Real-world web-app development is much different than the ideal of HTML5.
jQuery? Just including that puts your mobile web-app at 86KB of JavaScript and you haven't even written a single line of your app's code. Rendering on the client-side with jQuery templates is not anywhere near practical for the majority of web-enabled phones – if not all of them except the iPhone. More than 100-150KB of JS and most phones will enter sluggish territory.
Something you'll notice reading about HTML5 is most books and online articles are written by people who have never built or released web-app to be run on phones less powerful than an iPhone 4. They picked up an HTML5 book and regurgitated everything and every new nifty feature into their own words and examples.
My advice?
Don't use jQuery or any framework. Use native JS and only the functions you need, don't use CSS3, keep the DOM structure as simple as possible, and test from day one. My entire web-app's JS is 60KB minified, because I don't rely on framework's like jQuery (86KB) or Sencha Touch (250KB). These sizes don't matter on a desktop computer, but on an HTC from last year running Android 2.3, size greatly affects performance.
I agree that it is important that you keep your application small. However, using the application cache like the author suggests means that you will never have to download the specified content, like jQuery, again. Additionally, since you have 5MB of storage, I really don't think the 31KB of jQuery really matters.
I agree that client-side rendering is not optimal, but for different reasons. This isn't 2006 anymore, phones are more powerful now. Unless I absolutely have to, I'm not going to care about your Nokia from back in the Stone Age. My problem with client-side rendering is that it chokes on large data-sets. You also have to pay the penalty to download the dynamic content after you download the static content, but this can be mitigated by only updating the content, rather than using AJAX to load it all to begin with. That also kills the author's use of the plug-in that stores AJAX data locally, but that's a separate issue.
My advice:
Use jQuery, jQuery Mobile, and AppCache. When combined, jQuery and jQuery Mobile (including CSS) take up 58KB of space. Oh, and since you're storing it all locally from now on, there's no need to cry about the time (less than 1s on 3g) it takes to download all those fancy JS files.
This is exactly why I run my text editors in Windows 95 still. There is far less code than current versions of windows and I find a huge performance hit when using these newer versions.
Ok, I'm getting ready to write an HTML5 app for cross-platform use.
I grok what you're selling, but looking at JQuery, and especially Sencha Touch, I'm asking myself: how the heck could I make a native-looking app in plain JS without having to dive deep and beating the hell out of myself with all this crap that's already been solved?
Are you saying that the extra UI snazzle just isn't worth it? Or to buck up and start trying to replicate huge pieces of frameworks on my own?
It's not that making native-looking apps in plain JS isn't worth it – it's not possible unless you're only targeting iPhone.
There may be a place for jQuery, but stay away from Sencha Touch. It's a huge library, and barely runs on any device but an iphone. Don't be fooled by their slick marketing and examples.
Essentially, trying to replicate a native app in JS is a losing battle, unless your app is extremely simplistic. The majority of smartphones are simply too underpowered to replicate a native experience.
Aim for a site like http://m.twitter.com instead of trying to replicate a native experience. Test on real devices you intend to support as soon as possible, even if you've barely written the app.
Except for testing most this advice is awful. jQuery is 31KB not 86KB. CSS3 is or has plans to be hardware accelerated on all major platforms.
Also what exactly is your claim on how size is greatly affecting performance? That Android takes a long time to deflate a large zipped javascript file or that it takes a long time to parse a large javascript file or the large javascript libraries are ineffecient? Do you have any evidence on these things?
jQuery 1.6.2 is 94KB minified. Gzip compression is not applicable here because the code being executed is not 31KB, that's just the download size.
The amount of JS that has to be interpreted by the browser does noticeably affect performance. At least in my experience testing and building webapps for mobile phones, over 100KB of JS and many devices start to manifest performance issues. I find it interesting so many people here question JS size affecting performance, when this has been the main focus and selling point of browser development for the last year.
