Re: Client-Side Scripting using Meta-Programming

Здравствуйте, Mirrorer, Вы писали:

>Мы используем F# на клиентской и серверной части веб-приложения, что позволяет разрабатывать клиентский код на типобезопасном языке используя подмножество библиотеки F# , и обеспечиваем возможность писать сервеную и клиентскую часть кода как часть одного гомогенного типа описывающего логику веб-старницы. Код выполняется генерогенно, часть как JavaScript на клиенте, а часть как native код на сервере. Мы также используем монадический синтаксис для разделения клиентской и серверной части кода, отслеживая это разделение при помощи системы типов F#

Голубовский делает то же самое на Haskell:

Hi,

I have written a draft of the tutorial on Haskell programming for
client side of Web applications (related to Yhc/Javascript backend):

http://www.haskell.org/haskellwiki/Haskell_in_web_browser

Any feedback, suggestions, and additions (like your own code examples)
are welcome.

Thanks.

PS This part of the Tutorial does not cover XMLHTTP programming; this
is yet to be written, but there is Haddock documentation under
http://www.golubovsky.org:5984/_utils/yhcws/index.html

Yhc web service online (where you may experiment with your code):

http://www.golubovsky.org:5984/static/yhcws/MainGUI.html

Users guide:

http://www.haskell.org/haskellwiki/Yhc_web_service

--
Dimitry Golubovsky

и ещё одно письмо, <<[Haskell-cafe] "Hello World" in Haskell for web browser>>:

Hi,

I have noticed that some people tried to compile a traditional Haskell
program using IO monad to show "Hello World" using Yhc Web Service.

Thanks for your interest, and here are some clues.

The programming paradigm chosen for Haskell Web programming is not
based on monads. It is based on CPS-style calls to DOM functions by
Haskell code compiled to Javascript. Further on, additional layers may
stack up (such as Haskell Web Toolkit) to provide more convenient
APIs, but DOM is the basis.

So here is an example of "Hello World" program written this way:

-- begin pasteable code --
module HelloWorldDOM where

import CPS
import UnsafeJS
import CDOM.Level2.DomUtils
import DOM.Level2.Dom
import DOM.Level2.Html2
import DOM.Level2.HTMLElement
import DOM.Level2.HTMLDivElement

main = getHTMLDocument $ \doc ->
       documentBody doc $ \body ->
       mkText doc "Hello World" $ \txt ->
       mkDiv doc $ \dv ->
       addChild txt dv $ \_ ->
       addChild dv body $ id
-- end pasteable code --

The meaning of this:
* get reference to the HTML document node first (it is the parent of
everything),
* extract the <BODY> tag node, create at text element with "Hello World",
* create a <DIV> tag node,
* insert the text node into div,
* insert div into body

Or, same in HTML:

<html>
<body>
<div>
Hello World
</div>
</body>
</html>

but filled in dynamically.

Using Haskell Web Toolkit API, the same may be expressed in more
compact fashion:

-- begin pasteable code --

module HelloWorldHsWTK where

import DOM.Level2.HTMLDivElement
import Graphics.UI.HsWTK

main = docBodyC (mkDiv |<< textP "Hello World")

-- end pasteable code --

Here, docBodyC is roughly equivalent of the first two DOM calls, mkDiv
is same as above, |<< means "insert into container", and textP is a
wrapper around mkText.

Earlier, I posted the link to Haddock-generated documentation which
also includes the Haskell DOM API. Here it is again:

http://www.golubovsky.org:5984/_utils/yhcws/index.html

Hopefully this example along with the documentation provided helps
shed some light on Haskell Web programming techniques.

Feel free to ask questions.

Thanks.

--
Dimitry Golubovsky

	От:	BulatZiganshin
	Дата:	31.03.08 15:51
	Оценка:	9 (1)