Flash Sucks –- But everything else sucks more (2011) | Hacker News Reader

Flash Sucks –- But everything else sucks more (2011)

kristopolous.blogspot.com

20 pointsdredmorbius23 comments

The comments here and on the article itself both imply that flash is maybe a little less bad that the alternatives or that it isn't just the absolute worst. I does not escape me that this is 10-year old post. But - is it still the same now, really?

I'm a long time software engineer for UNIX and embedded, but I'm certainly not a web developer - the closest I get to a browser and related technologies is as a user/reader.

I think this may be part of the cause of a very different viewpoint - it seems like web developers actually bemoan the demise of flash? As a UNIX/Linux user especially, I was extremely happy to see the end of that. I danced a jig when Netflix and Prime dumped it. GAD it was awful! Only think worse was java in the browser. I had never considered anyone would be a flash apologist (even adobe employees).

So while using CSS, etc. etc. might be a pain for the developer, I think everybody else was more than happy to see it die. Just about everything from adobe seems to have the same basic stench.

I wonder what other things are popular for the creators but hated by the users? That's kind of an odd economic situation.

> I wonder what other things are popular for the creators but hated by the users? That's kind of an odd economic situation.

Electron, depending on how the developers use it. Most users doesn't knew that it's Electron because it's bundled with the application versus Flash which is visibly installed by the user, but most do indirectly complain of incomplete integration on their OS (usually they re-implement the notification system instead of using the native notification of the system) and very bloated memory requirements.

Speaking as a user, not a developer, I miss Flash because I had control over it. I could turn off "autoplay" everywhere with a global setting, turn it back on on a case by case basis, and the website owners couldn't do anything to stop me.

It has been 10 years, and the tooling for WebAssembly/WebGL is still quite far from what Flash was allowing in 2011.

Yes Unity and Unreal do support them as backend, yet those export formats remain as experimental prototypes.

The indie game development community moved into the mobile platforms, which gave them what the Web took away and is yet to properly replace.

And for what? WebGL + WebAssembly now have brought Flash back.

https://leaningtech.com/cheerpx-for-flash/

Flash was often misused - and you could argue it made it too easy for bad coders or malicious actors to load up a page with garbage. But as a game designer - especially for being able to code and work in the same environment as the artists producing assets - it was wayyyy ahead of its time. And it's true that even now, unless you use some other closed ecosystem like Unity, there is nothing even close to the workflow that Flash allowed for packaging a game that would work on the web, Android and iOS. And not just games - all kinds of interactive apps and simulations that could live outside a walled garden (which IMHO was why it was really killed).

Yeah, we've sort of got the tools now, ten years later, to be able to do what we were doing in the 2000s. Although it's still a lot messier. You now need to pick an audio lib and a drawing lib and a networking lib and hope they're all still supported in a few years, or be willing to maintain them yourself. But what really gets me is the way Adobe killed it in the end, by silently adding a self-destruct date to the plugin several versions in advance. I mean, people should still be allowed to somehow load my old games and apps if they want to. But now I can't even load them on my new machine. Literally hundreds of thousands of lines of hand-crafted code, three gaming platforms I built from scratch, several private business logistics apps that were still in daily use, and the majority of my work output from 2004-2014, was obliterated with very little warning.

Right now I'm still rewriting a critical AIR app that's still used by employees onsite all over the country, the last one I have in the wild. There's absolutely nothing wrong with the app other than that the language was killed, and it's basically impossible for me to push updates. Doing everything again is just a huge waste of time. The client can't understand why we're spending the money and time to completely rewrite something that will look and function exactly the same, only it'll be in TS instead of AS3.

The reason there's not an open source AS3 emulator out there that you can run in a web page is that Flash did so many different things, as related by this little article. It let you code sockets before there were websockets, and was still a friendly place for your artists to make their animations. It was like a swiss army knife with everything in one place, working pretty darn well.

What do you think of Cheerp, the project linked in a post above? Are there limitations that prevent it from being s replacement to the Flash engine?

https://leaningtech.com/cheerpx-for-flash/

Cheerp is an interesting project - for anyone who doesn't know, it's a cross-compiler that takes C/C++ and compiles to WASM. So in its Flash-specific version, it's basically running a complete Flash Player compiled inside another virtual machine in the browser. But it's not free to deploy. I don't know what their licensing scheme is with HARMAN, the new owners of the AS3/AIR SDK, but there are limitations on the binaries HARMAN can distribute on such a VM: Each binary is locked to a certain domain (a requirement of their agreement with Adobe, who insisted that Flash Player will not be deployable anywhere on the open web outside a whitelist); H.264 and AAC aren't available because HARMAN doesn't have a license; RTMP isn't allowed for the same reason. The need for Cheerp to get HARMAN to produce a binary locked to a domain likely explains why Cheerp's pricing is... extremely opaque. It might be a solution for keeping access to certain projects... but truthfully, without Adobe CC support for the design/dev ecosystem that Flash Builder provided, it's almost impossible to maintain a Flash/AIR project anymore, and it's hard to see a use case that would justify keeping one alive any longer than necessary, let alone building a new one. It's a shame. But these days I use a combination of PixiJS, Howl, ThreeJS, websockets, etc. to more or less do the kind of responsive, generative stuff I used to do in Flash.

