Re: [dev] Talk about sane web browsers

2009/9/8 Uriel <lost.goblin_AT_gmail.com>:
> On Tue, Sep 8, 2009 at 10:07 AM, Anselm R Garbe<garbeam_AT_gmail.com> wrote:
>> I think it is clear that the existing web stack can't be implemented
>> in a less sucking way.
>
> This is ridiculous, are you saying that a web rendering engine can't
> suck less than webkit or geko? Are you fucking kidding me?

Theoretically it can suck less than webkit, geko or what not, but it
still would suck much more than a web replacement engine that has not
to support the current web stack.

BUT it is nearly impossible to do considering the man years needed to
make the existing web stack "work" and the continuous effort to keep
it working (considered all the upcoming features such as html5,
geolocation, device apis, etc that are expected to be supported by a
browser engine).

Undertaking the development of a less sucking browser engine could
easily end up in a more sucking browser from the end user point of
view... because it won't support every web page. Or in other words it
will be the w3m or lynx experience when trying to access Google Maps
;)

> Any implementation will suck hugely, but there is tons of shit that
> suck in webkit and geko that are in no way required to suck to be 'web
> compatible'.

Have you considered that this is because of the web stack and not
because of stupid WebKit C++ developers?

I mean come on, that netsurf thing is barely comparable to WebKit atm,
otherwise you must be joking.

>> That's why it's quite clever and pragmatic to
>> let the WebKit folks do the ugly work.
>
> It might be pragmatic, but there is nothing 'clever' about it.

The "clever" aspect is that others do the work, mostly people that
have far more clue about the current web stack than ourselves...

> Chrome is not a black box, what I meant was that Chrome has put WebKit
> into something much closer to a real black box than what uzbl and surf
> are doing.

Ok

>> Second, I see nothing wrong with experimenting with browser chromes
>> like what uzbl or surf do.
>
> Experimenting with anything is cool, pretending it is something it
> clearly isn't is not cool. Uzbl and surf are not 'sane browsers' that
> 'follow the Unix Philosophy', and to claim so is to be either
> delusional or disingenuous.

I more and more dislike using the term "unix philosophy". At least the
official surf website doesn't claim surf to follow the unix philosophy
or to be "sane" in some way. I didn't follow the whole thread here
however.

>> BUT using Chrome as a base looks like much more effort to me, not to
>> mention its baroque build system.
>
> Perhaps, but Chrome has actually built something much closer to a real
> black box around WebKit, which is what some people here were claiming
> they had done, when it clearly isn't the case, and unless that is
> addressed not only will uzbl and surf be little more than very thin
> coats of paint on top of the sewage leaking WebKit turd, they will
> also be insecure and performance hogs.

I'd be careful with this "Google gets it right" trust. Theoretically a
very thin layer around WebKit should perform better, have less bugs
and be more secure than the rather thick layer of "Google Chrome" --
but this only holds unless the developers of both layers are similar
experienced. And one has to admit that Google has hired damn good
developers compared to the average IT world.

Kind regards,
Anselm
Received on Tue Sep 08 2009 - 09:50:25 UTC