Related
so when do we expect this stuff in our cooked roms?
-----------------------------------------------------------
http://wmpoweruser.com/?p=3962
"Datalight announced Reliance Nitro, a new tree-based file system with broad performance implications for embedded device manufacturers around the globe. Early tests show sequential and random write speeds as much as five times faster on Microsoft Windows Mobile than the default file system, allowing devices like smartphones and digital cameras to store and access large amounts of data without the lag time often experienced when dealing with multimedia and other large files.
To make Reliance Nitro faster than other file systems, the Datalight engineering team analyzed all I/O operations, eliminated all redundant calls to the disk, and optimized the file system’s disk operations. This translates into much faster throughput speed for read and write operations. In fact, the revolutionary tree-based allocation structure allows Nitro to be up to 100 times faster on file operation speeds, enabling a much more responsive user experience.
The new file system will initially be available for Windows Embedded CE, Windows Mobile, and Wind River VxWorks. An RTOS porting kit will be offered to allow porting to any 32-bit operating system. Like Reliance, the well-known transactional file system on which it is based, Reliance Nitro provides 100% data reliability and fast mount times.
“Our intent in building Reliance Nitro was to mitigate the sluggish performance associated with increasing data storage on devices,” said Roy Sherrill, Datalight CEO. “Nitro will allow users to finally enjoy the high speed capabilities of flash. We expect fast adoption as manufacturers vie for the loyalty of customers seeking better performing devices.”
Reliance Nitro is available immediately from Datalight and the Datalight worldwide network of channel partners. Please visit them at http://www.datalight.com/companyinfo/resellers.php to find a reseller near you.
The Reliance family of file systems and FlashFX family of flash media managers comprise the Datalight flash file system solution. Reliance was designed from the ground up for high reliability applications. Dynamic Transaction Point™ technology provides 100% immunity from file system corruption, even after unexpected system interruption. Embedded applications can benefit from faster boot times that remain consistent for the life of the product, regardless of disk size. FlashFX™ Pro features pre-written support for over 300 flash parts, works with virtually any NAND controller, and features wear leveling, bad block management, and background compaction for unrivaled performance.
For more information, visit http://www.datalight.com
via Windowsfordevices.com
"
Yes, this looks interesting! Can't wait (but I guess anyone on this forum would want such a performance boost)
Hmm, how will this be implemented in , let say ... Windows Mobile?
We buy this as application and then install it?
Or we now pray that Microsoft will license it and put it in the next Win Mo 6.5 / 7?
Maybe this is real but i'm not sure because today is 1st april
yeah i forgot about that. may not be april fools though http://www.datalight.com still has the info. they should have picked a better time to announce hard to swollow exaggerated 5X performance claims
fatso485 said:
yeah i forgot about that. may not be april fools though http://www.datalight.com still has the info. they should have picked a better time to announce hard to swollow exaggerated 5X performance claims
Click to expand...
Click to collapse
give it a rest its most probably just Aprils fools. if it was true it would have been posted EVERYWHERE. ive only seen it posted in your link
ok ok people, seems fake stop rubbing it in my face
Not an April Fool's
Ok am late to reply here but I am the product manager for this product and Datalight and this is a real product that we are selling to WM OEMs. We announced it on March 30th but the press picked it a day late and ran the story on April 1st.
The flash management feature cannot be cooked into ROMs since it has close hardware ties as well as is specific to the flash part on a specific device (a same model may have different flash parts).
The file system also requires custom integration though we are considering having a pre-built driver to make available to the enthusiast community for non-commercial use. if there is interest, please let me know in this thread.
madmak said:
Ok am late to reply here but I am the product manager for this product and Datalight and this is a real product that we are selling to WM OEMs. We announced it on March 30th but the press picked it a day late and ran the story on April 1st.
The flash management feature cannot be cooked into ROMs since it has close hardware ties as well as is specific to the flash part on a specific device (a same model may have different flash parts).
The file system also requires custom integration though we are considering having a pre-built driver to make available to the enthusiast community for non-commercial use. if there is interest, please let me know in this thread.
Click to expand...
Click to collapse
If this is real, I DEFINITELY have a LOT of interest, as I'm sure most of this community would.