CSS3 animations and gradients are not really feasable for anything but the iPhone at this point. On HTC devices running Android 2.3, just adding a CSS3 gradient to each item in a list slows my scrolling speed significantly and introduces stutters in the scroll, versus using a PNG of the same gradient I'm using CSS for. This is just one example of many, but CSS3 support is very shoddy on devices outside of the latest iPhones.
I'm curious what browsers you're supporting and experience you're aiming for with your hand rolled javascript.
Targeting phones outside of iOS and Android with anything but the most basic functionality is a difficult task without help. A prime example being the 500+ ms delay created when adding event listeners for click instead of touch events. Click covers everything but it creates a perceptibly slow ui for the user.
Browsers supported are IE6+, IE Mobile, Opera Mini, Safari 3+, Chrome, Firefox 3+, blackberry browser, android browser, symbian's web browser, pretty much everything that renders the important parts of CSS1/2 and has JS support.
The click delay workaround is like 15 lines of JS. It's one of the few issues (along with foreach) requiring a major workaround, if you design your HTML/CSS layouts properly. Also, don't forget you can take what you need from jQuery or other libraries. For example, the click delay workaround I think I grabbed from HTML5 mobile boilerplate, and my foreach function I grabbed from jQuery.
Asking from nearly complete ignorance here but: with a JS compiler/optimizer like Google's Closure [1] wouldn't it be possible to still use jQuery but have all the unnecessary code that's not used striped out?
That might be the case. This blog post [1] shows some examples of what the advance optimizations do and it seems that, while it'll shorten everithing up, things that are put in the global namespace (window) will not be stripped even when they aren't called. As jQuery adds the 'jQuery' and '$' variable to the global namespace and all it's functions are attached to that object, I would assume that the Closure compile connot remove any of them.
Of course, this is all pure speculation, it'd be much better to test this out.
Closure can definitely remove functions that are never called - but I wonder if its static analysis can figure out jQuery style code. My guess is that it'll do a decent job, but jQuery would need many changes to get it to compile.
All of your points are well taken. The purpose of this article was to collect and explain a number of things I wish I knew when I started working with mobile HTML5. I stated in the article that the topics are basic ones, tho I think valuable, I fully agree they are not ment for a serious scale application.
One note of caution: Rendering on the client side using jQuery Templates as the author suggests is great for small amounts of content, but even a list with 40 items will nearly freeze the UI while its rendered. I've experienced this on both iPhone and Android.
Also, using a cache manifest will cache your AJAX/JSONP calls. Either explicitly list these URLs in the network section of your manifest or add a timestamp to each request so its unique.
Was this a straight HTML list or were you using a jQueryMobile listview to pretty it up? When we've used templates to render a list (using either underscore or ejs) it was taking 10-20ms to generate the initial HTML but around 10-15x as long to call listview on it.
I was using jQuery Mobile. The listview framework is a big part of the performance problem.
Needless to say, the best advice I would give (that others have said here as well) is to avoid frameworks like jQuery Mobile. There many other reasons too...not least of which is that jQuery Mobile doesn't work on T-Mobile or O2 in the UK and parts of Europe. It appears these networks intercept all traffic and try to minify the JS and CSS before they deliver it to the phone. There is some part of jQuery Mobile that they choke on and the page doesn't render correctly. I've tested extensively using Perfecto Mobile and even the jQuery Docs & Demo pages don't work right (so I know its not just me).
Thanks for this info. I wasn't aware of potential carrier-specific issues so I'll make sure we take this in to account while testing. Have you raised this with the networks? I guess there's a chance that it could also affect non-jqm powered code.
I recently visited a site from HN on my phone, but its text was a bit too small for me to read, so I tried zooming in. To my surprise, it didn't zoom at all! From this article, the website must have set user-scalable=no.
I don't really see the point of this option. The website can (presumably) scale to many different devices already, surely it can then scale with zooming as well. I'd think you could use percentages to create any kind of 'static' ui, if you really wished to do so.
As annoying as it is, it makes sense that some web developers would choose this option. Although, limiting the experience for everyone just because one mobile browser has a bug? I'm pretty sure something similar happened with a certain browser called ie6. From the link you gave, it looks like you can just detect an iPad and disable zooming in that case (although, that requires some javascript or server side programming).
