LiveCode Forums.

bobcole wrote: Thu Aug 18, 2022 4:33 am I was curious as to what the minimum html file, with the minimum javaScript, might be.
I put this into a .html file (I called it "Minimum.html") and put it in the standalone folder on my server.
I navigated the browser to open that file on the server. It worked for a simple LiveCode stack I'm playing with at the moment (-dp-3).

I tried it with my current web app test (LC 10 dp-4) and after updating the source link it works fine for me too. Very nice.

bn wrote: Thu Aug 18, 2022 8:34 pm This version runs much better on Safari Version 15.6.1 MacOs 12.5.1 than the one before.
Especially resizing th browser window vertically and horizontally and both is now smooth.
Congratulations.

Thanks Bernd! Yes, it works better and is more interesting to test now.

So far, I've tested it as follows:

- On MacOS 12.5 (Monterey), everything works including animation of moving stacks in Firefox, Chrome, Safari and Opera (although animation can be a bit jerky).
- On Windows 10, everything - except animation - works in Edge and Chrome.
- On iOS 15, everything - except "grab me", moving stacks manually and animation - works in Safari and DuckDuckGo.

HI,

I have a stack I am attempting to deploy on HTML5 (using Livecode 9.6..
Everything works fine in the Mac standalone build.
I get nothing that works from the HTML5 build.
If I run it in Test build it comes up with "Err #1".

Any chance this stack you have works in 9.6.8?

Thanks,
Andrew

aetaylorBUSBnWt wrote: Tue Jan 24, 2023 4:47 am Any chance this stack you have works in 9.6.8?

Nope, no chance, I'm afraid. The html5 output of LC9 is a dead end.
You'll need LC10 to produce a useful html5 standalone (it uses Web Assembly (WASM), unlike whatever older tech LC9 uses).

Hi,

Anybody have any ideas about LC 10 release?

The LC 10 DP 4 web page comes up with a 404 error.

So is ALL HTML5 code dead on LC 9.6.8 or just this stack because it relies on newer technology?

LC 10 dp 5 should appear quite soon, if I understand correctly, and then they aim for stable launch sometime during spring.

You can get some info on the progress here:
https://livecode.com/tag/livecode-10/

They had an offer a while back (2021?) where you could buy an LC 10 one-year license, which expires one year after the official launch, so it can be used and tested all through the development process. I don't know if that is still available. I can recommend taking it if it is.

My web app example stack is unlikely to work at all within the LC 9 html5 context.
But I suppose it should still be possible to deploy 'something' html5-ish with LC 9. Haven't tried that for years though, so can't remember really.

Andreas -
Just a quick note to say thanks for your work here. And thanks to the forum for being - well - a forum!

I only get to spend a handful of hours a month working on LC stuff for a little side business. When 10 came out I was excited but alas until MeetingSpace was used to at the 2022 devcon, there wasn't much for me to go on relative to structure and approach, etc.

Got off to a good start with the meetingspace model and setting up backend api's. Then onto the VIEW part... Made some stupid mistakes and convinced myself polygrids wouldn't scroll on the web. Came to the forum last week and posted. Bernd replied quickly and demonstrated simply that it does work - which sent me digging into my code where I found an artifact I'd stuck in there where I turned the scrollbars off if the platform was web (sigh).

Clearing that hurdle, the next one was getting the design to work responsively on the web like it does on the desktop (my desktop responsive design was doing perfectly what I wanted).

I haven't figured out the issue (and I haven't had a chance to digest your work yet), but seeing this example I'm greatly encouraged that I'll get there soon. I've downloaded your stacks and bookmarked the WhenInSpace site.

Again - thanks to you - and everyone - for restoring my hope in running with LC 10 (may we get to RCs and a production release soon!).
Onward and upward,
Ray

SWEdeAndy wrote: Tue Jan 24, 2023 12:07 pm

aetaylorBUSBnWt wrote: Tue Jan 24, 2023 4:47 am Any chance this stack you have works in 9.6.8?

Nope, no chance, I'm afraid. The html5 output of LC9 is a dead end.
You'll need LC10 to produce a useful html5 standalone (it uses Web Assembly (WASM), unlike whatever older tech LC9 uses).

LC9 transpiles its engine to JavaScript, which is certainly mature. But given its role as the most-used scripting engine on the planet, with the most powerful tech companies pouring millions into enhancing it daily, I'd hesitate to call it "old".

Moreover, while WASM enhancement is promising in some areas, I would caution against perceiving it as a universal panacea.

In this case of responding to window resizing, given that responsive design has been the norm among web devs throughout the decade LC has been working on their web export, this would seem more of a messaging issue than the isolated performance enhancement the WASM transpilation aims for.

