Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Jens Axel Søgaard]

The goal for the core Urlang language is to be almost 1-to-1 with
JavaScript.
Therefore I haven't used any JavaScript libraries. As Dan's examples
shows (see urlang/urlang-examples) it is straightforward to use existing
JavaScript libraries.

There are a few predefined macros that one can optionally use:
cond, case, for etc which are Urlang macros that expand into
core Urlang (think JavaScript) constructs. The ability to extend JavaScript
with a sane macro system was what I wanted to achieve.

The source code for Space Invaders is a port of an HtDP-style program.
The attempt was to match the original HtDP-style and I think it turned out
pretty well.

It contains an ad-hoc implementation of structs in Urlang (JavaScript), so
when
you want to access the x-coordinate of a player p, you will see (player-x p)
in the code. If the program was written from scratch I would have used p.x
which is the standard JavaScript notation of access x in an object (or
array) p.

If anyone wonders why I chose (var [x 42] [y 43]) rather than adding
an internal define to Urlang - it is due to the scoping rules. The
var-form follows the scoping rules of JavaScript - which means it
has block-scope.

If normal local scope is needed one can use the the let-macro (or letrec*)
from urlang/extra.

/Jens Axel




2016-03-23 12:42 GMT+01:00 Daniel Prager :

> On Sun, Mar 20, 2016 at 9:44 AM, Matthew Butterick  wrote:
>
>> When you put it that way, subjectively it still sounds good. I'd use it.
>> But objectively I can't foresee that it would be a wise investment of
>> anyone's Racket time. Let's face it: any web framework is lucky to live 5
>> yrs before developers tire of it and move onto the next thing, or browser
>> changes make it obsolete.
>>
>
> The shifting of the underlying platform is inevitable -- perhaps
> eventually it will settle down -- and empirically your point about
> web-frameworks checks out, suggesting that it would be unwise to bake in
> reliance on any particular framework.
>
> That said, I reckon that there are gains to be made via making use of
> existing JavaScript frameworks and libraries.
>
> OTOH: Here's a new example from Jens that has zero dependencies on
> external JS - a simple Space Invaders game in Urlang:
> http://soegaard.github.io/urlang/space-invaders.html
>
> Dan
>
> --
> You received this message because you are subscribed to the Google Groups
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>



-- 
-- 
Jens Axel Søgaard

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Daniel Prager]

On Sun, Mar 20, 2016 at 9:44 AM, Matthew Butterick  wrote:

> When you put it that way, subjectively it still sounds good. I'd use it.
> But objectively I can't foresee that it would be a wise investment of
> anyone's Racket time. Let's face it: any web framework is lucky to live 5
> yrs before developers tire of it and move onto the next thing, or browser
> changes make it obsolete.
>

The shifting of the underlying platform is inevitable -- perhaps eventually
it will settle down -- and empirically your point about web-frameworks
checks out, suggesting that it would be unwise to bake in reliance on any
particular framework.

That said, I reckon that there are gains to be made via making use of
existing JavaScript frameworks and libraries.

OTOH: Here's a new example from Jens that has zero dependencies on external
JS - a simple Space Invaders game in Urlang:
http://soegaard.github.io/urlang/space-invaders.html

Dan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Andrew Gwozdziewycz]

On Wed, Mar 16, 2016 at 5:56 AM, Daniel Prager 
wrote:

> Awesomely, Jens has been working on Urlang: a Racket-ish syntax for
> JavaScript, using the nanopass compiler infrastructure:
>
> https://github.com/soegaard/urlang
>
> and more ambitiously, a Racket (subset) ->JavaScript compiler (rjs),
> taking a distinct approach from Whalesong.
>
> * * *
>
> Just using plain Urlang together with sxml, Bootstrap (for css), and
> notably Ractive.js  for templating and data-binding I
> made the following demo / proof-of-concept:
>
>
> http://htmlpreview.github.io/?https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html
>
> Note: no heavyweight js runtime required: actually no runtime at all.
>
>
> Racket source:
> https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.rkt
>
> HTML output (including embedded JavaScript):
> https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html
>
> The abstractions are a bit leaky, but I'm quietly excited about where this
> sort of approach could lead.
>
> What do you think?
>

Very neat! I had to make some changes to it before I could see it. This is
mostly around "mixed-content" due to mixing HTTPS with HTTP for assets.
Otherwise, all this is *very* promising.



-- 
http://www.apgwoz.com

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Daniel Prager]

Awesomely, Jens has been working on Urlang: a Racket-ish syntax for
JavaScript, using the nanopass compiler infrastructure:

https://github.com/soegaard/urlang

and more ambitiously, a Racket (subset) ->JavaScript compiler (rjs), taking
a distinct approach from Whalesong.

* * *

Just using plain Urlang together with sxml, Bootstrap (for css), and
notably Ractive.js  for templating and data-binding I
made the following demo / proof-of-concept:

http://htmlpreview.github.io/?https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html

Note: no heavyweight js runtime required: actually no runtime at all.


Racket source:
https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.rkt

HTML output (including embedded JavaScript):
https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html

The abstractions are a bit leaky, but I'm quietly excited about where this
sort of approach could lead.

What do you think?

Dan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Matthew Butterick]

When you put it that way, subjectively it still sounds good. I'd use it. But 
objectively I can't foresee that it would be a wise investment of anyone's 
Racket time. Let's face it: any web framework is lucky to live 5 yrs before 
developers tire of it and move onto the next thing, or browser changes make it 
obsolete. I have a shoddy `css-tools` library that I sometimes think about 
cleaning up for public consumption, and then I get dizzy and move on to 
something else. 


