why gnustep doesn't have a browser for it?

[g***@mail-on.us]

I recently learned about GNUStep from blog "Why Did GNUstep Never
Really Take Off?" <pinderkent.blogsavy.com/archives/134#> and begin to
study and try it a bit.

I am trying to figure out why gnustep does not have a browser yet. I
searched google and found this blog post explain current progress of
browser development:

http://multixden.blogspot.com/2008/01/browser-and-webkit-progress.html

The general impression I get is "wait, there are still a lot of things
to do before you can use a GNUStep brower".

I am totally newbie in the browser area so I am confused that it seems
a lot of work on the browser development went to parse HTML, render
pages in CSS etc. As far as I know this is the job of a browser
engine, and we already have many browser engine that is working pretty
well, what is the point to make a new different one? We have QT
browser that uses gecko engine, gtk-based browser that usese webkit
engine, which gives me an impression that a brower does not have to
have its own engine.

In my humble opinion a browser works for gnustep would be as simple as
building user inteface for accessing addressbar, managing bookmarks,
managing tabbed browsing (or, in gnustep with the fact not all windows
have to be on the task bar, tabbed browsing may not even be a
must-requirement), manage user's preference etc. Although I am not a
developer, I couldn't imagine this a very difficult work because I saw
other people did such thing in opensource world as well as commercical
software (where they wrap-in IE and call it something different with a
different user interface).

In short, my question is: is the reason we don't have a browser /for/
gnustep yet because the developer community want to have a browser
/of/ gnustep, that is entirely built with objective-C and gnustep
tools and libraries? Is there a believe having a non-pure browser is
worse than having no browser? If that is the case, it would be stupid,
because if you don't give user something to use now, at the pace
currently software industry goes, they will never need something from
you, because they are fullfilled with what others can offer, like
KDE/GNOME/Windows/Mac OS/XFCE etc.

[David Chisnall]

There is a browser in Étoilé svn which works as you describe, by
simply taking a Gecko X11 view and reparenting it. This is not a good
solution, however, because it means that things like text fields do
not have the same behaviour as other GNUstep text fields, buttons do
not look like GNUstep buttons, copy and paste / drag and drop do not
work in a GNUstep-friendly way and so on.

If you look at any of the GTK/Qt browsers that use WebKit (for
example) you will notice that they all use a branch of WebKit which
uses the native toolkits for the rendering work. There have been a
couple of efforts to port WebKit to GNUstep in the same way that it
has been ported to Qt, GTK, S60, Windows, and so on, but they are not
in a mature state (although Nicolas said he would take a look at
WebKit from this perspective at the hackathon next week, so we may
have some progress soon...)

David

Post by g***@mail-on.us
I recently learned about GNUStep from blog "Why Did GNUstep Never
Really Take Off?" <pinderkent.blogsavy.com/archives/134#> and begin
to study and try it a bit.
I am trying to figure out why gnustep does not have a browser yet. I
searched google and found this blog post explain current progress of
http://multixden.blogspot.com/2008/01/browser-and-webkit-progress.html
The general impression I get is "wait, there are still a lot of
things to do before you can use a GNUStep brower".
I am totally newbie in the browser area so I am confused that it
seems a lot of work on the browser development went to parse HTML,
render pages in CSS etc. As far as I know this is the job of a
browser engine, and we already have many browser engine that is
working pretty well, what is the point to make a new different one?
We have QT browser that uses gecko engine, gtk-based browser that
usese webkit engine, which gives me an impression that a brower does
not have to have its own engine.
In my humble opinion a browser works for gnustep would be as simple
as building user inteface for accessing addressbar, managing
bookmarks, managing tabbed browsing (or, in gnustep with the fact
not all windows have to be on the task bar, tabbed browsing may not
even be a must-requirement), manage user's preference etc. Although
I am not a developer, I couldn't imagine this a very difficult work
because I saw other people did such thing in opensource world as
well as commercical software (where they wrap-in IE and call it
something different with a different user interface).
In short, my question is: is the reason we don't have a browser /
for/ gnustep yet because the developer community want to have a
browser /of/ gnustep, that is entirely built with objective-C and
gnustep tools and libraries? Is there a believe having a non-pure
browser is worse than having no browser? If that is the case, it
would be stupid, because if you don't give user something to use
now, at the pace currently software industry goes, they will never
need something from you, because they are fullfilled with what
others can offer, like KDE/GNOME/Windows/Mac OS/XFCE etc.
_______________________________________________
Discuss-gnustep mailing list
http://lists.gnu.org/mailman/listinfo/discuss-gnustep

[Markus Hitter]

Post by g***@mail-on.us
In short, my question is: is the reason we don't have a browser /
for/ gnustep yet because the developer community want to have a
browser /of/ gnustep, that is entirely built with objective-C and
gnustep tools and libraries?