Thanks!
jim256 said:
If this is real, I DEFINITELY have a LOT of interest, as I'm sure most of this community would.
Thanks!
Click to expand...
Click to collapse
wot he says
How about a demo or something? Some benchmarks... give us something to go on. What OEMs are you talking with?
For more details about the benchmarks, please refer to this whitepaper (registration required, you won't be spammed - i promise).
http://www.datalight.com/mydatalight/active/download_resource.php?resourceid=830
We also have a demo video comparing TFAT (with exFAT) with Nitro on the Beagle Board (which has the Omap 3530 processor). You can check that on Youtube at http://www.youtube.com/watch?v=z_Husw6oxT8
any update on this?
btw, this also claims to get major performance boost by doing some intresting tricks http://forum.xda-developers.com/showthread.php?t=530710
how soon can this be implemented in smart phones? compatibility? cost?
hope this is not dead
THE GRIZZ said:
hope this is not dead
Click to expand...
Click to collapse
Dont hold your breath. It wont be coming.
cyberhern said:
Dont hold your breath. It wont be coming.
Click to expand...
Click to collapse
Looks like somebody else beat them to it: http://forum.xda-developers.com/showthread.php?t=530710
And it is free!!
Android apps are about to get 3 times faster!
Myriad Group AG, a member of the Open Handset Alliance, unveiled yesterday a brand new high performance virtual machine, named Dalvik Turbo. It would replace the standard Dalvik engine that Android uses to run applications.
Dalvik Turbo increases application execution speed by up to three times ‐ allowing OEMs and operators to bring smoother delivery and more complex applications to Android phones, while also providing substantial battery life improvements when running resource intensive tasks. It also enables developers to take full advantage of Android, allowing them to create games boasting advanced graphics and complex models while retaining full compatibility with existing software.
You can bet the guys at Google are all over this right this moment. A three times speed increase on Android apps would allow even the G1 to run the most complex apps. It would also improve battery life, something smartphones tend to suck at. However, the best part about this is “full compatibility with existing software”, meaning we could see the benefits of this new Dalvik virtual machine very, very soon. This is one of the beauties of Android, things like these could only happen on an open environment where several companies innovate together.
The guys from Myriad will be at the Mobile World Congress in Barcelona from 15th‐18th February 2010. That is just around the corner. This is going to be one of the biggest news during the event, can’t wait for it.
Late to the party.
http://forum.xda-developers.com/showthread.php?t=630523
i could live with my apps running 3 times faster, but wont bealive it till i see it
prettyboy85712 said:
Late to the party.
http://forum.xda-developers.com/showthread.php?t=630523
Click to expand...
Click to collapse
Lame response
I think this is great news especially since they're a member of the Open Handset Alliance.
I knew that the VM had to be improved if we wanted any big Android performance improvements. Why haven't any devs tried optimizing DalvikVM?
With all the source available, why hasn't someone modded the window manager to implement real multitasking?
I've only seen tweaks here and there... if the source is available to both homebrewers and manufacturers, why can't we make something as good as what HTC has done to Android for the Hero?
PSP_Hacker said:
I think this is great news especially since they're a member of the Open Handset Alliance.
I knew that the VM had to be improved if we wanted any big Android performance improvements. Why haven't any devs tried optimizing DalvikVM?
With all the source available, why hasn't someone modded the window manager to implement real multitasking?
I've only seen tweaks here and there... if the source is available to both homebrewers and manufacturers, why can't we make something as good as what HTC has done to Android for the Hero?
Click to expand...
Click to collapse
1. we are not a company where we can have deadlines of stuff and devs get paid at least 150,000 a year.
2. we have devs that do this for their own enjoyment and their own liking. If the little tweaks here and there are enough for them to enjoy the phone, its their call since they are the publisher of the product.
3. lastly, i'm sure pay all these devs at least $150,000 a year on top of what they earn on their regular work and I'll tell you, we would have a blast of new improved stuff.
4. It's not that all devs want money. But devs won't do something so hard, esp if you have to do research and stuff, and it'll prolly take months in research at a company, for free. If companies are researching on improving those fields, why would our devs spend time improving the same said fields?
just thank our devs for the hardwork they do and for all the overclocking and stuff they spend their free time on.
PSP_Hacker said:
I think this is great news especially since they're a member of the Open Handset Alliance.
