What knowledge should every app developer have about Android.

  • In today’s quick paced versatile business sector, purchasers have no tolerance for portable applications that bargain their experience. “Crashes” and “Not working” are the most widely recognized input on Google Play for unsteady or drowsy applications (counting diversions). Those remarks and evaluations make a huge number of potential downloaders avoid those lousy applications. Sounds cruel, yet that is how it is.
  • An application succeeds not by possibility. It is the right’s consequence choices made at the correct time. The best versatile application engineers comprehend the significance of execution, quality and vigor over the variety of cell phones that their clients use. Samples proliferate of exactly how effortlessly an engineer can turn out badly, with their application.
  • An application can act distinctively on an assortment cell phones, even ones running the same OS variant and indistinguishable equipment segments.
  1. Amid Q1 of this current year (1 January to 31 March), we assembled a lot of information from Testdroid Cloud on tests done by numerous portable application and diversion designers.
  2. Testdroid Cloud is an online cloud-based stage for versatile engineers to test that their applications work superbly on the gadgets that their clients use.
  3. Amid this period, more than 17.7 million tests were keep running on 288 particular Android equipment models.
  4. To be clear, diverse variants of some well known models were tried yet are considered in the information one unmistakable gadget, (for example, the Samsung Galaxy S4 GT-i9505 running 4.2.2, API level 17).
  5. Some prominent gadgets likewise had diverse variants of a solitary OS, for example, the Google Nexus 7 ME370T with Android OS form 4.1.2 (API level 16), 4.2.2 (API level 17) and 4.3 (API level 18).
  6. All tests were mechanized, utilizing standard Android instrumentation and distinctive test-mechanization structures. On the off chance that you are not acquainted with instrumentation, Android has an instructional exercise that clarifies fundamental test robotization.
  7. Likewise, the tests got issues through logs, screenshots, execution investigation, and the achievement disappointment rate of test runs.
  • Note: The information incorporates all test results, from the most punctual stage (=APK prepared) to when the application gets “concluded.” Therefore, it incorporates the accurate issues that designers experienced amid this procedure.
  1. The objective for this exploration was to distinguish the most widely recognized issues and difficulties that Android designers face with the gadgets they manufacture for.
  2. The 288 novel Android gadget models speak to a huge volume of Android use: roughly 92 to 97% of worldwide Android volumes, contingent upon how it gets measured and what locales and markets are incorporated.
  3. This exploration speaks to amazing scope of Android use all around, and it demonstrates the most evident issues and also the status of Android equipment and programming from a designer’s perspective.
  4. We’ll plunge profound into two regions of the exploration: Android programming and equipment.
  5. The product segment concentrates on OS renditions and OEM customizations, and the equipment area separates to the segment level (showcase, memory, chipset, and so forth.).
  • Android Software Link
  1. Your application is programming, however a considerable measure of other programming is included in cell phones, as well.
  2. What’s more, this “other” programming can make your product perform in an unexpected way (read “off-base”).
  3. What would you be able to do to verify your product functions admirably with whatever is left of the product in gadgets today?
  4. We should first take a gander at the absolute most basic programming issues experienced by application engineers.
  5. Android OS has been rebuked for stage fracture, making things extremely troublesome for designers, clients and essentially every player in the biological community to keep up.
  6. Nonetheless, that is not precisely the situation. The OS seldom makes things break independent from anyone else — particularly for application engineers.
  7. All the more regularly it additionally includes OEM overhauls, equipment and numerous different components — notwithstanding the OS.
  8. One motivation behind why Android is colossally well known among versatile fans and has immediately jumped in front of Apple’s iOS in piece of the overall industry is on account of it comes in gadgets of all shapes, costs and structures, from many distinctive OEMs.
  9. Additionally, the infrastructural programming of Android is strong, furnishing an astounding stage with a surely understood Linux portion, middleware and other programming on top.
  10. In the accompanying areas, we’ll take a gander at the aftereffects of our examination, separated by OS rendition, OEM customizations and programming conditions.
  • Instructions to READ THE GRAPHS 
  1. The discoveries from this study are plotted on charts. In those diagrams, the dull dark line is the middle (half) of the disappointment rate of distinctive gadgets in that gathering.
  2. The lines over the bars check the upper quartile (75%), and the lines beneath imprint the lower quartile (25%). The dashed lines are the most extreme (top or right) and least (base or left). The circles speak to anomalies.
  • OS And Platform Versions 
  1. We will begin with the most vital component in the issues experienced by application engineers: Android OS adaptations.
  2. This significantly influences which APIs engineers can utilize and how those APIs are upheld in assortment of gadgets.
  3. Numerous engineers have encountered the greater part of the immense new components, upgrades and increases that Android has brought form by adaptation.
  4. Some late-comers have played with just the most recent adaptations, however some have been creating applications since the good ‘ol days as days of Cupcake (1.5) and Doughnut (1.6). Obviously, these adaptations are not important any longer.
  5. The accompanying table demonstrates the discharge dates for the different OS variants (alongside their code names) and notes why certain renditions were prohibited or were not utilized on the gadgets being tried.
  6. With regards to the most recent variants of Android — and how about we number Ice Cream Sandwich (4.0.3 to 4.0.4), Jelly Bean (4.1.1 to 4.3) and Kit Kat (4.4 to 4.4.2) among them — Ice Cream Sandwich is unmistakably the most vigorous stage. OEM upgrades were clearly the minimum dangerous on this form, and the larger part of tried applications worked exceptionally well on API level 15 (4.0.3 to 4.0.4).
  7. The move up to Jelly Bean and API level 16 didn’t essentially present new incongruently issues, and the middle disappointment rate remained generally low.
  8. Be that as it may, API level 16 had numerous exception cases, and for good reasons. For example, more issues were accounted for with Vsync, extendable warnings and particularly with backing for the lock screen and home screen turn.
  9. Programming interface level 17 conveyed changes to the lock screen, and this form was for the most part stabile up to form 4.2.2.
  10. The disappointment rate went up when the lion’s share of OEMs acquainted their redesigns with this variant. Evidently, it turned out to be more dangerous for clients than past forms.
  11. Maybe most shockingly, the disappointment rate went up when Kit Kat API level 19 was discharged.
  12. The normal disappointment rate came to almost the same level as it was with Gingerbread.
  13. Google fixed Kit Kat rapidly with two discharges (4.4.1 and 4.4.2). Of those, 4.4.2 appeared to live any longer, and after that the 4.4.3 upgrade turned out more than a large portion of a year later.
  1. By and large, 23% of applications carry on contrastingly when another adaptation is introduced.
  2. The middle blunder percent was the littlest with Ice Cream Sandwich (1%), then Jelly Bean (8%), Honeycomb (8%), Kit Kat (21%) and, at last, Gingerbread (30%).
  3. This is really lower than what was found in the study directed with Q4 information (a 35% disappointment rate).
  4. In spite of old Android variants being utilized almost no and Gingerbread being the most seasoned effectively being used, a few applications (40% of those tried) still take a shot at even more established renditions (beneath 2.2).