What's the urgent point of having a browser written in GNUstep if you
can run GNUstep apps just nicely along with apps coded in other
toolkits? GNUstep's premier goal is to be a development environment,
so all you could ask for is a NSWebView class.

Markus

- - - - - - - - - - - - - - - - - - -
Dipl. Ing. Markus Hitter
http://www.jump-ing.de/

[Zhang Weiwu]

Post by David Chisnall
There is a browser in Étoilé svn which works as you describe, by
simply taking a Gecko X11 view and reparenting it. This is not a good
solution, however, because it means that things like text fields do
not have the same behaviour as other GNUstep text fields, buttons do
not look like GNUstep buttons, copy and paste / drag and drop do not
work in a GNUstep-friendly way and so on.

Is it better having a non-full fledged product than having no product
for certain purpose, or worse?

I hope Étoilé release their browser soon and I'll try it first. I hope
Étoilé has something for the users NOW during when gnustepers working
hard to get a full feature browser engine that uses gnustep widgets.
When a gnustep-widget based rendering engine is available, full feature
and stable, Étoilé have no reason to not use that. For the moment I
think Étoilé people are doing the right thing.

Post by David Chisnall
If you look at any of the GTK/Qt browsers that use WebKit (for
example) you will notice that they all use a branch of WebKit which
uses the native toolkits for the rendering work. There have been a
couple of efforts to port WebKit to GNUstep in the same way that it
has been ported to Qt, GTK, S60, Windows, and so on, but they are not
in a mature state (although Nicolas said he would take a look at
WebKit from this perspective at the hackathon next week, so we may
have some progress soon...)

Expecting good news!

[Riccardo Mottola]

Hi,

as you cite my own blog, I feel compelled to write. I thought that blog post
was clear, but apparently it is not. Also, if you check later blog news, you
will see that SimpleWebKit and Vespucci made a certain, albeit slow,
progress compared to more than a year ago.

You are correct: there are two distinct things to develop: the engine and
the browser itself. Historically the two things have been tied together, but
in recent years some engines are reusable.

SimpleWebKit is a from/scratch implementation of a an engine and it adheres
as much as possible to the WebKit API, aiming to be a drop-in replacement.

Vespucci is a browser being developed on Macintosh and GNUstep to use a
WebKit-compatible engine.

As I already have written, you can compile Vespucci on Cocoa against WebKit
or SimpleWebKit just by changing the target in XCode: it will materially
just use one or the other Framework.

I think this setup is very interesting and has many advantages
1. it allows the parallel development of Vespucci without needing to wait
that SWK is complete
2. it allows to compare WebKit and SWK under the same condition and thus
discover where SWK does not follow correctly the WebKit API
3. it will allow in the future, if somebody will port WebKit to GNUstep or
write even another engine compatile with its API, to run Vespucci against
WebKit on GNUstep or another engine
4. it will allow to use SimpleWebKit on the Macintosh, this could be of
interest if some advantages arise

in any case, you see the current setup leaves a lot of freedom. The only
thing which I miss is Gecko, which I consider a heavy but high-quality and
industry-standard engine. Using gecko - once ported/wrapped to gnustep -
would require the usage of a different API or some sort of gecko to
WebKit-api wrapper.

Currently I welcome our choice, which comes from Nikolaus Schaller, because
after all GNUstep (and myStep) are close cousins to OpenStep and Cocoa, so
following their API makes sense.

Given that, I like the idea of SimpleWebKit a lot, for technical and
philosphical reasons. As you also can see, no time put in Vespucci is wasted
and a copmlete SimpleWebKit will jsut give more freedom of choice.

A GNUstep browser won't happen in a day and there are many places where help
is needed..

Regards,
Riccardo



----- Original Message -----
From: <***@mail-on.us>
To: <discuss-***@gnu.org>
Sent: Friday, May 08, 2009 11:20 AM
Subject: why gnustep doesn't have a browser for it?


I recently learned about GNUStep from blog "Why Did GNUstep Never
Really Take Off?" <pinderkent.blogsavy.com/archives/134#> and begin to
study and try it a bit.

I am trying to figure out why gnustep does not have a browser yet. I
searched google and found this blog post explain current progress of
browser development:

http://multixden.blogspot.com/2008/01/browser-and-webkit-progress.html

The general impression I get is "wait, there are still a lot of things
to do before you can use a GNUStep brower".