DNK Jump In Blog - Latest Comments

Re: Going Faster: Sexy numbers and full post-mortem

Drew Peterson — Thu, 26 Jan 2012 11:03:35 -0000

Thanks!
As for the Firefox lag, I'll look into it. We could have a poorly performing jQuery selector or something.

Re: Going Faster: Sexy numbers and full post-mortem

Guest — Thu, 26 Jan 2012 11:01:57 -0000

Thanks!
As for the Firefox lag, I'll look into it. We could have a poorly performing jQuery selector or something.

Re: Going Faster: Sexy numbers and full post-mortem

$18045708 — Wed, 25 Jan 2012 16:09:21 -0000

Great work, Drew! I'm noticing a significant speed-up here in Britain.

It's still lags slightly on Firefox sometimes, but that's much more bearable. (And it could be my side anyway.)

Re: Going Faster: How not to suck at serving web content

$18045708 — Thu, 12 Jan 2012 17:26:25 -0000

Excellent work. The site really is noticeably snappier for me in the UK now! :-)

Re: Going Faster: How not to suck at serving web content

Drew Peterson — Sat, 07 Jan 2012 19:21:34 -0000

A ended up putting several hours into fixing these problems (some of it due to code cruftiness), and some of it due to finding even more issues once I started digging.

I'm going to write a full postmortem once the fixes go live, but to answer your question we're using https://github.com/jethered... to compress/combine our javascript and css files and the gzip/proxy fixes were just changes to the web.config file.

Re: Going Faster: How not to suck at serving web content

Bret Ferrier — Thu, 22 Dec 2011 11:17:26 -0000

This is a classic list of things that you are currently doing wrong. I would be interested in hearing how you are fixing some of them. Like how are you enabling gzip and how are you leveraging proxy caching. Just curious if you found any tools to help solve these issues or...

Re: Going Faster: How not to suck at serving web content

Vijay Thirugnanam — Thu, 22 Dec 2011 10:16:31 -0000

Nice points on improving performance. Optimizing Scripts + CSS will definitely improve performance quite a bit.

Re: HTML5 Takes on the King of Plugins

Vijay Thirugnanam — Thu, 20 Oct 2011 12:40:25 -0000

Silverlight is good, especially v4. The reason Silverlight does not get a mention is that the earlier versions, esp Silverlight 2, had a lot of shortcomings.

Even in v4, using RIA services with MVVM, is somewhat of a pain. I am sure if the same amount of effort goes to the next version of Silverlight, some of this can be addressed.

Re: HTML5 Takes on the King of Plugins

CodingBandit — Wed, 19 Oct 2011 11:24:42 -0000

Agreed that tooling has a long way to go. But the opportunities to reuse said animations will be across all platforms, not just limited to a Flash audience-only.

Re: HTML5 Takes on the King of Plugins

CodingBandit — Tue, 18 Oct 2011 17:43:25 -0000

There will be need for Flash developers for some time to come.It would also be good for Flash Developers to look into learning HTML5 as well, potentially using Adobe Edge as an IDE. HTML5 is going to be a very marketable skill set.

Re: HTML5 Takes on the King of Plugins

Vijay Thirugnanam — Tue, 18 Oct 2011 12:14:18 -0000

HTML5 creates animates which are Flash like. But, there are still no standard tools available from vendors which can produce the standard set of animations that Flash does. Maybe, I am not aware of those tools :)

Re: HTML5 Takes on the King of Plugins

Kiruba Sankars — Mon, 17 Oct 2011 23:48:23 -0000

What about adobe flash developer?

Re: HTML5 Takes on the King of Plugins

nick — Mon, 17 Oct 2011 18:48:01 -0000

More like, network engineers are not programmers and should beware about job security. There will always be viruses and malware, but HTML5 might be hackable someday too.

Re: HTML5 Takes on the King of Plugins

nick — Mon, 17 Oct 2011 18:45:09 -0000

Adobe Edge is top notch from what I am seeing. Visual Studio 2010 comparably has the least favorite css compliant styling component I have seen, but that will change then the standards are 100%.

As for HTML5 vs Flash, it depends which platform creates the best mix with Jquery and frame per second. Once that there are options, I would still go with the one that is free.

Re: HTML5 Takes on the King of Plugins

CodingBandit — Mon, 17 Oct 2011 18:17:02 -0000

Alex - agreed Silverlight is an amazing technology, I see it's future predominantly on the Windows platform. The discussion above is targetting presenting content on the web specifically.

Unfortunately, like Flash, being plug-in based in the browser, it is also subject to some of the realities mentioned above with regards to Reach.

IMO, it really comes down to use case, if you want to target more devices, on the web, with more reach, you'd be better off to use HTML5. If you want to target the Windows desktop specifically, then target Silverlight or WPF, if you are looking for a Windows 8 Metro/tablet UI, you can target XAML or HTML5 on WinRT.