Hi noduerme,

Do you have a email or another form of contact I can message you on?

or send me an email listed on the about of my profile.

No one's reading this thread anymore. Ask me whatever you want here.

At its core the beauty of Flash was the extremely tight interplay between vector graphics in nested timelines, and the the programming environment. Many people compare Flash to Unity, but I really don't find the two all that comparable. You don't create your assets in Unity, and by the time those assets get into Unity you have lost most potential for dynamic behaviour. Flash's tight integration allowed for extremely fast iterations, and its hierarchal model was very flexible and allowed you to work very fast and creatively.

I'm surprised no one has had a good crack at re-creating the core functionality of Flash circa 2004, based on web tech, or something like Kha[1]. I've thought many times about starting such a project. I think the biggest challenges are developing a solid vector rendering runtime alongside the vector drawing tools, but just about everything else that ex-Flashers want can be reduced down to some pretty simple functionality. You could even forego all the cruft of the display list, the event model, etc, and just go with a simpler immediate mode renderer and I think you'd still be retaining those core components that made Flash special.

There are so many use cases (eLearning experience, Digital Signage, Touch Screen Kiosks) that Unity isn't particularly suitable for, and for which HTML/Javascript is just clunky, that such a tool could far better accomodate.

1. https://github.com/Kode/Kha.

Flash's vector drawing tools were the best in the business, and I think would be very difficult to clone. It was all based on doing boolean operations between bezier splines ("planar maps" I think is a good search term), which I believe (I did a small amount of research into this) has no generally-applicable analytic solution, so you sometimes have to fudge it with approximations. And Flash's tools did spaz out sometimes; occasionally your drawings would disappear or turn inside out or otherwise corrupt themselves. To be honest I think it's amazing that it was as stable as it was.

AFAIK this drawing model came from Flash's predecessor FutureSplash Animator. In fact, it might have already been there in SmartSketch. It really was the most innovative part of what became Flash, and it's not surprising to see all the commercial and open-source would-be-replacements opt for the easier-to-implement vector model used in stuff like Illustrator, where you just have a stack of separate shapes.

I think any Flash replacement that does vector graphics the easier way will have a hard time feeling as natural to create in as Flash did.

I'm not sure it is particularly desirable to clone Flash's vector drawing tools. They were certainly nice to use but also pretty glitchy as you say. Functionality like vector shape tweening, while nice to have, is also not that important IMO. If the tools optimised more for artists and animators, they would mostly be using graphics tablets for more illustrative graphics. Vector tools for UI could perhaps be more along the lines of what Sketch offers.

I reckon taking something like Godot's vector draw tools, and combining that with a Flash timeline/library, with a Javascript runtime, and you would win over all those animators that keep saying "why can't they just re-make Flash".

Hey look, it's me from 10 years ago. I honestly think I was kinda crazy back then.

I find that your perspective from back then holds up. Flash critics are mostly blinded by their dislike of ads, Adobe in general and by the points made in Steve Jobs's 2010 open letter. The latter was central to Apple's strategy of cordoning off iOS into a walled garden and a 30% gatekeeper fee.

The state of the creative web diminished dramatically during this time because there were pretty much no tools that could replace Flash. Making something creative became the exclusive domain of web developers who understood the miasma of different technologies (JS, SVG, CSS, WebAudio, etc). What came of that? Nothing, because devs that know those things work for boring companies making boring software.

I picked up your earlier comment referring to that post, BTW.

The amount of creative freedom and how enabling flash was for anyone to create anything is still lacking greatly.

Sure, we still get.. some of it, but the amount of skill and dedication required is a big limitation.

I still consider the loss of flash a great cultural loss.

Sure, the web rendering technology has caught up, you can certainly do what flash did, and more, better, faster on modern browser platforms, but you can't do it nowhere nearly as easily, the bar for entry has been set so much higher, to the detriment of all.

There is some rust open source effort to keep old flash content working, besides cheerpx: https://ruffle.rs/

> Ruffle is a Flash Player emulator written in Rust. Ruffle runs natively on all modern operating systems as a standalone application, and on all modern browsers through the use of WebAssembly.

I think we can safely say that the web ten years later is a lot more serious and grown-up, but just as slow and a lot less fun. Less playful. And hidden malware, dark patterns, tracking and fingerprinting are all going strong.

By way of a comment to an earlier post on autoplay video in present-generation websites:

https://news.ycombinator.com/item?id=29455411

I really miss Flash. Particularly the vector drawing within it, there was something nice about the aesthetic that the brush tool created and the simplicity behind drawing with it.

All the "responsive" auto-stretch stuff turned bicycle science into rocket science. Often I think it would be easier to create 2 diff apps rather than one with responsive UI because it would be quicker and wouldn't have to cater to the lowest common denominator of finger-vs-mouse.

Still true.

Reply on news.ycombinator.com