (if you are interested, just leave a comment) or not…

That’s it for now.

[non-Linux, but UNIX] iPhone 3G

Apple is a company I like more than other companies. I like the design of their computers and operating systems – which was almost always ahead of (affordable) PC design.

I have never bought an Apple product up to now, my parents bought one years ago, it was the first computer in our house and the first I used. No games besides a puzzle, System 7 and (after I found the solution for the puzzle due to being bored by having a look at all these configuration options the system crashed, we got a ram upgrade (12MB, huge!) and System) 7.6.1.

I don´t think that there are much people in these world which are interested in mobile phones and who don´t know the “iPhone”, which is the third business of Apple Inc. next to “Macs” and “Music”.
While the first iPhone already was a breakthrough, featuring some new and “different” features – mainly the multitouch touchscreen and the acceleration sensors, Apple didn´t change that much with the iPhone 3G. Due to more competition, they made it more affordable, they added 3G (many people (most of them, I suppose, never really used mobile internet) complained about the lack of that feature before) and GPS for location enabled services. I think that GPS is the more “interesting” feature, as there seem to be many ideas of what to do with that.
3G only means a faster “everywhere” internet connection (but remember: go to the countryside, and … probably there is no 3G), Apple didn´t include this nasty video telephony feature – I don´t know anybody who uses this and I read somewhere that people don´t use due to looking ugly on the camera as it has a “wrong” angle (internet video telephony hasn´t) and often bad colours – besides that it is still expensive. But I am pretty sure there will be many guys which are dissapointed by the lack of this feature – all these “feature list” readers, which just compare the features – “Well, I don´t use it, but I might someday…”
iPhone 3G has the same camera. I am pretty sure we will (or do we yet) a lots of complaints about that: “SonyErricson, Samsung, LG, Nokia, HTC, Motorola (soon..) they all have better cameras. 5 megapixel is state of art, why doesn´t Apple include a better camera?
Back when iPhone (2.5G) was introduced, I had a look at pictures at flickr – a comparison of Nokia N95s´camera to the one build into the iPhone. I have to say that they where equal, sometimes I would have prefered the iPhones “photos” if I had to choose.
Besides: If you wan´t to take real pictures you should (still) get a real camera. Those guys who just have a look at the feature list, not at the results won´t buy an iPhone (and it´s OK like that ;) ).

This was all about the somehow less important part – the hardware. The main improvements – at least I think so – are on the software side. iPhone 2.0 software (which can run on both variants of the iPhone – something that sounds not that special to us opensource-people, but hey, it is, have a look at the Symbian or Windows Mobile world, where (official) upgrades to the next version of the smartphones OS are relatively rare) has some features for businessmen (e.g. support for CISCO-secured networks (great for me as well —> university)) and a SDK which seems to work pretty well and which reminds me lot of the SDK demoed in this NeXTSTEP video. And apparently, there will be no charge for “freeware” developers to use the Apple AppStore – which is good news as well, I suppose it is possible to create free software for the iPhone (read this article which says it is not, but distributing source code is never impossible, legal issues are ugly), which uses free software – webkit and (correct me if I am wrong) SQLite.
Unfortunately (I think it is like that) there is no free software SDK – development on Mac only using Xcode (like QuantumSTEP development).

So, what do I dislike about the iPhone?
Well, it isn´t that open as I would like it to be. It has crippled bluetooth (at least I believe it as there where no announcements), no support for bluetooth HID or things like A2DP, PAN. No official “free as freedom” SDK. The way it is marketed seems to change, if there will be legal unlocked iPhones available without contract, this is no more point to complain about – but as we don´t know this yet I have to list that point.

By the way: Since wine is close to 1.0, it would be nice to see some more efforts on projects like GNUstep, it would be “huge” (as “his Steveness” always says) to be able to develop for Apple´s Mac (natively) and free software as well, i would highly appreciate a better integration of GNUstep-software into the standard DEs, or the ability to run “Mac” software (not only free software) on free software.