Re: HTML5 Takes on the King of Plugins

daris2000 — Mon, 17 Oct 2011 17:42:06 -0000

Thanks for talking back, great discussion here. To be clear, I'm not into bashing any technology, nor did I think you were.

First, I'd like to say that Flash is technically superior to HTML5. Silverlight as well. I fully agree. But the "reach" argument is the deciding factor over time. So yes, there are still use cases for Flash right now.

As for the "safe bet" for businesses, I only partly agree. I work for a huge company and two web apps based on Silverlight were rewritten from the ground up because some important people demanded it to work on iPads. That's a few 100K of lost capital. This influx of devices isn't coming, it has happened already. Therefore, I hardly consider it "safe" to create anything new that has a lifecycle of a few years in Flash right now. In specific cases it may still be the best choice, but safe, no, not really. Again, I cannot emphasize this enough...the revolution isn't on the horizon, we are in the middle of it.

On vendor extensions: in your CSS, you include the vendor-extension property, followed by the official one. That's it. As you upgrade your browser, at one point the vendor-extension will be ignored and the official one will be used, thereby the CSS is future-proof.
That doesn't mean there's no issues though. The most important one being the extreme differences in the performance of running these animations, as well as the age-old problem of users on IE6-8. So I agree, CSS animations are not mature enough if you demand exact behavior across all browsers. Increasingly though, app builders are letting go of that thought to serve all users equally, even if they are on a decade old browser.

Re: HTML5 Takes on the King of Plugins

$18045708 — Mon, 17 Oct 2011 17:28:29 -0000

Well done Vijay for at least bringing up Silverlight...! I mean, this is DNK and no mention of the .NET rival to Flash heh? :-) Silverlight really is far superior to Flash in my view; the platform and associated programming language above all. I wish HTML5/CSS3 *were* more like Silverlight in fact. Adobe Flash is still more ubiquitous though, so can't have it all my way soon, regardless of HTML5...

Re: HTML5 Takes on the King of Plugins

CodingBandit — Mon, 17 Oct 2011 16:32:29 -0000

Again, adoption of HTML5 has to start somewhere - the more demand devs make for it, the quicker browser providers will come around to implementing the functionality. I'd say if you're a flash shop, implement in Flash first (since that's what you know), then at the same time invest in creating an HTML5 version of the identical requirement - it sounds like a lot of work, but it accomplishes a couple things, 1) it prepares your workforce for transition and gaining the benefits of reach, 2) it allows people that are currently using an HTML5 browser the opportunity to experience it. IMO, it's fun just to get out there with a capable browser and find out what others have done with HTML5. I think I'll be spending more time looking into Modernizr myself.

Re: HTML5 Takes on the King of Plugins

CodingBandit — Mon, 17 Oct 2011 16:22:52 -0000

Vijay, with HTML5, you now have the ability to create those animations using HTML/CSS/JavaScript - I think it's an attractive technology even for animation-only components.

Re: HTML5 Takes on the King of Plugins

Drew Peterson — Mon, 17 Oct 2011 16:22:45 -0000

I didn't think you were trying to belittle anything ;-) This is a debate, you'd have to be pretty mean to get under my skin!

I'll agree that we're moving in the direction of HTML for these types of applications. I think however, as I pointed out in my response to Ferdy's comment, that it does make good business sense to stick with Flash until there truly is feature and support parity. I'm not advocating shying away from HTML5 wholesale, just that there are compelling reasons not to make the switch.

Re: HTML5 Takes on the King of Plugins

Drew Peterson — Mon, 17 Oct 2011 16:17:00 -0000

Well, first off, thanks for the response. You clearly put a lot of thought into it and that's what we love to see.

"I agree though that Adobe can maintain a very strong position in rich web content if it includes HTML5 as output."
That seems to be the consensus from some of the other comments as well. I probably should have focused more on Flash as a development environment as well as the technology.

"Correct, but perhaps you missed the memo"
I did, actually. I also didn't intentionally leave out context from Hickson's quote to prove a point. I had only seen the quote and not in its original context, I wasn't aware it was part of a larger discussion. Oops.

"I think this is a lack of understanding of what vendor-extensions are"
Well, I understand that they are the standard way to be non-standard, but for HTML5 to compete against Flash in, say, animation capabilities, we'd require some things from CSS3. Given that some of those required things may only be implemented as vendor extensions currently, it does cause issues for the programmer. I'll agree with your point that there are solutions to hide those problems from a code maintainability standpoint, but it doesn't account for any differing behavior between the two implementations.