I knew that the VM had to be improved if we wanted any big Android performance improvements. Why haven't any devs tried optimizing DalvikVM?
With all the source available, why hasn't someone modded the window manager to implement real multitasking?
I've only seen tweaks here and there... if the source is available to both homebrewers and manufacturers, why can't we make something as good as what HTC has done to Android for the Hero?
Click to expand...
Click to collapse
1) I think the majority of us don't like Sense.
2) Why don't you do it?
3) Read the post above mine.
PSP_Hacker said:
Lame response
Click to expand...
Click to collapse
Well not really. We are a huge site and avoiding multiple threads on the same subject really helps when searching for answers.
So please use:
http://forum.xda-developers.com/showthread.php?t=630523
Thanks
Mike
http://www.computerworld.com/s/article/9218075/Mystery_of_the_missing_Honeycomb_apps
Really good, looks like they did their homework...not that I feel much better about hoping more apps come along shortly after reading this-lol
Question: Tegra Zone is mentioned as a place to, also, get apps. Is anyone familiar with them as in do I have to be rooted, etc....I've never heard of them.
sunsetbay said:
http://www.computerworld.com/s/article/9218075/Mystery_of_the_missing_Honeycomb_apps
Really good, looks like they did their homework...not that I feel much better about hoping more apps come along after reading this-lol
Question: Tegra Zone is mentioned as a place to, also, get apps. Is anyone familiar with them as in do I have to be rooted, etc....I've never heard of them.
Click to expand...
Click to collapse
They sure did, and as you say...there won't be a rush for more apps on the market.
For the tegra zone:
You do not need to have your TF rooted. It's listed at googlemarket
https://market.android.com/details?id=com.nvidia.tegrazone&feature=search_result
sparven89 said:
They sure did, and as you say...there won't be a rush for more apps on the market.
For the tegra zone:
You do not need to have your TF rooted. It's listed at googlemarket
https://market.android.com/details?id=com.nvidia.tegrazone&feature=search_result
Click to expand...
Click to collapse
Thanks for the reply, sparven89
I think I will give this app a try and I see they have a couple of games I might want to try.
Thanks!
I enjoy crapshooting and refunding apps. I bought over 500 apps, refunded over 300. It is a fun process
frosty5689 said:
I enjoy crapshooting and refunding apps. I bought over 500 apps, refunded over 300. It is a fun process
Click to expand...
Click to collapse
+1, if only Google gave you a bit more than 15 minutes before you could refund the app you don't want .
Yup it's odd how Google are known for their great search engine, yet have the worst search function ever in the android market. I don't buy that bit in the article about there not being enough demand for developers to want to create apps for HC though. There's loads of demand for apps in just about every section (other than for twitter, enough already!).
Everyone,
I found another new article about Android apps vs HP Touchpad having 400 at release on July 1st - (Android around 200+):
"Does the HP TouchPad Have More Tablet Apps Than Android?" (Apparently it may have and how they did it...yet I wonder if this is true)
http://www.pcmag.com/article2/0,2817,2387999,00.asp
I do like this part in the article "Google, as I've said before, wouldn't comment. But I've heard through the grapevine that they're working on improvements for the Android Market—it's just nothing they'll confirm. It's an obvious move, though."
On the personal side, I love my TF and number of apps isn't an issue for me-I have plenty(s). But, I am following the posts of others and some apps they want, so I find the whole subject interesting.
Google has so much riding on this...changes have to coming shortly that helps the developers and what they're waiting for. Well, just my 2 cents(s) and interesting to follow.
frosty5689 said:
I enjoy crapshooting and refunding apps. I bought over 500 apps, refunded over 300. It is a fun process
Click to expand...
Click to collapse
jonitfcfan said:
+1, if only Google gave you a bit more than 15 minutes before you could refund the app you don't want .
Click to expand...
Click to collapse
Seems more behind the scenes on this front happening, at least in Taiwan:
http://www.taipeitimes.com/News/taiwan/archives/2011/07/01/2003507141
Apple is complying with Taiwan's 7 day refund policy on their stores and Google
has pulled their paid apps, so far, rather than comply...well!?? Google has to July 14th, certainly hope they don't handle it like this.