  5. At the end of the day, if an application takes a shot at adaptation 2.2, then it will work 40% of the time in even more seasoned variants too.
  6. More than half of Android OS overhauls presented issues that made applications fall flat in testing.
  7. Testing fizzled 68% of the time when 5 applications were haphazardly chosen of 100.
  8. The normal length of time of testing was 57 cycles for each stage. Old variants were tried not exactly new ones: Gingerbread (12 test cycles), Ice Cream Sandwich (17), Jelly Bean (58) and Kit Kat (95).
  9. A normal testing cycle diminished 1.75% of bugs in the code general.
  • Note: A test cycle constitutes an emphasis of a specific test script executed in one application adaptation. At the point when an application is changed and the test continues as before, that is considered another test cycle.
  1. For maximal scope either topographically or all inclusive, utilizing however many physical gadgets as could be expected under the circumstances is suggested.
  2. Track your intended interest group’s utilization of diverse OS forms. The worldwide status of renditions is accessible on Google’s dashboard.
  3. All OS renditions above 2.3.3 are still applicable.
  4. This won’t likely change soon on the grounds that clients of Gingerbread and Ice Cream Sandwich speak to almost one quarter of all Android clients, and a considerable lot of them don’t upgrade (or would have done as such as of now).
  5. On the off chance that you need to cover 66% of OS volume, then testing with Kit Kat (4.4.x) and Jelly Bean (4.1 to 4.3) is sufficient (covering API 16 to 19).
  6. To cover 75% of OS volumes, then test from rendition 4.0.3 (API level 15).
  7. We prescribe testing the accompanying gadgets to augment scope:
  8. Amazon Kindle Fire D01400 — 2.3.4
  9. HTC Desire HD A9191 — 2.3.5
  10. Huawei Fusion 2 U8665 — 2.3.6
  11. Sony Xperia U ST25i — 2.3.7
  12. Asus Eee Pad Transformer TF101 — 4.0.3
  13. LG Lucid 4G — 4.0.4
  14. HTC One S X520e — 4.1.1
  15. Motorola Droid XYBOARD 10.1 MX617 4.1.2
  16. Acer Iconia B1-A71 — 4.2
  17. BQ Aquaris 5 HD — 4.2.1
  18. HTC One scaled down M4 — 4.2.2
  19. Samsung Galaxy Note II GT-N7100 — 4.3
  20. LG Google Nexus 5 D821 — 4.4
  21. HTC One M8 — 4.4.2
  • Note: These gadgets were chosen in light of the fact that they are a decent base to test certain stage variants, with diverse OEM customizations included. These gadgets are not the most dangerous; rather, they were chosen in light of the fact that they give incredible scope and are illustrative of comparable gadgets (with the same form OS, from the same maker, and so on.).
  • OEM Customization
  1. One hindrance with Android — like any open-source venture — is its adaptability, which uncovered the whole stage to issues.
  2. What is called “fracture” by engineers would be viewed as a state of separation for OEMs.
  3. As of late, all Android OEMs have anxiously constructed their own particular UI layers, skins and other middleware on top of vanilla Android.
  4. This is a noteworthy wellspring of the fracture that influences designers.
  5. Notwithstanding UI layers, numerous OEMs have presented legacy programming — custom-made to Android — and it, as well, is keeping designers from building their applications to work indistinguishably crosswise over diverse brands of telephones.
  6. Drivers likewise cause real issues, numerous identified with illustrations.
  7. Certain chipset makers have made a particularly terrible showing with redesigning their illustrations drivers, which makes the hues in applications, diversions and any realistic substance conflicting crosswise over telephones.
  8. Engineers may experience totally distinctive shading plans on different Android.
  1. Nothing unexpected, Samsung gadgets are among the most hearty and the most dangerous.
  2. For instance, Galaxy Nexus GT-I9250 is a standout amongst the most hearty gadgets in all classifications, while the Samsung Infuse 4G SGH-I997 fizzled the most in those same classifications.
  • Asus gadgets, alongside Xiaomi gadgets, are the most hearty. Xiaomi actualizes Android in an unexpected way, in any case; for occasion, pop-ups make the controllability of a few gadgets unthinkable.
  • Coolpad has, by volume, the most issues. Among the greatest brands, HTC has the minimum mistake inclined gadgets.
  • The majority of the enormous brands — HTC, Samsung, Sony and LG — have customizations that are risky for specific sorts of utilizations. For instance, Sony’s customizations breaks some essential usefulness, for example, seeing PDFs. Samsung’s customizations has issues with bringing photographs with the camera and intruding on calls.
  1. 33% of applications incorporate with no less than three or more online networking stages.
  2. 48% of applications incorporate with no less than two online networking stages.
  3. 65% of applications incorporate with no less than one online networking stage.
  4. 92% of applications incorporate with the offered stage to show promotions.
  1. Check whether the product that your application relies on upon is introduced on all gadgets.
  2. Try not to accept that those outsider applications and other programming exist on each gadget!
  • Android Hardware
  1. The Android gadget eco-framework keeps on developing and advance.
  2. Numerous handset producers keep on producing gadgets with astounding details and equipment and with diverse structure components.
  3. Normally, the sheer number of conceivable gadget arrangements shows a test to designers.
  4. Guaranteeing that an application functions admirably on the amplest scope of gadgets is pivotal and is the most straightforward approach to abstain from baffling end clients.
  5. Most engineer precisely measure the advantages and disadvantages of testing on emulators and testing on genuine gadgets to take after the right system.
  6. Ordinarily, emulators are utilized as a part of starting phases of advancement, while genuine gadgets are gotten later in the diversion.
  7. Your decision of stage on which to fabricate your next huge thing ought to be as legit as could be expected under the circumstances — from the very first moment.
  8. As far as we can tell, that is a foundation of making an effective application — and picking up those a huge number of downloads.
  1. The way to accomplishment with any application — particularly recreations — is to get the UI and illustrations right.
  2. This is a test in light of the fact that there are many resolutions, a considerable measure of approaches to present substance and, normally, a variety of equipment.
  3. With new top of the line gadgets picking up fame among clients, the Android eco-framework is by all accounts immediately headed towards high-determination shows.
  4. High-determination screens clearly improve for a client experience, yet to exploit this, designers need to redesign their applications.
  1. Anything can happen when an Android gadget comes up short on memory.
  2. Things being what they are, how does your application work when a gadget is low on memory?
  3. Engineers manage this issue rather regularly.
  4. Numerous applications won’t keep running on certain Android gadgets on the grounds that they devour an excess of memory.
  5. Normally, the most prevalent applications — ones evaluated with four and five stars — don’t have this issue in light of the fact that memory administration has been actualized the right way, or else they have been barred from being downloaded on low-end gadgets altogether.
  6. An excess of today’s applications were initially created for top of the line gadgets and can’t be keep running on low-end gadgets. What is clear is that you can without much of a stretch handle this issue by checking how your application functions with distinctive memory limits.
  7. Memory appears to essentially influence how a gadget adapts to an application.
  8. On the off chance that a gadget’s RAM is equivalent to or under 512 MB, then the gadget will crash 40% of the time when certain applications are running.
  1. The middle of blunder was least with RAM of 1024 MB (1%), then 1536 MB (3%) and next 768 MB (16%).
  2. It was most elevated with RAM of 336 (45%) and 168 MB (44%).
  3. Roughly 10% of applications keep running on gadgets with more than 512 MB crash due memory-related issues.
  4. Numerous OEMs don’t give their gadgets the 512 MB of RAM that Google suggests as the base. Such gadgets are 87% more prone to have issues than gadgets with more memory.
  5. The likelihood of disappointment drops 41% for gadgets that contain more than 576 MB of memory.


