World: r3wp

The browser as the launch platform for applications has always been 
an interesting idea. The fundamental problem of the sheer complexity 
of it can be solved with R3. If done right, it can completely wipe 
the floor with browsers and AJAX. I think the problem is that we 
haven't been speaking in a language that people can understand, such 
as "browser", "web2.0" and "webserver", but instead "dialects", "VID", 
"Viewtop" and "X Internet" and people go "huh?".

Some things I believe are needed to do this right:


- Browser form factor. People are used to browsers, not Viewtops. 
What's always the first thing a complete newbie computer user uses, 
when wanting to do anything on the internet? A webbrowser. I don't 
want a desktop inside my desktop. There are tens of solutions for 
such things and they are almost all forgotten. Carl is doing the 
REBOL browser. When you fire up R3, you will get what looks like 
a webbrowser and acts like one. The concept has to work equally well 
for people like us, as well as 5-year-olds and 95-year-olds.


- Do apps that are similar to webapps, like GMail. That's a quick 
way to compare. Don't you think a 50k GMail look-a-like inside a 
REBOL browser running at native speeds would be _slightly_ impressive? 
Remember to say that you can serve 5 times more users with the same 
bandwidth. REBOL can help make raw numbers look better without much 
effort. Google would have to use it as a content platform. They have 
no other choice. :-) Chrome? What's that?


- Plugins suddenly are very flexible. You don't have plugins as in 
Firefox, but helper scripts that can enhance/change your browsing 
experience. 15k full screen document reader that prettifies plain 
text files? Sure thing. Blog posts presented in that would be much 
nicer to read. Out goes the PDF reader.


- Do apps that are completely out of the league of AJAX, such as 
multithreaded P2P systems. In fact, why not build P2P capabilities 
right in? Have different instances of the browser allow users to 
connect and chat, when they are visiting the same "Rebsite". It's 
sort of like going into a physical store and chatting with the other 
customers and you decide to exchange business cards. Initial contact 
without needing email. Do the same thing with chat support for an 
article that you bought at that  "rebsite". Current websites are 
almost completely anonymous. You don't feel you are entering a live 
community. Coded in REBOL/Services.


- Webpages are now REBOL scripts. In R3, scripts can be closed and 
encrypted, so you can't read the source and you can sell scripts 
and have them signed. The best you can do right now is some kind 
of code obfuscation.

- Windows, MacOSX and Linux version.


- "A webbrowser that directly supports OpenGL without obscure/limited 
3rd party plugins." Say that again in your head.


- It's very important that the public get to see that creating REBOL 
scripts for the browser is very similar to creating plain HTML pages. 
REBOL scripts can be served off a plain webserver. All the infrastructure 
is already there. Or how about serving scripts from the browser itself? 
AltME can both be a client and a server. It's that P2P thing again.

- Browser would run wherever R3 runs.


- Market it as Web 4.0. Market it as a direct competition to current 
webbrowsing.


- Browser would be a 500-600 kb downloadable exe that starts immediately 
without installation. From deciding to get it, to be using it to 
browse "Rebpages", it should not take more than 30-45 seconds.

- We need AltME in that browser (Altissimo?) as well as QTask.

For developers:


- It's easy to create an HTML file in notepad and display it in your 
favourite browser. It's going to be equally easy to create a REBOL 
script in notepad and see it running in your REBOL browser. A 5-year-old 
who has just learned to type, should be able to create a script and 
display it.

- One language for everything.


- Everything is free. You can start out with notepad. The barrier 
for creating content is about as low as it can get.


- You wanna code slow web 2.0 apps or fast web 4.0 apps? Hard choice, 
I know.

That's a great vision Henrik.

Henrik that is very interesting.  Is that your idea or what we should 
be seeing when R3 is released?

A REBOL browser is a great way to really get REBOL out there and 
might be used by non-developers just as their browsing tool instead 
of current browsers.

Paul, I began rambling about replacing the viewtop with a browser-like 
deployment platform in the r3-alpha world. I didn't expect that 30 
seconds after posting it, Carl wrote something akin to "Henrik, that's 
basically what I'm doing now." and at that point he hadn't said anything 
for over a week. :-)

He has since only talked more about the new VID, so I don't know 
if he's leaving the browser implementation up to us or if he's actually 
doing the browser itself. But it looks like the plan is for a browser.

This is basically what I meant earlier about being "psyched about 
a REBOL browser". I left out things like video playback and advanced 
audio, because I don't know yet what the approach for making those 
things possible will be. But if they are possible, they would be 
equally possible, like OpenGL would be possible.


And if it turns out that he won't do the browser himself, then it 
can easily be a community effort, not hard to build.

Not hard to build, but hard to design. Graphics models, interaction 
models, security issues, trust issues, resizing and reflow, these 
are all tricky problems. I can see why it would be taking a while 
for Carl to think through the implications.