I agree with you both 100% that 15 minutes is not enough time. I've already
been caught on an app Moon+Reader Pro $4.99 causing my screen to flash repeatedly when I used the Text to Speech, so I immediately deleted it.
Since it took me almost an hour to get to this point, I wrote the developer for a refund...very polite & I'd be willing to try again if it was fixed. I waited 3 days and no response, then I left a 1 star review in the marketplace, again polite & details.
My review days later never showed up either. Call me skeptic, but I have wondered if I was censored by the 1 star rating & possibly because I said the developer hadn't responded to refund request...ah, guess that's a clue-lol.
So what are you going to do, just being honest & following the rules Google set in place, didn't work for me.
Shame because other's experience is what I reply on when purchasing an app.
Actually, this has been my one and only problem. But, it was a biggie the way it fell through the cracks when I actually asked for a refund.
sunsetbay said:
Seems more behind the scenes on this front happening, at least in Taiwan:
http://www.taipeitimes.com/News/taiwan/archives/2011/07/01/2003507141
Apple is complying with Taiwan's 7 day refund policy on their stores and Google
has pulled their paid apps, so far, rather than comply...well!?? Google has to July 14th, certainly hope they don't handle it like this.
I agree with you both 100% that 15 minutes is not enough time. I've already
been caught on an app Moon+Reader Pro $4.99 causing my screen to flash repeatedly when I used the Text to Speech, so I immediately deleted it.
Since it took me almost an hour to get to this point, I wrote the developer for a refund...very polite & I'd be willing to try again if it was fixed. I waited 3 days and no response, then I left a 1 star review in the marketplace, again polite & details.
My review days later never showed up either. Call me skeptic, but I have wondered if I was censored by the 1 star rating & possibly because I said the developer hadn't responded to refund request...ah, guess that's a clue-lol.
So what are you going to do, just being honest & following the rules Google set in place, didn't work for me.
Shame because other's experience is what I reply on when purchasing an app.
Actually, this has been my one and only problem. But, it was a biggie the way it fell through the cracks when I actually asked for a refund.
Click to expand...
Click to collapse
am i being daft ? ive got no 'optimised for tablet' section in my market app ????
wilbur-force said:
am i being daft ? ive got no 'optimised for tablet' section in my market app ????
Click to expand...
Click to collapse
I'm not aware of anything like this either. Is there supposed to be one ?
I think the tablet section is only for the USA.
Damn those Yanks, getting everything first...well, almost .
Misleading
There are several things mentioned in the article (like every other article decrying the state of HC apps) that are false or misleading:
The number of Honeycomb-optimized apps remains in the low hundreds. By comparison, there are over 100,000 apps optimized for the iPad.
Click to expand...
Click to collapse
Of course the iPad has been out for a full year longer, that number was much lower three months after its release (which is approximately how long its been since the Xoom was released). It looks like it was about 10,000 based on a quick googling. That's quite a big difference from 100,000 (like, an order of magnitude).
There is also the question of how they are counting these "Honeycomb-optimized apps". There is no provision in the market, or in the metadata that describes an app to determine if it is "optimized" or not. Unless a developer explicitly denies the app for a certain screen size, a certain device, or requires a feature that the device doesn't have, then it will most likely run. If its just a question of "will it run?" then we have tens (maybe hundreds) of thousands of Honeycomb apps. Beyond that, where can you draw the line to say that an app is "optimized"? This is a subjective question that can't be used as the basis for such an empirical evaluation.
I can offer my own apps (linked in sig) as an anecdotal example. Both worked on day one on Honeycomb, with no changes needed. All the text and graphics were crisp, clean and appropriately sized. They did not look ugly, just not as efficiently designed as they could have. Shortly after HC dropped, I updated them with a minor change to make use of the native Action Bar in HC for menu items. The rest of the UI (mostly list views), while great for a small phone screen, wastes a lot of space on a tablet screen that is several times larger. I'm currently working on refactoring the UI components into fragments, to improve the HC experience. At this point, all the fragments are in, and working, but I have not yet implemented a new activity/layout for HC devices which will make use of the extra screen space to place multiple fragments on the screen at once. At what point is my app then "optimized"? Day one (when it worked)? Once it made use of the action bar (more consistent UI)? When it was refactored into fragments (easier to detect this automatically)? Or when I add separate layouts that make use of a larger screen's real estate (most of us would say this is "optimized", but the point is we don't have any way to tell besides looking at it and making a judgement).