  1. The distinction in execution between silicon chipsets (CPU, GPU, and so on.) is really stunning.
  2. This is not as a matter of course evident to end clients.
  3. A few individuals give careful consideration to the CPU’s clock rate, as opposed to the chipset and different components that influence the gadget’s execution.
  4. Envision creating something that is focused at top of the line gadgets yet that keeps running on low-end equipment. It essentially wouldn’t work.
  5. The client experience would clearly endure on the grounds that a top of the line application running on a low-end chipset with a low clock-recurrence rate would endure in execution.
  6. Numerous applications endure seriously in light of the fact that the exercises on screen are matched up with the clock, and the UI can’t revive rapidly enough to stay aware of it. For clients, this interprets into inadequately actualized illustrations, flickering screens and general gradualness.
  7. Not at all like the CPU construction modeling in chipsets — which is fabricated essentially by ARM — the representation part is produced by various sellers, which gives some semiconductor organizations the adaptability to pick and pick which GPU runs best with the CPU in their chipsets.
  8. Once upon a time, the essential employment of the design card was to render top of the line graphical substance and 3-D pictures on the screen.
  9. Today, GPUs are utilized for a great deal more than diversions and are as pivotal as the CPU, if not all the more so.
  10. Today, the late’s majority Android OS forms — Ice Cream Sandwich, Jelly Bean, Kit Kat et cetera — depend intensely on the GPU on the grounds that the interface and all activitys are rendered on it, which is the manner by which you’re ready to accomplish move impacts that are rich smooth.
  11. Today’s gadgets have numerous more GPU centers than CPU centers, so all illustrations and rendering-concentrated undertakings are taken care of by them.


