Why the Vivaldi tablet never came to market
lwn.net
lwn.net
It is just too much cruft on top of what I care about. So if Vivaldi will not come to be, and the alternatives are so so (I did see that project to make a RPi into a tablet with a custom shell; very cool), I am definitely going to pony up the money for a tablet laptop with a touch screen and just spend a little more and save myself the hassle.
I have seen decent reviews of the Yoga 11s. I think I am going to splurge on that. Honestly, it is not much more epxensive than a high end of the line Android tablet, but with no one of the flexibility I waste my time on Linux for so.
So yeah. If someone makes Vivaldi I would be super happy. But I have come to see the truly open tablet as the year of the Linux desktop, in terms of them being pipe dreams, and I would rather focus on things I can use instead of praying for hostile industry players to begrudgingly give me and then make me do a song and dance to get back to root for.
I have fears that the software porting effort is lagging far behind the hardware effort, but I'm putting the odds of getting a usable jessie or Maemo phone a bit higher than the odds of Jolla retreating from the iOS/Android strategy of sandboxing the user for business purposes.
If I had money, I would pledge but I am tight on cash and 500-900USD promise of a phone is out of my league for the coming year.
I think that enabling that kind of thing - running an alternative environment on already existing hardware is a better goal than making your own hardware from scratch. Better as in more realistic to gain a userbase, I mean.
I agree, and we know for a fact that it's a good goal because it's pretty much been the strategy of every desktop Linux distribution for decades. You don't need to own the hardware, you just need to make it work on someone else's.
So I do think at this point desktop Linux is at a crossroads where someone really needs to throw the money in to take over the personal computer space that Chromebooks are already dominating. Except instead of ChromeOS, throw around some Fedora, Lubuntu, Ubuntu, OpenSuse, etc devices. Same general idea, just without saying "were Google, use our cloud stuff because your computer can't do anything else", instead "heres a computer that does all the things the Windows machines do, except on ARM CPUs that cost a third as much and don't compromise".
I guess for that, though, you need freedreno and lima to get some love so we can finally have some first class mobile GPU Mesa drivers, because the binary blobs are hopeless wrecks even if they ran fine in a classic X or Wayland stack.
It might not be less painful in terms of getting it to work, but it would be less painful in terms of getting it to work in a resource-constrained system like an ARM-powered smartphone.
And frankly, I like Xserver support across a variety of devices.
Just as it happened with the OpenPandora.
I can also totally understand the position they were in. My understanding is the vast majority of parts manufactures are over there, and surely all the cheap ones. The Pandora was already going to cost over $300 USD, and I remember the project was taking on some of the costs themselves when reality didn't match with their projections. If they had any hope of getting it out the door at a price that people would willingly pay, you have to take on some risks.
But you can also find a few US or EU based contract manufacturers who have very good partners in Asia such that you start low volume runs and work out the issues close to home and then transition to Asia for the low cost labor once all your manufacturing is working well and has good yields. But this is also not free in cost or time and some companies cannot afford to bear it.
Brilliant! You want a thermal camera with that (muoptics)? or maybe a smart watch (ostro)? or a magic nano ultra capacitor AA batteries?
Main problem was economies of scale, no experience with manufacturing, NIH syndrome, and no market.
1 NOBODY wants a tablet with 2 year old specs.
2 almost nobody wants a KDE/gnome/firefox tablet.
There was nothing stopping them from simply ordering 1000 (pretty standard moq for crazy low prices in $30-50 range) already shipping tablets, and porting their code over.
They have been stuck with x86 based hardware since the ARM GPU drivers are a collective wreck and they use X, so the shitty Android blobs aren't even close to being an option.
I think it was just premature. We need Lima / Freedreno / that open Tegra project to mature into their own, so we can have first class Mesa GPU acceleration on ARM systems, before we can try bringing desktop Linux to said devices. It is the same mistake Canonical made with Mir, and I think it is a mistake with Wayland - the Android GPU stack is a lost cause, and trying to work around it and have the tumor of Qualcomms / ARMs shitty blobs is doomed from the start.
When undertaking a project of this kind, you will need capital. It is clear from this post, the they ran out of personal (?!?) money. Heck, to even "save" money, the manufacturing was taken to a UK company (?!?!?). Not manufacturing hardware of this kind in China is a really really bad idea for several reasons.
> 1) NOBODY wants a tablet with 2 year old specs.
You are mistaken. the low cost tablet market exists. I would rephrase that to most people do not want a slow / unresponsive tablet. those are two very different things.
> 2 almost nobody wants a KDE/gnome/Firefox tablet.
How can you even make that statement when you don't even know what that would be like? People want a tablet that responds well, to browse the web, send email and check their Facebook acct. For some segments of the world population the ecosystem of 3rd party apps is irrelevant since they never ever download an app.
(photo not mine)
[1] http://www.gregwar.com/posts/debian-on-a-low-cost-tablet-50
There is no market for Linux devices because everyone wants them to be "cheap". I would like a high end tablet that runs Linux myself...
I read the grandparent comment along the lines of 'I want to see how to put Linux on a tablet. People will be hacking this with the possibility of bricking the device. I'll search for cheaper devices'
Bus stops: local transport people experimenting with e-paper displays for real time timetables. Kindle sized (might even be modified kindles?).
The display/info space looks like a possibility.
As a potential consumer of a kiosk, I don't particularly care what OS it's running.
I think you'd be wrong if you knew how much the total sum of all Windows licenses cost in a large-scale deployment. That is why they usually install the kiosk with a decade-old Windows OS, usually Win2K or XP embedded which arguably costs less than a Win7 or 8.1 license.
> As a potential consumer of a kiosk, I don't particularly care what OS it's running.
I am afraid we are not particularly talking about consumers, but the companies that make these things.
Also, mobile is the perfect direction for a perceived "top heavy" desktop to move towards, because it requires you to simplify your UI to accommodate it.
The thing in the KDE world to understand is that QML changes everything. The Plasma widgets we have now were built on top of Qt4 series technologies and were often language hodge podges of Python, Ruby, Perl, C++, etc. They weren't hardware accelerated, and the standard toolkit was lacking compared to the QtQuick Controls of today.
And the entire QML thing is about fluidity. Whereas old Qt Widgets let you build wildly complex UIs that were pretty much static, QML is best suited to simple parts doing crazy transitions and fancy effects. Even if you convey the same complexity, the appearance difference can be enough to be intuitive for some where, say, kdenlive is not.
So I don't think Plasma Active is a "mistake". It is a great place to see the next geenration of applications for KDE in general come from, since they have to be focused. It is slow, and it is not going to see any market adoption any time soon, but trying to aggressively take the mobile market I feel is a mistake right now - it needs mindshare that requires Qt5 adoption which requires the KDE5 series to get into its stride first.