Livecode & prestashop

[Richard Gaskin]

jbv wrote:

 > Is there any connexion between LC and e-commerce frameworks
 > like Prestashop ? I thought I've seen a thread about something
 > like that a couple of years ago on this very list but can't
 > retrieve it, and searching the archives is not an easy task...
 > Long story short, a friend of mine is working on a website
 > with prestashop, and needs a specific module. I was wondering
 > if I could build it with LC, export it as html5 and use it as
 > a prestashop plug-in...
 > Any advice ?

I know of at least one developer using the current version of LC's HTML 
export to deliver a product.  But it's a fairly specialized case in 
terms of business needs, so while it seems to be a good fit there I 
would hesitate to suggest it as a general alternative to native web 
development.

It still *may* be a good fit for what you have in mind, but of course 
that would depend on the specifics of what you're envisioning.

LC's HTML export relies on a JavaScript library created by translating 
most of the LC engine's C++ code base via Emscripten, and then using the 
HTML canvas object similarly to how it uses a window content region on 
the desktop.

The result is that it requires what is effectively the entire LC engine 
(in JS form) to be downloaded before the page can be rendered.

And it means that we have an interpreted language being interpreted 
within an interpreted language, resulting in performance that may not be 
bad but in many cases doesn't match what we're accustomed to on the desktop.

Moreover, being somewhat self-contained within this JS version of the 
engine and the canvas object, many things we take for granted in web 
development require re-thinking, where they can be done at all.

For example, in an HTML page we just use an <a> tag to link to another 
page.  But LC has no native understanding of HTML tags, and requires 
scripting to handle those.

In other cases, there are capabilities in the desktop that for good 
reason don't exist within the confines of a browser, such as arbitrary 
file access, registry manipulation, and more.

And responsive design for a great experience on screens of all sizes is 
much easier in many layouts with CSS rather than LC.  For example, the 
features provided in CSS3's Grid and FlexBox are pretty easy to learn 
and use, but would require hundreds of lines of LC code to attempt to 
emulate.

Bottom line:

If the module you're imagining has substantial features that would be 
costly to implement in native HTML/CSS/JS,

- and -

...the audience is already dedicated enough to what you're delivering 
that a wait time of between 20 and 60 seconds to load the JS lib LC 
engine (depending on network factors, CPU, etc.) would not be prohibitive,

...then LC's HTML export may be a good fit.

For other cases I tend to favor native web development. It's gotten so 
much easier with CSS3 that it's sometimes even fun.

When WebAssembler becomes universally adopted, and as network speeds 
continue to increase along with CPU, things may tilt the other 
direction.  But even then, those changes will mean native web apps will 
run even faster too, so the Emscripten method may at best become viable 
for a broader range of applications but for most is likely to remain a 
second choice to native web development.

-- 
  Richard Gaskin
  Fourth World Systems
  Software Design and Development for the Desktop, Mobile, and the Web
  ____________________________________________________________________
  Ambassador at FourthWorld.com                http://www.FourthWorld.com