  • Single center
  • The middle of mistake was most minimal in Intel Atom Z2420 (0.2%), then in Qualcomm Snapdragon S2 MSM8255T (1.1%).
  • Double center
  • The middle of mistake was most minimal in MediaTek MT8317 (0.3%), then in Intel Atom Z2560 (0.4%).
  • Quad center
  1. The middle of mistake was most minimal in both MediaTek MT8125 (0.2%) and Qualcomm Snapdragon 600 APQ8064AB and APQ8064T (0.2%).
  2. Some top of the line gadgets are upgraded to the most recent adaptation of Android and get the most recent OEM redesigns, which makes them more powerless against issues (67%) than low-end gadgets with the first OS variant.
  3. Clock rate doesn’t straightforwardly correspond with execution. A few benchmarks show huge change in execution, notwithstanding when the client experience and the applications on those chipsets and gadgets don’t enhance fundamentally.
  4. Power utilization is actually a more concerning issue in a few batteries, and a few gadgets come up short on battery life rapidly.
  • Conclusion
  1. In this article, we’ve concentrated on two ranges — Android programming and equipment — and what application designers need to think about both.
  2. The Android eco-framework is continually changing and new gadgets, OS forms and equipment turn out each week.
  3. This normally exhibits a major test to application designers.
  4. Testing on genuine gadgets before dispatch will make your application fundamentally more hearty.
  5. In the following article, we will plunge profound into this theme, covering the most fundamental testing strategies, structures, gadgets and other base required to boost test.

About Author

Annyesha Bakly

Annyesha Bakly

Ms.Annyesha Bakly has rich experience in building eCommerce and Rich Internet Applications and has helped build some of the leading eCommerce websites. Besides eCommerce, she also has a passion for Usability, Human Computer Interaction and Data Visualization.

For business enquiry, please contact us

© Copyright 2013 Veltrod Scroll Top