All Android 2.x apps will run on Android 3.x, but oftentimes the text just looks awkwardly small and clunky; this is because the Android SDK does not automatically scale fonts, so developers must create their own code to get their fonts to scale up or down within their apps, which can be extremely tricky.
Click to expand...
Click to collapse
As long as fonts are specified in "sp" (scale independent pixels) units, this happens automatically.
When graphics are involved, you must use higher-resolution graphics and program them to be scalable, so that when they are on a larger screen they won't become pixilated. However, this requires the size of your app to be larger, and it may put more of a strain on the device's processor.
Click to expand...
Click to collapse
Nope. Having different resources for different screen densities or sizes is well documented and as easy as putting them in specially named folders. This was already done to make apps look good on LDPI, MDPI, and HDPI devices. It does increase the size of the app, but it will have virtually no performance penalty.
Many pundits have speculated that perhaps the holdup is in anticipation of Ice Cream Sandwich, which will theoretically merge Android's phone and tablet platforms, helping to end fragmentation, and which is due to arrive this fall.
Click to expand...
Click to collapse
Even if this were true, its not the release of ICS developers are waiting on, but the point where adoption is widespread enough that they can safely drop support for previous versions (this wont happen for quite a while, as we've seen from the last several releases of Android).
To the author's credit, they acknowledge this and state that none of the developers they talked to felt this was a roadblock, but I have seen other articles promoting this idea.
This article was one of the more fair and balanced looks at the subject I have seen, but still upholds some myths about Android app development. The fact is, most any app that was built following the Android UI guidelines worked and looked fine without any changes for Honeycomb. Their UI may just not have been the most efficient for a tablet sized screen. In comparison, iPhone apps, using the 2x scaling on the iPad, look pretty terrible. This, combined with faster hardware adoption by consumers, is what prompted developers to port more apps to the iPad quicker than we are seeing for Honeycomb. The fact that my apps looked "good enough" on HC is why I didn't feel like there was much of a rush to put in the work to "optimize" them. Google made the right call by making Android as resolution/density agnostic as possible, as opposed to iOS's initial assumption of a single screen size and resolution. While iOS developers had to scramble to get iPad optimized versions of their apps out quickly, Android developers are able to take their time and (hopefully) do it right. There will be more HC "optimized" apps, as more tablets are sold, and more developers decide its worth their time to invest in it.
maxpower47, Thank you as a developer jumping in and adding clarification. I've reread your post several times and now I'm clearer on understanding the article and difference between optimized and 'working apps' as the article seemed to implied to be very low in number.
"Even if this were true, its not the release of ICS developers are waiting on, but the point where adoption is widespread enough that they can safely drop support for previous versions (this wont happen for quite a while, as we've seen from the last several releases of Android).
To the author's credit, they acknowledge this and state that none of the developers they talked to felt this was a roadblock, but I have seen other articles promoting this idea."
Interesting!
"This article was one of the more fair and balanced looks at the subject I have seen, but still upholds some myths about Android app development. The fact is, most any app that was built following the Android UI guidelines worked and looked fine without any changes for Honeycomb."
This sure is a myth that I was holding, so again I appreciate your contribution to understanding this better...much appreciated!
sunsetbay said:
Everyone,
I found another new article about Android apps vs HP Touchpad having 400 at release on July 1st - (Android around 200+):
"Does the HP TouchPad Have More Tablet Apps Than Android?" (Apparently it may have and how they did it...yet I wonder if this is true)
http://www.pcmag.com/article2/0,2817,2387999,00.asp
I do like this part in the article "Google, as I've said before, wouldn't comment. But I've heard through the grapevine that they're working on improvements for the Android Market—it's just nothing they'll confirm. It's an obvious move, though."
On the personal side, I love my TF and number of apps isn't an issue for me-I have plenty(s). But, I am following the posts of others and some apps they want, so I find the whole subject interesting.
Google has so much riding on this...changes have to coming shortly that helps the developers and what they're waiting for. Well, just my 2 cents(s) and interesting to follow.
Click to expand...