On Mar 18, 2016, at 1:27 PM, Daniel Prager  wrote:

> Matthew writes:
> > The idea of a subset of Racket that compiles to this kind of statistically 
> > probable subset of JS is very appealing. 
> 
> Agreed. The way Jens has split it up is that Urlang is a thin, cleaned-up 
> Racket-ish syntax for ES5, with a bit of sugar and a macro capability.
> 
> Then there's rjs, which includes a runtime, and aims to compile a substantive 
> subset of Racket. Of course there's going to be a trade-off between 
> completeness and runtime-size, etc.
> 
> > Especially if it can obviate the need or should I say "need" for a JS 
> > framework.
> 
> So, I'm currently looking at some of the "better" JS / HTML / CSS frameworks 
> / libraries with a view to being able to hide the detail and relatively 
> easily spin-up browser-hosted applications. E.g.
> CSS / reponsiveness: e.g. Bootstrap
> DOM manipulation & binding: e.g Ractive, Mithril
> Data structures: e.g. mori
> Graphics: e.g. D3
> I take it obviating the need for frameworks would essentially mean pulling 
> necessary functionality up into Racket: analogous to writing in Racket rather 
> than using interop to access external e.g. C/C++ libraries from regular 
> Racket.
> 
> Is that what you mean? Sounds good to me, but more of a "longer term" 
> project. Elm might be a good model for this.
> 
> Dan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

['John Clements' via Racket Users]

> On Mar 16, 2016, at 5:56 AM, Daniel Prager  wrote:
> 
> Awesomely, Jens has been working on Urlang: a Racket-ish syntax for 
> JavaScript, using the nanopass compiler infrastructure:
> 
> https://github.com/soegaard/urlang

Looks very cool. I’ll use this the next time I write JS code.

Thanks!

John

> 
> and more ambitiously, a Racket (subset) ->JavaScript compiler (rjs), taking a 
> distinct approach from Whalesong.
> 
> * * *
> 
> Just using plain Urlang together with sxml, Bootstrap (for css), and notably 
> Ractive.js for templating and data-binding I made the following demo / 
> proof-of-concept:
> 
> http://htmlpreview.github.io/?https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html
> 
> Note: no heavyweight js runtime required: actually no runtime at all.
> 
> 
> Racket source: 
> https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.rkt
> 
> HTML output (including embedded JavaScript): 
> https://github.com/danprager/urlang/blob/master/urlang-examples/ractive/ractive-bootstrap-example.html
> 
> The abstractions are a bit leaky, but I'm quietly excited about where this 
> sort of approach could lead.
> 
> What do you think?
> 
> Dan
> 
> 
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Racket Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to racket-users+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Matthew Butterick]

> and more ambitiously, a Racket (subset) ->JavaScript compiler (rjs), taking a 
> distinct approach from Whalesong.


The short answer: +1

The longer: I am no fan of JavaScript, which I consider a brain-eating virus. 
[1] But IMO the "frameworkification" of JS has made it even worse. Yes, the 
wheel-reinvention critique is also leveled at Lisps. But never have so many 
devoted so much effort to creating ever-larger elephants to crush ever-smaller 
peanuts. [2] 

And yes, I maintain a package called `sugar`. Dry humor intended. Would that JS 
frameworks acknowledged their place in the world. Which is mostly to create 
mutually unintelligible dialects of JS.

Today, I use Pollen as a JS preprocessor. [3] That's a lot better than JS 
alone. But I notice that 90% of my JS code uses 10% of the language. The usual: 
variables, loops, functions. Manipulation of the DOM. Async requests. The idea 
of a subset of Racket that compiles to this kind of statistically probable 
subset of JS is very appealing. Especially if it can obviate the need or should 
I say "need" for a JS framework.


[1] https://youtu.be/20GGVNBykaw?t=128
[2] http://stackoverflow.com/questions/3811678/add-two-variables-using-jquery
[3] http://pollenpub.com

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [racket-users] Racket -> HTML+JavaScript using Urlang and Ractive

[Daniel Prager]

Matthew writes:
> The idea of a subset of Racket that compiles to this kind of
statistically probable subset of JS is very appealing.

Agreed. The way Jens has split it up is that Urlang is a thin, cleaned-up
Racket-ish syntax for ES5, with a bit of sugar and a macro capability.

Then there's rjs, which includes a runtime, and aims to compile a
substantive subset of Racket. Of course there's going to be a trade-off
between completeness and runtime-size, etc.

> Especially if it can obviate the need or should I say "need" for a JS
framework.

So, I'm currently looking at some of the "better" JS / HTML / CSS
frameworks / libraries with a view to being able to hide the detail and
relatively easily spin-up browser-hosted applications. E.g.

   - CSS / reponsiveness: e.g. Bootstrap
   - DOM manipulation & binding: e.g Ractive, Mithril
   - Data structures: e.g. mori
   - Graphics: e.g. D3

I take it obviating the need for frameworks would essentially mean pulling
necessary functionality up into Racket: analogous to writing in Racket
rather than using interop to access external e.g. C/C++ libraries from
regular Racket.

Is that what you mean? Sounds good to me, but more of a "longer term"
project. Elm might be a good model for this.

Dan

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.