10 ms·

Are mobile web apps slow?

38 points by Unregistered 13y ago

ndnichols 13y ago
Q: "Are mobile web apps slow?" A: "A web app's CPU budget is spent in three different buckets" Q: "That... doesn't answer my question." A: "Javascript is the puppet master of the browser engine." Q: "That doesn't either." A: "I'm the director of engineering for the Web Platform team at Adobe. I sort of have to write blog posts like this." Q: "Ah, ok. No hard feelings then. You're in a tough spot." A: "Yes. Yes, I am."
- markdown 13y ago
  > Are mobile web apps slow? A meaningless question. Slow relative to what? Slow for whom? What sort of app are we talking about?
  - pistoriusp 13y ago
    You're able to infer the answer to all those questions.
  - CmonDev 13y ago
    Relative to native. Slow for users. A typical non-super-dynamic app. Use common sense.
- Roboprog 13y ago
  No, it doesn't answer the question, it suggests more questions to ask to make an informed answer. It's just a short cautionary tale to make the user consider what type of app he/she is building, and to consider if significant activity within the app is, or not, likely part of the GC/JavaScript (or HTML/CSS) that would suffer in some cases vs local, native, code. Personally, I tend to agree with the original article that GC is an expensive thing to put in most apps. BUT, you need to be pragmatic about determining if the additional margin of execution speed from not having a GC justifies the coding effort.
  - Roboprog 13y ago
    Having said that... you're right in that mostly this posting is to fulfill somebody's job requirements, rather than a substantial amount of content :-)
- zobzu 13y ago
  You mean "how'd that link get upvoted?" yeah..
DevKoala 13y ago
I also felt like he evaded the question he posed to himself.
thekingshorses 13y ago
Hacker news webapp for mobile - http://hn.premii.com http://hn.premii.com I wrote this client. It is as fast as most other native HN mobile apps, but mine works on iOS, Android browser, and windows phone 8. It also works on regular browsers. I have wrapped this in hybrid app using phonegap, but only submitted to Google play. One day I will submit it to Apple too. https://play.google.com/store/apps/details?id=com.premii.hn https://play.google.com/store/apps/details?id=com.premii.hn But it is not a simple task to make fast performing HTML5 app that works on all mobile browsers.
- gketuma 13y ago
  great app. Did you use any framework to build the app (I mean frameworks like sencha, enyos, etc)?
  - thekingshorses 13y ago
    No, I didn't use any framework, but I am using Zepto library. I may remove zepto in the future as all these libraries are very slow on mobile. When I replaced $.html() with dom .innerHTML, rendering time improved by 50% in some cases.
    - panabee 13y ago
      this is an awesome insight. any chance you could share more of these lessons?
    - nreece 13y ago
      Native fragment technique is even faster than innerHTML in Chrome/Firefox: http://jsperf.com/innertext-vs-fragment/19 http://jsperf.com/innertext-vs-fragment/19
- ux-app 13y ago
  Awesome work! Thanks for creating such a great site, I've added it to my iphone home screen. The performance is great even on an older iphone4 which is quite the achievement. This will definitely be the way I consume HN on the go from now on.
voltagex_ 13y ago
Yes. http://sealedabstract.com/rants/why-mobile-web-apps-are-slow/ http://sealedabstract.com/rants/why-mobile-web-apps-are-slow...
- lukifer 13y ago
  In fairness, the word should be "slower". Performance is always relative and/or subjective.
bsimpson 13y ago
Drew's post was very well explained, and relied heavily on concrete examples. This rebuttal was very much the opposite.
eonil 13y ago
This article written to advocate current direction of Adobe, but fails on everywhere by evading single question on title.
droidist2 13y ago
IMO it's not so much a question of performance, but of shaky browser support (especially on mobile) for things like real-time audio, video, vector graphics, and socket communication. Thankfully Adobe Flash / AIR really shine in these areas. It's sad that Adobe doesn't seem to realize it.
xiaomai 13y ago
This doesn't directly address web apps, but is a much more substantial rebuttal of some of the original article's claims: http://www.excelsior-usa.com/blog/java/5plus-garbage-collectors/ http://www.excelsior-usa.com/blog/java/5plus-garbage-collect...
akanet 13y ago
I (strongly dis)like how the article's titular question actually goes almost completely unanswered throughout the article. The article seems like a round defense of Javascript's productivity and interoperability benefits, but the original article (Why mobile apps are slow) had almost nothing to do with those issues. If you are going to respond to someone else's essay, please at least address the prompt!
weixiyen 13y ago
Yes, for the most part (relative to native apps).