Click to collapse
Android SDK sucks balls.
WebOS SDK does not.
Result = More WebOS apps.
JCopernicus said:
Android SDK sucks balls.
WebOS SDK does not.
Result = More WebOS apps.
Click to expand...
Click to collapse
Um, er... what? More webOS apps than Android apps? In which universe?
I'll tell you, I stuck with webOS for 18 months before giving up because the app situation over there is such a mess. I think there's still only about 6,000 official (non-homebrew) apps for webOS, and many (most?) of those don't run on the TouchPad outside of a silly little emulation window. And there's as much junk in the webOS App Catalog as there are good apps. I think the only category with good apps is Twitter clients, which is why you'll find most webOS entusiasts constantly talking about the latest Twitter app.
I'm also astounded that there are only 300 or so TouchPad-optimized apps, given that HP is the world's largest tech company and has a huge vested interest in seeing webOS succeed. Couldn't they do SOMETHING to get more apps written before launching the TouchPad?
My golly, but saying that the webOS SDK being good (which is hugely debatable, given all of the SDK's limitations since webOS was announced) resulting in more webOS apps is just laughable. I have thousands of apps available for my TF, across every major category. If I were to swap my TF for a TouchPad, I'd be giving up a wide variety of useful apps and would essentially give up most of the value of owning a tablet.
Sent from my Transformer TF101 using Tapatalk
I meant in terms of "tablet" apps.
(which I know just means XL v11 support).
maxpower47 said:
Or when I add separate layouts that make use of a larger screen's real estate (most of us would say this is "optimized", but the point is we don't have any way to tell besides looking at it and making a judgement).
Click to expand...
Click to collapse
This is definitely "Optimized" to me. As much as function is important, I think it's equally as important for an app to utilize screen real-estate properly.
The fact is, most any app that was built following the Android UI guidelines worked and looked fine without any changes for Honeycomb. Their UI may just not have been the most efficient for a tablet sized screen. In comparison, iPhone apps, using the 2x scaling on the iPad, look pretty terrible. This, combined with faster hardware adoption by consumers, is what prompted developers to port more apps to the iPad quicker than we are seeing for Honeycomb. The fact that my apps looked "good enough" on HC is why I didn't feel like there was much of a rush to put in the work to "optimize" them. Google made the right call by making Android as resolution/density agnostic as possible, as opposed to iOS's initial assumption of a single screen size and resolution. While iOS developers had to scramble to get iPad optimized versions of their apps out quickly, Android developers are able to take their time and (hopefully) do it right. There will be more HC "optimized" apps, as more tablets are sold, and more developers decide its worth their time to invest in it.
Click to expand...
Click to collapse
This all makes a lot of sense.
What doesn't make sense, as a consumer, though... there is no way for us to search for apps that *are* optimized for tablets. Even with "optimized" being hard to define, there needs to be a way for a user to head into the market and search for apps that work nicely on their tablet. It kills me that I have to depend on forums to find apps that fit my screen well.
HA!!!
I just find what I like from my phone and use that.
Or I think of other things to be done with my amazement machine I: study, download, try, and if I like it buy! If your not an informed consumer then you dont belong on android anyways. Sorry if that sounds crass, but if you want someone to tell you what to by then by all means have fun following the heard.
haensgn said:
HA!!!
I just find what I like from my phone and use that.
Or I think of other things to be done with my amazement machine I: study, download, try, and if I like it buy! If your not an informed consumer then you dont belong on android anyways. Sorry if that sounds crass, but if you want someone to tell you what to by then by all means have fun following the heard.
Click to expand...
Click to collapse
*I* am obviously fine, as I'm here and other places researching apps.
But from a general user population point of view, it's extremely consumer unfriendly to not have a way to find "native" (using that term loosely) apps for your new "hot" products.
Also, there is no reason to be rude about it, imo. It's simply an opinion on customer service, which I think Google has a poor track record of, unfortunately.
Honeycomb is absolutely great, some of the tablets released also great (I love my TF).... but the Android Market needs some serious work, imo, especially in terms of the search engine (ironic hah).