But a REBOL Browser will introduce REBOL to others that have never 
used REBOL.  And if REBOL gains some acceptance then it means that 
other browsers will have to begin to integrate some compatibiliity 
with REBOL.

yes, the point is to say "hey, this is like a webbrowser, only much 
faster."

Henrik: "When you fire up R3, you will get what looks like a webbrowser 
and acts like one." - not just that, I want R3 not just to look like 
a webbrowser and act like a webbroser but actually TO BE a webbrowser 
- download R3 (few hunderts kB), run it and be able to browse REBOL 
pages - and if you enter *.html - just show some window that says 
"downloading" and download some REBOL plugin that can display webpages 
(being based on Gecko, Webkit, whatever) - it will be few megs download, 
but people are used to it. This is definitely possible - it's possible 
to display OpenGL etc in View window so I believe there are some 
libraries to do this ("somebody" just needs to make an interface 
to them ;)'


The thing is that R3 browser (and just a R3 browser) will be once 
again a great platform without apps (Be Inc etc...). If R3 browser 
can display classic HTML+JS+CSS+DOM+XML+AJAX+WHATEVER combo it's 
win-win situation. Lots of apps available and we can improve them 
one after one to show it can be done much easier and faster.

yes, I agree

But we also have to realize that REBOL is only the means, not the 
end. I don't know if the REBOL browser should directly go under the 
REBOL name. Perhaps the name should be more inline with what KHTML 
or Gecko means for other browsers.

And simply say "we made a new kind of webbrowser. it's much faster 
than your old webbrowser.". Perhaps market it as a side product of 
REBOL. This would bring up the old discussion again of what REBOL 
is.

Good luck. Firefox receives $57 milion annually from Google alone. 
Go big, or go home.

Well, even the Firefox had not that backup from begining.

It was just an independent brach of Mozilla bloat.

what about the new documentation of vid in rebol3 ?

Making another web browser won't help - new web browsers are made 
every day,and most don't catch on. Google's Chrome is an exception 
mostly because of the reputation of Google, and some nice features, 
but even with that it is unlikely to make much headway against the 
browsers that people are already using. Web browsers are commodities.

Firefox wasn't an independent branch of Mozilla, it was a branch 
of Mozilla (the software) written and supported by Mozilla (the organization 
with corporate sponsors and backing).

Has the API (or header file) for the R3.dll been published anywhere 
yet?

No, nor has any R3 dll.

At least not for a while. We have had only standalone builds for 
almost a year.

The API isn't set at this point, because of the core changes and 
more.

I only have the public alpha which has a dll.

It sounds as though there is still a huge amount of work to be done 
before R3 is going to be ready.

Well the API wasn't finalized even then, and then would have needed 
to change drastically with the Unicode changes. That doesn't even 
include Carl's current rewrites or the unfinished tasking model.

New blog!  http://www.rebol.net/r3blogs/0147.html

Henrik - very good POV ....

Terry - why are you constantly demotivated? You may as well take 
another aproach - use rebol free technologies, you will be safe this 
way. Watch R3, it can only surprise you, or it can fail, but you 
will not be hurt ...

As for FF, they receive money from Google only because Google is 
preset as default search engine. We might do similar if R3 browser 
gets popular :-) At OSNews, I also suggested Google to adopt REBOL, 
as it is THE ONLY small RIA technology along to Flash, SilverLight., 
which is a complete platform ...

terry's cup is always half empty

It is good to have Terry here otoh, or I thought I am the most depressed 
person about REBOL here :-)

After reading Henrik's post, I also think we found RT a good REBOL 
advocate ;-)

Henrik - re new name for REBOL browser. You mentioned "side" effect 
or so .... I think that the best so far is FireSide, even if .com 
domain is chosen ... it suggest Fire as in FireFox, FireBird, and 
we aproach it from side. Of course we could come-up with anything 
else ...

call it WMD browser

Graham, that won't work. USA users wouldn't find it. :-)

LOL

The rest of the world would believe it wasn't real if we claimed 
to find it.  ;-)

Rebol as being 'half empty or half full' .. either way, is a good 
metaphor.

Be serious. There's no way you can pull the necessary resources together 
to build anything like a browser. Just won't happen.  Call that 'half-empty', 
if you want.. I call it 45:1 odds against it every happening.. any 
takers?

Terry - you are starting to be boring with your attitude. Those naysayers 
like you, always claim anything like that before some product becomes 
eventually popular.

If by "like a browser" you mean implement HTML rendering and styling, 
a JavaScript interpreter and all of that, then I agree. If you want 
to implement a REBOL browser, then you are dead wrong. It's not the 
browser part that is the hard part.

I meant the former

I don't see the point to the former.

If the html browser part is seperate from the rebol "bowser" (as 
a plug-in using web kit?) it would not be too tuff.

That is also not what Carl is doing.

I don't see the part to implementing an HTML browser at all - we 
already have those, and they suck.

part -> point

I can see the point to implementing a compiler from a REBOL dialect 
to HTML/CSS/JavaScript though.