That messaging enhancement does seem to be a focus for the team's v10 efforts, so yes, v10 will eventually bring LC web apps on par with web-native apps for responsive layouts, in ways the more limited messaging in v9 doesn't support.

I bring up the limited role of WASM here because from time to time we see posts where it's presented as a solution broader than it provides, and it's in everyone's interest to refine expectations to be meetable.

The perceived scope of WASM as a general overarching solution is understandable, as we've seen what compiled languages can do on the desktop.

But in all fairness, much of the feature richness of compiled apps is a function of their libraries and other tooling, where compilation really brings just performance.

With WASM, we can expect much of what LC needs to do as a rich-media platform to remain in JavaScript. The WASM spec prioritizes portability over browser-specific integration, so for example it offers no direct way to manipulate page elements like layout and UI controls, requiring JS as the intermediary for that.

And even where performance can be factored to isolation suitable for handling with WASM, observable gains depend on many factors beyond just script interpretation.

Even in desktop LC, performance is more often bound to the rendering/compositing requirements of the engine, in ways not even true machine-code execution of script interpretation can address. At best, in the subset of cases where WASM can help, as a portable tokenization it's always at least one step removed from true machine-code performance.

The Wikipedia page on WASM is a good intro, richly sourced with high quality links worth following:
https://en.m.wikipedia.org/wiki/WebAssembly

Among those links is this tale somewhat similar to LC's usage, outlining the challenges encountered when expecting the benefits of WASM to be more pervasive than discovered in practice:
https://zaplib.com/docs/blog_post_mortem.html

In short:

As long as LC's web export remains bound to using Emscripten to transpile the engine for emulating native controls and behaviors within the HTML canvas object, the move to adopt WASM to enhance some of that is a good one.

Just keep in mind that the "older technology" of JavaScript is likely to continue to play a strong role. And since JS is the most-used scripting engine, and that it's well maintained by well-funded teams, that's also a good thing.


As for the task at hand which started this thread:

If this is a form-based UI that needs to integrate with an e-commerce backend, as much fun as LC is, we see things like that across the web using a wide range of alternatives, some requiring no code at all.

With some flexibility in the design of the layout, you may find that the WordPress ecosystem offers a good set of NoCode building blocks for that, as LiveCode Ltd uses to drive their e-commerce solutions. My personal favorite these days is Drupal, which I'm using to rebuild LiveCodeJournal.com, arguably offering more fine-grained tooling, at the cost of a higher initial learning curve.

But if there's no business need which would benefit from immediate implementation, LC is always a good time and v10 may be worth waiting for on this.

Thanks for these interesting insights, Richard. While I wrote that comment quite some time ago, I may indeed have been a bit blunt in my labelling of new and ’old’ tech (though I wasn’t referring to JavaScript as such, of course).

Still - what I saw, and still see, is the very immediate practical aspect of it:
- The LC9 web deployment feature was a disappointment to me. I abandoned it.
- The LC10 web deployment feature impressed me from day 1. Suddenly I could bring so many of my LC projects to web more or less out-of-the-box!

The ”less” part of that is still annoying of course, and indeed requires the power of good ol’ (heh) JavaScript to remedy, if at all possible.

Some of which I feel could have been easily added to the LC10 dp’s early on, such as the resize message.

But I would argue that many business needs need no waiting at all on LC10, they can be realised (or at least tested) today. At least some of the shortcomings of the currently available LC10 version are bridged by my WebDeployHelper tool, so that less JavaScript knowledge is required by the LC dev to get some basic features working that end-users would expect to see.

Of course, as you and other wise developers often point out: Use the right tool for the task. If you’re building an actual website, with a need for web service integration, then a tool for that (not LC) is probably a better choice.

My ’mission', as it were, is to help LC users get their LC stacks deployed as web apps (commercial or not), and my credo is 'Do it now - try LC10, don’t wait (and here’s a little help on the way)’.

rayjbenet wrote: Thu Jun 20, 2024 1:13 pm Andreas -
Just a quick note to say thanks for your work here.

Thank you very much, Ray! I'm happy my examples and tools may be of help to you.
Since you seem to have the responsive design working in your stack already, it should be fairly easy to get it to behave well on web as well, if you test it with WebDeployHelper.

SWEdeAndy wrote: Thu Jun 20, 2024 8:28 pm Thanks for these interesting insights, Richard. While I wrote that comment quite some time ago, I may indeed have been a bit blunt in my labelling of
My ’mission', as it were, is to help LC users get their LC stacks deployed as web apps (commercial or not), and my credo is 'Do it now - try LC10, don’t wait (and here’s a little help on the way)’.

Yours is a valuable mission, very much appreciated by many. The value of a tool like LC is realized in deployment, so anything that removes pain points from the deployment process is very valuable to the company and the community.

As the CEO at my day job says often: "We reserve the right to get smarter."

You guys are awesome. Thank you.