In all honesty (and potential ignorance), I'm floored by the difference in commonly accepted time frames for testing desktop tweaks vs. mobile tweaks ("tweaks" used in the most general sense) and the influencing factors. Installing a new kernel on my Fassy feels like installing a new CPU in my desktop in terms of potential changes (good or bad) in stability, features, general performance, and tweaking potential, minus the thermal grease under my nails. In my PC OCing days though, people would commonly balk at anyone claiming a stable config without the 24h Prime95 runs, repeated SuperPi results, etc. as evidence; I really haven't seen anything like that in my limited smartphone experience, but I understand some of the reasons why, aside from being impractical for most users.
I find it easier to get spoiled on the mobile platform. It seems the majority of users (myself included) are less inclined to put a new kernel through its paces for more than a day or two before reaching a conclusion. With kernel and rom releases being far more frequent and accessible than, say, expensive user-upgradeable hardware, it makes perfect sense. Still, I often wonder if a given kernel or rom release has been deeply and broadly tested enough to truly provide the developers with reliable feedback. Dedicated beta testers are a godsend, but I'd venture a guess that such groups are frequently unavoidably too small to serve as a truly representative sample set. I'd be an idiot to overlook all of the amazing development progress many of us have grown used to (hence, "spoiled"), but I feel it would be shortsighted to assume that only a rare few serious devs out there have given up their free time on red herrings.
At last, there is an actual question here. I don't expect a concrete or "right" answer, though. To anyone reading, I ask: do you feel that developers and their projects (and thus the rest of us) would see any appreciable benefit from the widespread adoption of longer testing times? Obviously you don't need a week to figure out you're in a boot loop, but I feel like a lot of cumulative minor errors end up rearing their ugly heads despite a user thinking their configuration is "stable" after only a day or two.
NOTE: Despite the presence of a question, the overall intent of the post didn't seem to fit the Q&A section. I apologize if this was an incorrect assumption.
No. When something is released on xda it is immediately in "testing". For a developer to have user feedback, logs, and the discovery of potential app conflicts is what xda is all about. And you can see that type of "growth of an app" happen in android apps section. Many apps there started as v.01 alphas and are now full releases on the market (thanks to testings from courageous xda addicts). But for the non commercial side of things roms/kernels/themes. Being on the cutting edge comes with risks or if you're not sure of something, resist the urge and monitor the thread for a week or two and see if the modification suits your needs.
good day.
Chopper, I agree in principle but not completely in practice (though you have a lot more experience to speak from). As far as the "in testing" nature of projects posted here, I suppose my uncertainty goes back to the question of the quality and consistency of said testing. The kernel.org fiasco comes to my mind as a potentially interesting case study if significant time and resources could feasibly be allocated. I was intrigued to see a situation where such a large and varied (device-wise) group of users had to "settle" with the existing builds for so long. This allowed more time than usual for users to test and agree on the stability of a given build. Whether that was truly widespread or meaningful, I cannot say. To your closing point though, I was happy I decided to stick with 35 for a few days once the nightlies resumed.
Sent from my SCH-I500 using xda premium
Hi all
MY QUESTION
How much work is it to get Ubuntu working on a cheap tablet, (in terms of weeks and stress/reliability)? I'm about to spend a year writing tablet software that needs cheap hardware. If I find a capable Android tablet going cheap, is it reasonable to consider getting Ubuntu working on it, instead of restricting myself to the Android OS to use cheap tablets? Would Ubuntu C++ apps still kill performance? (Ubuntu will save me lots of development in other ways.)
ALL comments pleease, however brief and knee-jerk.
BACKGROUND (all feedback gratefully received)
I'm at the design stage of a project to use tablets to improve education in poor countries. Extremely briefly, the tablets will use elements of social media to enable children to collaborate remotely and asynchronously on projects, and game aspects to get the kids excited, who have probably spent the day working on the fields, and to welcome kids with special needs. The system will enable education to continue in complex emergencies, such as droughts and conflicts. Currently most kids drop out after grade one as the education they're offered is so poor. Tablets can support teachers and enable kids to get more out of their classroom and homework time.
The system needs some clever back-end engineering to operate a local social network if there is no internet connection. (I'm thinking something like NodeJS acting as both a p2p client and a server.) It also needs to run on cheap devices, if it is to be adopted by third world Ministries of Education.
