|
|
Posted by Jim Higson on 03/14/06 19:54
Next wrote:
> Hi folks,
>
> Years ago, it occurred to me that a lot of the trouble
> of writing web browsers is caused by the upside-down
> arrangement of things: Javascript code exists inside
> a document, when really it should be the other way around.
How so? You seem to be saying the content should be contained in the
logic...?
> And yet, although this seems fairly obvious to me,
> having tried myself to write a web browser and given up,
> I don't see a lot of movement by major web browser
> projects in a direction that might TRULY fix the problem.
> I do see a few slow-moving projects: HTML5 and Web Applications.
>
> These are not hobby projects however, and it does seem
> that "industry" always has and always will have a
> preference for messy, bloated applications and poorly
> conceived standards because these things keep people
> buying new computers and justify companies' existences.
> We as consumers and/or hobbyists however should seek
> a better solution, and create it ourselves if necessary!
>
> I would suggest to fix the original problem. Here is my GUIML idea:
> GUIML would encode essentially the basic features of a
> modern GUI widget system, with enhancements to support
> fancier features like animated sprites that you see in some web pages.
> But basically it would reverse the fundamental problem with browsers,
> namely bad design caused by the code-in-document flaw
> that has led to enormous bloat and which has effectively
> made many perfectly usable computers obsolete,
> because web browsing is a vital app.
I don't think web browsing requires a very modern PC. A friend of mine still
uses a 800MHz machine from about six years ago with no problems, and I have
a 233Mhz machine that runs Firefox pretty well for most sites. There are
computers older than that still working out there, but not very many, and
there are lightweight browsers such as Dillo (http://www.dillo.org) that
run fine on them.
Despite what Intel tell us, web browsing is one of the things a modern PC
doesn't do much better than an older one.
> I would welcome any support or criticism of this
> idea but first take a look below at a sample GUIML web design.
> My initial idea is to simply take a familiar GUI like Motif or Java's
> GUI
> and use that as inspiration.
>
> And, to set things right I would completely remove from HTML
> any ability to run Javascript. HTML itself need not even be
> supported but could be replaced with any number of
> document formats such as RTF or something SGML based.
I don't think RTF will be popular for the web because it is difficult to
read and generate the code using scripting/templating languages. Besides,
it is set up for display on the printed page.
If you want to make older machines viable, at least base it on XML rather
than old SGML. XML is much simpler (and therefore faster) for the computer
to parse.
> Imagine the following webpage:
>
> <GUIML>
> <head>
> <title> Test </title>
> </head>
> <code>
> <!-- insert here javascript code for initialization, callbacks etc
> -->
> </code>
> <design>
> <MainWindow> <!-- widget that takes a menu, frame, and scrollbar(s)
> -->
> <PulldownMenu location=top preferredPointer="hand">
> <Menu title=About code="javascript_about()">
> </Menu>
> <Menu title=Products>
> <MenuItem code="javascript_callback1()"> First
> </MenuItem>
> </Menu>
> </PulldownMenu>
> <Frame name=main >
> <Table width=100% height=100% rows=1 columns=3>
> <tr>
> <td><PushButton code="js_button_callback()" /> </td>
> <td><Image code="js_img_callback()" /> </td>
> <td><Document url="foo.html" /> </td>
> </tr>
> </Table>
Using a table like this suggests you don't really understand the aims of the
standards you wish to replace.
> </Frame>
> <ScrollBar location=right type=vertical callback="js_callback2()" />
> </MainWindow>
> </design>
> </GUIML>
>
> Comments?
> z808x@yahoo.com
Navigation:
[Reply to this message]
|