Re: [dev] [ubase] Announcing release 0.1

From: Christoph Lohmann <>
Date: Fri, 02 May 2014 01:51:30 +0200


On Fri, 02 May 2014 01:51:30 +0200 FRIGN <> wrote:
> On Thu, 1 May 2014 11:47:26 -0700
> Charlie Kester <> wrote:
> > Now I need to get busy and scrub my scripts, getting rid of or rewriting
> > anything that depends on GNUish cruft and other "improvements". (That's
> > not a complaint about ubase or sbase, btw. I'm actually looking forward
> > to this task.)
> Improving the world, one step at a time ;).
> If everybody just respected the standards and had common sense, we
> would face much less problems in today's computing.
> POSIX was heavensent and I wonder why we don't have something
> comparable for graphical UIs (One API for X, Wayland-compositors (by
> choice), ...), and I'm not talking about fancy compositing, but just
> basic ways of building and integrating user interfaces.
> What do you think?

That’s what the web did. With small religious fights over languages the
GUI frameworks emerged with all of them taking away resources to add
configurability the web easily took over by having CSS and millions of
web developers.

I was thinking of adding some »app mode« to surf. You run a website like
inetd, use surf as a frontend and so allow the integration of all kinds
of javascript libraries as a local application. The benefits are: The
frontend is suckless, you can easily create new applications because now
nearly everyone knows the web. BUT, this wastes a lot of RAM. For work‐
ing on this issue surf might need to support multiple rendering engines.
Simple websites use netsurf and bigger ones decide between gecko and we‐

Another way would be to have an intermediate bytecode for all the web
stuff and then interpret it efficiently on an X11 window. The DOM level
of web browsers is already the whole interpreted website. Having a com‐
mon bytecode here would save much interpretation time and resources.
Just export the DOM tree, save it and restore the state later on. Inbe‐
tween you use simple dom utilities to add elements, change the style or
copy out a subtree into another one. Clearly, that's the striking idea a
»graphical standard« must have. And this reminds me of omero [0].

Will a »simple DOM bytecode« standard suffice to what you wanted?


Christoph Lohmann

Received on Fri May 02 2014 - 01:51:30 CEST

This archive was generated by hypermail 2.3.0 : Fri May 02 2014 - 09:12:07 CEST