I am currently torn between Android and Ubuntu for tablets. Android will presumably be the cheapest platform for the foreseeable future - tablets now go for as little as $40 wholesale. However Ubuntu for tablets now offers the ability to bring a proper IT education to these children, as they can learn office software, desktop OS, etc. Ubuntu also provides source code I can customize, eg, GCompris, Tux4kids, KDE and Epoptes. I can see Ubuntu on other tablets here, but it seems the Nexus 7 is the cheapest tablet I can currently get Ubuntu on and performance is still an issue. Is that fair to say?
I can write everything using C++ and OpenGL to squeeze as much as possible out of every processor cycle. I have been a developer for 15 years, but am pretty ignorant when it comes to hardware/OS level.
My alternative is using something like Titanium and Unity, (I don't think HTML5 will perform well enough), so I have a bit more platform flexibility, at the price of having to develop everything from scratch, and using technologies for the back-end stuff that aren't as ideal, (such as Android Java and/or Titanium JS). But perhaps that keeps more options open for me?
My feeling is I should go for Ubuntu, but the price needs to reliably reach considerably below $100 to become a nationwide system in a poor country.
Obviously any comments or thoughts on any aspect very gratefully received. Don't restrict your comments to my question - I want all your wisdom!
Huge thanks for reading all this and any contributions
Chris
Re-post
perhaps you would get more of a response if you made this a bit shorter, and re-posted on ubuntu.stackexchange.com, android.stackexchange.com, and programmers.stackexchange.com
Also, perhaps a little off topic, but have you considered using coffeescript? :cyclops:
Thought it might be a stackoverflow question, this forum is amazing for tablet OS dev though. As ever I blather on too much...people have complained in the past.
All three? Wouldn't that be bad netiquette?
That's kind of a tricky question because technology is always evolving and prices fluctuate so much that in a year you might be able to get a device for half the price. I'm not knowledgeable about the new Ubuntu options but if your gut says go Ubuntu, than do it.
Thanks, I'd love to say gut instinct served me well, and I'm all for intuition, but I wouldn't trust it enough to dedicate 6 months of development on its hunch. However these replies and the act of writing the question has crystallized my view a little so I now have more targeted questions.
found this excellent guide on the hassles of porting an OS to a new device...
http://www.linuxfordevices.com/c/a/Linux-For-Devices-Articles/Porting-Android-to-a-new-device/
and this
http://stackoverflow.com/questions/...droid-kernel-porting-from-one-device-to-other
So looks like a month, best case, with expert developers and a device well-known for being hackable. So to port ubuntu to a $40 device, I'm thinking three+ months, plenty of risk, and much pain.
Some other interesting posts:
A little gritty detail on porting kernels: http://stackoverflow.com/questions/...droid-kernel-porting-from-one-device-to-other
A tutorial on building (compiling, not developing) a kernel: http://forum.xda-developers.com/showthread.php?t=2110842
The best post I found on porting ROMS: http://forum.xda-developers.com/showthread.php?t=1941239
Porting modules from within kernels: http://forum.xda-developers.com/showthread.php?t=1751966
Porting cyogenmod, (a ROM, not a Mod, as anyone on this forum probably knows): http://forum.cyanogenmod.org/topic/15492-general-cyanogenmod-porting-discussion/
A new kernel developer: http://forum.xda-developers.com/showthread.php?t=2173411
A vocab for noobs like me: http://www.talkandroid.com/guides/beginner/android-rom-and-rooting-dictionary-for-beginners/
I was wrong. Ubuntu Touch is based on the CyanogenMod kernel, which is widely ported.
From Canonical's FAQ on the bits of CyanogenMod used: "The kernel and a few low level drivers for network, video, audio and some other hardware features are taken, all the higher level parts have been taken out. On top of this the whole Ubuntu is started in an chroot environment." ( https://wiki.ubuntu.com/Touch/FAQ#How_is_Ubuntu_Touch_connected_to_Android.3F)
As a result it has already been ported to about 40 devices, and porting to a further 30 is work in progress, listed here: https://wiki.ubuntu.com/Touch/Devices
CyanogenMod officially supports 172 devices, and unofficially supports another 59.
Officially supported devices: http://wiki.cyanogenmod.org/w/Devices
Unofficially supported devices: http://wiki.cyanogenmod.org/w/Unofficial_Ports