All-in-all, I agree that HTML5 is workable today. It wasn't my intention to bash HTML5 in my article, and I hope it didn't come across that way. I was simply presenting my case for choosing Flash at this time. I do think that HTML5 is the way forward, and I think Adobe probably realizes that, but I simply can't ignore the huge support for Flash currently. I don't think that HTML5/CSS3 is quite up to par in terms of performance, support, and features yet.
As for my good reason to recommend Flash, many many businesses would agree that Flash has the best support and therefore can provide the best market penetration for their product. While some large players are using HTML5, they also must support browsers that don't support HTML5 yet. When it comes to spending money, I bet most businesses would go with the safe bet, which is currently Flash.

Thanks again for taking the time and effort to respond, we love to see participation like this from the community, and is the reason we put together these pieces.

Re: HTML5 Takes on the King of Plugins

CodingBandit — Mon, 17 Oct 2011 16:15:50 -0000

I agree with you entirely. Building native applications for the proliferation of devices out on the market today (and in the future) will be ridiculously costly. With HTML5 you have an open, cross-device technology at your fingertips

Re: HTML5 Takes on the King of Plugins

Mark Doyle — Mon, 17 Oct 2011 15:03:32 -0000

I agree with Carey. Flash was a huge success because website operators and developers made it so, leveraging functionality that was not available in HTML4. HTML5 will replace Flash because this same group will choose to do so, leveraging functionality that does not require a plug in, and that reaches more platforms.

Re: HTML5 Takes on the King of Plugins

daris2000 — Mon, 17 Oct 2011 14:55:45 -0000

I think the decline of Flash over time is inevitable. There will be a legacy for years and even right now new Flash applications may be created in areas where HTML5 falls short. The main reason for its fall will be reach, as mentioned in the article. Consumers and businesses want their apps to work on iPads, and the amount of devices is only going to increase. Open is the only way to go.

I'd like to add that for many people Flash brought video, ads, and that's about it. It can do a whole lot more but I see few successful examples outside of some niche areas. I agree though that Adobe can maintain a very strong position in rich web content if it includes HTML5 as output.

In addition, it is interesting to look at Microsoft's decision as to move HTML5 in front of Silverlight for public-facing websites and applications. They realized they cannot get their runtime on enough platforms.

Finally, I'd like to make some remarks on Drew's text:

"HTML5 is currently a working draft. It’s susceptible to change at any time and is incomplete."

Correct, but perhaps you missed the memo. HTML as of a few months is now declared a living standard. Meaning it is never done and always evolving. They're getting rid of baseline versions. Likewise, browsers will no longer implement baseline versions. In fact, they never really did. This change of policy is too complex to explain here, but the bottom line is that both the standard and the implementation will move in increments now (except for Microsoft, who continues to baseline implementations).

"In fact, Ian Hickson, a maintainer of the HTML5 specification, has state that HTML5 would not be a published standard until 2022"

If you quote him on that, you should include the whole context. The context being that the W3C considers a standard to be published when two 100% implementations are in the market. We currently don't even have that for CSS2.1, although the majority of the web is built upon it. Therefore, this quote is taken out of context and was put forward by Ian to emphasize that HTML5 can be used way before that date.

"If I were building a large scale interactive application on HTML5, I would be very concerned when hearing statements like that."

Currently, it depends on which parts you are actually using. Google, Facebook and many other major services are using parts of HTML5 already. There's sophisticated strategies for progressive enhancement and polyfills to fill the gaps. Yes, if you'd like a pixel perfect 3D virtual reality webshop working across browsers, Flash remains the best choice, but in many areas you can use HTML5 as of right now. Many have gone before you, it's not pioneering work.

"Except a new kind of vendor lock-in is starting to pop up now. With HTML5 being an unfinished draft, vendors are left to interpret and implement certain features in their own way"

I think this is a lack of understanding of what vendor-extensions are. Vendor extensions are a standard way to be non-standard. It is a way for vendors to offer an experimental implementation of a feature right now, eventually to be replaced by the real implementation that lacks the prefix. This means that web developers can make better experiences right now, and eventually their vendor extension rules will be ignored. It's not a maintainability nightmare since there are solutions which hide these rules entirely.

One other aspect little known about HTML5 is that it has made error handling part of the standard. Before, vendors had to decide for themselve how to render invalid markup, which lead to most of the browser quirks we know today. HTML5 changes all that, the error handling is standardized.

Concluding, whilst I agree that HTML5 is currently far from perfect, it is a lot better and more workable than most people realize. Therefore, I strongly disagree with the advise to recommend Flash for anything you plan to build new, unless you have some damn good reasons.

Re: HTML5 Takes on the King of Plugins

Tridus — Mon, 17 Oct 2011 13:55:49 -0000

The biggest problem right now with Flash is that it's the single largest security hole on the Internet. We've reached a point where far more malware is installed via Flash exploits then via Browser or OS exploits. That has to change and it has to change quickly if Flash is going to survive.