Re: [dev] rebooting the web (it was: surf rewrite for WebKit2GTK)

From: FRIGN <>
Date: Fri, 31 Oct 2014 16:48:44 +0100

On Fri, 31 Oct 2014 15:31:44 +0100
Christoph Lohmann <> wrote:

> This is what this discussion boils down too. The web is too big to be
> changed with new established cultures.


> In the sense of the communist movement and the suckless interpretation
> don’t go the hard way, which is not possible. China has shown to let
> free development happen and use it as needed. The next step is to have
> guerilla and revolutioners in key positions. Convince them and allow
> simple access to the information of the web, where Javascript is the
> current enemy.

Or we go the Stalin-way, kill all members of the W3C and dictate our own

> The first priority for changing the web is to make it more usable for
> *us*, the people trying to use the commandline to use the web and script
> the data returned. Here web developers won’t change and aren’t consid‐
> ered qualified in their mass to be convincable. Next is the web
> browsers, which are software and on *our* level of concern. The browsers
> are a VM for a common interface now standardized and know as »the Web«.
> In this VM the DOM is the interesting part and how it’s represented.

To be really feasible, you would need some kind of artificial intelligence
which can untangle today's websites into a logical system.
The main issue I see is that people don't understand what the DOM is and
use it non-semantically. The DOM-noise how I like to call it is the main
reason why it is so painful to do web-dev nowadays.

> What's needed for the next steps:
> * Someone who knows any of the popular web rendering engines very well
> and can modify them without ending up in psychiatry.
> * The rendering engine needs to be modified to be reused for the
> points below.

Well, we'd need the big players on-board to change the rendering-engines.
A cumulative effort could change things (look at SPDY).

> * An abstraction of the DOM to be reused more easily on the commandline.
> (Maybe some daemon mode to run a browser and control it.)
> * Appropriate tools to handle the different exchange formats (images,
> audio, XML, JSON etc.)
> * Be able to easily switch to »normal mode« to have the a backward-compa-
> tibility.

Someone already said it, but if a profound change has to happen anyway, why
not rethink the whole idea of a DOM?

I think the basic model of a web-browser should not have to deal with
objects of any kind (images, video, web-application, webgl, ...).
The basic standard should be very strict in separating semantics and
presentation (kinda like LaTeX), but not define too much.
Instead of a stacked approach, it should be an axiomatic model.

To put it simply, what has the W3C been doing all these past years?
Right! Stacking more and more stuff on top of what was there. However,
writing (X)HTML hasn't become simpler in any way!
Or who can possibly remember this every time he writes a new page:

<?xml version='1.0' encoding='utf-8'?>
<?xml-stylesheet type='text/css' href='style.css' ?>
<!DOCTYPE html PUBLIC '-//W3C//DTD XHTML 1.1//EN' ''>
<html xmlns='' xml:lang='en' lang='en'>

And that's the basic stuff. Now, you can tell me that there are default
values for that, but often, this breaks (one word: charsets) and most people
have no idea what the fuck is going on.
Another aspect is site layout. You have to be a CSS-wizard to know how
to align div's vertically next to each other. I bet many people would
already jump on the train if we designed a new type of markup which
made stuff like this simple (again, an axiomatic approach).
(X)HTML could be considered axiomatic, but it isn't. Even though everything
is block-oriented, the entire system is full of contradictions.

So let me know what you think.


Received on Fri Oct 31 2014 - 16:48:44 CET

This archive was generated by hypermail 2.3.0 : Fri Oct 31 2014 - 17:00:09 CET