So, which version of iOS do they plan on fixing this (or did fix this)? But then, that doesn't factor the people who don't know about OS updates and never sync.
Is this really the way forward? Why the assumption that the only reason I might zoom in is 'accidental'?
Also I am uneasy about populating all content via JSON but I can't put my finger on the precise technical reasons (but then neither could Gawker wrt desktop browsers so they went ahead with it anyway)
I use the HTML5 app cache. I have tens of millions of users on this in-production product. I only really feel safe doing this because I have a custom web browser, and as I only target jailbroken devices I can hack at WebKit a little to make this work.
When I started using it, it seemed awesome. Then I played with it a little, and it started failing. Given that I have friends who actually give talks at conferences on the HTML5 application cache and how cool it was, this didn't make any sense: I believed I must be doing it wrong.
Then, one day, I decided to read the source code for WebKit's implementation of the HTML5 application cache, and found it to have tons of race conditions that would lead to whole-browser crash conditions (which I was even experiencing in my app). I even found a bug on the bug tracker with my most-occurred bug on it.
Only, that bug had been filed only months before, and this feature of WebKit had been "in production", with these bugs, for a couple years. My only guess is that no one really uses this feature very often.
Reading into the bug tracker some (to find more bugs), staring at the commit messages for the previous year, and even reading some of the posts on the mailing list, I found choice quotes from the people maintaining it that explained that no one knew how this code works, and that the one guy who did forgot.
In iOS 4.3 (I believe it was only 4.3, not 4.2, but I'd sort of believe 4.2, /maybe/ even 4.1), the crash bugs have been "fixed" (or I'd have given them to comex to maybe build another jailbreakme.com), but the race conditions remain: the code doesn't actually /work/, it just sort of doesn't fail.
Specifically, the issue is that the application cache loader for a page is tied to a specific frame: the one that started loading the cache. If that frame is freed, such as by the simple act of the user hitting the reload button, while the download is occurring, the entire thing gets wedged, and you can no longer get any parts of your page cached, nor can you recover or fix the process without the user totally closing the web browser and reopening it.
In practice, this actually happens constantly: if the user has a reasonably slow connection, and things are taking a while in the background to cache, they will hit reload. This deletes the old frame, but associates a new frame with the same cache builder. When the current download stops, the frame it was using is gone, and it ends up in this horrible wedge state.
Ok, I no longer even remember why I'm bothering to type this here (does anyone care? does anyone actually intend to try to use this feature? ;P), but I also no longer seem to have anything more to say, so I'm going to click "add comment" ;P.
I've experienced this too and no longer use the app cache either. I use the browser cache and that means that yes, the user must have network connectivity to load the page, but at least it always works and the cache doesn't get into a jacked up state.
In addition to powering full blown web apps, web technologies can also be used from within web views to create hybrid native/web apps.
I have a post on idevblogaday.com today on calling Objective-C from JavaScript, calling JavaScript from Objective-C, and logging to the Xcode console from JavaScript.
jQuery? Just including that puts your mobile web-app at 86KB of JavaScript and you haven't even written a single line of your app's code. Rendering on the client-side with jQuery templates is not anywhere near practical for the majority of web-enabled phones – if not all of them except the iPhone. More than 100-150KB of JS and most phones will enter sluggish territory.
Something you'll notice reading about HTML5 is most books and online articles are written by people who have never built or released web-app to be run on phones less powerful than an iPhone 4. They picked up an HTML5 book and regurgitated everything and every new nifty feature into their own words and examples.
My advice?
Don't use jQuery or any framework. Use native JS and only the functions you need, don't use CSS3, keep the DOM structure as simple as possible, and test from day one. My entire web-app's JS is 60KB minified, because I don't rely on framework's like jQuery (86KB) or Sencha Touch (250KB). These sizes don't matter on a desktop computer, but on an HTC from last year running Android 2.3, size greatly affects performance.