Re: [dev] Talk about sane web browsers

On Mon, Sep 7, 2009 at 9:35 PM, markus schnalke<meillo_AT_marmaro.de> wrote:
> [2009-09-07 11:50] Uriel <lost.goblin_AT_gmail.com>
>> On Mon, Sep 7, 2009 at 1:50 AM, Ray Kohler<ataraxia937_AT_gmail.com> wrote:
>> > On Sun, Sep 6, 2009 at 7:30 PM, Uriel<lost.goblin_AT_gmail.com> wrote:
>> >> You can't have a "sane web browser"[1] with an insane rendering
>> >> engine. All you are doing otherwise is giving a turd another coat of
>> >> paint.
>
>> >> [1]: Of course there can't be a sane web browser, but a different coat
>> >> of paint on top of webkit is not going to be any saner than Chrome,
>
> I agree on: There can't be a sane web browser, because the modern web
> is broken.
>
> But I don't agree that something sucks completely if it sucks in one
> part (that does not influence the rest).
>
>
>> > I think it's more like, "You can't have a sane web browser with an
>> > insane web".
>
>> > It didn't
>> > seem worthwhile to continue the thread there, since I don't really
>> > want to convince browser programmers that their cause is hopeless -
>> > and that's really what "victory" for me would look like on that front.
>
> The point is: It is simply not possible to have sane web browsers. But
> you both come to bad results IMO.
>
> Uriel says: Okay, so we'll not have a sane web browsers, thus we use
> one of the bad ones or better don't use the web at all.

No, I'm saying that I wish people would write or help write a browser
that sucks less. My point is that adding a coat of paint on top of an
existing browser (>90% of the browser is the rendering/js/etc. engine)
is not the same as writing that sucks less at all, it is a lazy and
meaningless gesture of little practical value and that contributes
little to the sorry state of things.

> Ray says: You web browser developers fight against windmills, that's
> senseless; you'd better stop. (You think this, but you don't tell,
> right?)
>
> Both oppinions don't help us.
>
>
> IMO `uzbl' _is_ an improvement, and in a kind sane. Because it makes
> (broken) web browsers fit into the Unix environment!
>
> Read my slides,

I would read them, if they were written in the standard language used
by the software industry (and the internet as a whole). People that
write stuff in marginal historical languages for nationalistic or
other barbaric reasons should not be surprised if their work is
ignored.

> I clearly state this: Take the broken render engine as
> black box and add sane interfaces around.

Well, first, the 'black box' is >90% of what the user, and code has to
interface with, so the box is not black at all, it is quite
translucent and with plenty of holes and leaks, and it is filled with
liquid feces.

Note that the Chrome people, while far from perfect, have built a much
more sensible 'black box' around a rendering engine which actually
does have serious benefits for security and performance, it still
sucks, but the approach is way more useful than the 'I will just write
another GTK UI and change a few keybindings'.

> IMO `uzbl' (without the render engine) is software that complies with
> the Unix Philosophy as it (1) does one job, (2) uses software leverage
> (works together), and (3) works with text streams.

You seem to have a very questionable definition of "Unix Philosophy",
can you point to a *single* Unix command with a remotely similar
interface or design to uzbl?

>
> It's the web browser that complies better with the Unix Philosophy
> than any other I've seen.

That is not saying much, and that is ignoring mothra and abaco.

> Only the render engine sucks (of course).

And the GUI toolkit, and many other things.

> But it may get exchanged ... it's only a black box anyway.

No, it isn't. No matter how much you repeat it, doesn't make it true.

uriel

P.S.: All I say about uzbl applies *exactly* the same to surf, which
seems to be an even more pointless "me-too" project.

>
>
> meillo
>
>
> P.S.
> I see `uzbl' as an example. It may not exactly be as I describe it.
> (In fact, it has it's dark corners. E.g. `--geometry'.)
>
> There might also be other browsers that are similar (surf).
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
>
> iD8DBQFKpWCV6aFpZ+X9qBIRAsS4AJ0RgH39y96pqhYLcqdkRPNsZ96UVgCfefkC
> ZtfYqkq+z5IwMLd9SIjF74k=
> =Hopt
> -----END PGP SIGNATURE-----
>
>
Received on Mon Sep 07 2009 - 23:16:10 UTC