38 points by Unregistered 13y ago

Q: "Are mobile web apps slow?" A: "A web app's CPU budget is spent in three different buckets" Q: "That... doesn't answer my question." A: "Javascript is the puppet master of the browser engine." Q: "That doesn't either." A: "I'm the director of engineering for the Web Platform team at Adobe. I sort of have to write blog posts like this." Q: "Ah, ok. No hard feelings then. You're in a tough spot." A: "Yes. Yes, I am."

markdown 13y ago
> Are mobile web apps slow? A meaningless question. Slow relative to what? Slow for whom? What sort of app are we talking about?
- pistoriusp 13y ago
  You're able to infer the answer to all those questions.
- CmonDev 13y ago
  Relative to native. Slow for users. A typical non-super-dynamic app. Use common sense.
Roboprog 13y ago
No, it doesn't answer the question, it suggests more questions to ask to make an informed answer. It's just a short cautionary tale to make the user consider what type of app he/she is building, and to consider if significant activity within the app is, or not, likely part of the GC/JavaScript (or HTML/CSS) that would suffer in some cases vs local, native, code. Personally, I tend to agree with the original article that GC is an expensive thing to put in most apps. BUT, you need to be pragmatic about determining if the additional margin of execution speed from not having a GC justifies the coding effort.
- Roboprog 13y ago
  Having said that... you're right in that mostly this posting is to fulfill somebody's job requirements, rather than a substantial amount of content :-)
zobzu 13y ago
You mean "how'd that link get upvoted?" yeah..

DevKoala 13y ago

I also felt like he evaded the question he posed to himself.

thekingshorses 13y ago

Hacker news webapp for mobile - http://hn.premii.com http://hn.premii.com I wrote this client. It is as fast as most other native HN mobile apps, but mine works on iOS, Android browser, and windows phone 8. It also works on regular browsers. I have wrapped this in hybrid app using phonegap, but only submitted to Google play. One day I will submit it to Apple too. https://play.google.com/store/apps/details?id=com.premii.hn https://play.google.com/store/apps/details?id=com.premii.hn But it is not a simple task to make fast performing HTML5 app that works on all mobile browsers.

gketuma 13y ago
great app. Did you use any framework to build the app (I mean frameworks like sencha, enyos, etc)?
- thekingshorses 13y ago
  No, I didn't use any framework, but I am using Zepto library. I may remove zepto in the future as all these libraries are very slow on mobile. When I replaced $.html() with dom .innerHTML, rendering time improved by 50% in some cases.
  - panabee 13y ago
    this is an awesome insight. any chance you could share more of these lessons?
  - nreece 13y ago
    Native fragment technique is even faster than innerHTML in Chrome/Firefox: http://jsperf.com/innertext-vs-fragment/19 http://jsperf.com/innertext-vs-fragment/19
ux-app 13y ago
Awesome work! Thanks for creating such a great site, I've added it to my iphone home screen. The performance is great even on an older iphone4 which is quite the achievement. This will definitely be the way I consume HN on the go from now on.

voltagex_ 13y ago

Yes. http://sealedabstract.com/rants/why-mobile-web-apps-are-slow/ http://sealedabstract.com/rants/why-mobile-web-apps-are-slow...

lukifer 13y ago
In fairness, the word should be "slower". Performance is always relative and/or subjective.

bsimpson 13y ago

Drew's post was very well explained, and relied heavily on concrete examples. This rebuttal was very much the opposite.

eonil 13y ago

This article written to advocate current direction of Adobe, but fails on everywhere by evading single question on title.

droidist2 13y ago

IMO it's not so much a question of performance, but of shaky browser support (especially on mobile) for things like real-time audio, video, vector graphics, and socket communication. Thankfully Adobe Flash / AIR really shine in these areas. It's sad that Adobe doesn't seem to realize it.

xiaomai 13y ago

This doesn't directly address web apps, but is a much more substantial rebuttal of some of the original article's claims: http://www.excelsior-usa.com/blog/java/5plus-garbage-collectors/ http://www.excelsior-usa.com/blog/java/5plus-garbage-collect...

akanet 13y ago

I (strongly dis)like how the article's titular question actually goes almost completely unanswered throughout the article. The article seems like a round defense of Javascript's productivity and interoperability benefits, but the original article (Why mobile apps are slow) had almost nothing to do with those issues. If you are going to respond to someone else's essay, please at least address the prompt!

weixiyen 13y ago

Yes, for the most part (relative to native apps).