[REBOL] Re: REBOL Web Browser Plugin (Alpha)
From: gjones05:mail:orion at: 6-Jun-2001 13:06
Comments interspersed...
From: "Carl Sassenrath"
> That's the advantage of going the plug-in route to doing this.
> But, Sterling and I have been discussing this. If instead of
> installing a plugin, we just made an application association
> to the MIME type, we would not leave the blank page.
>
> But, that would break EMBED tags.
>
> Putting it another way: it's probably not a good idea to be
> both an app assoc and a plug-in. The app assoc allows links
> to .r files to launch REBOL. The plug-in allows pages that
> contain EMBED with REBOL to launch.
Somehow the creators of the TclTk plugin found what appears to be a
middle ground. One difference is that the TclTk does embed the applet
window within the browser window. I offer the follwoing information as
food for thought, and as a means for comparison. (Great time to learn
from their mistakes.)
Here are some examples of how the code is referred to in the Tclet
series. The first example demonstrates in-lined code that runs embedded
in the browser window.
<embed type="application/x-tcl"
script='
button .b -text "Click Me!" -command changeBg
place .b -relx .5 -rely .5 -anchor c
proc changeBg {} {. configure -bg [format #%03x [expr
int(rand()*(1<<12))]]}
'
width=120 height=80>
The following example is also embedded, but the source is pulled from a
file.
<embed type="application/x-tcl"
src="console.tcl"
width=400 height=400>
Having a link to code causes the Tclet to run in the whole browser. The
downside is that if the browser window is not large enough, the Tclet
face is clipped, and resizing the browser window does not redraw the
face. The only way to fix the problem is to enlarge the window and
refresh. (Of course, the window could be resized appropriately in the
html.)
<a href="http://dev.scriptics.com/software/plugin/tktetris.tcl"
target="tcl-app-example"> TkTetris</html>
However, for files stored locally on my Win98 box, the .tcl extension is
properly recognized and run in the Tcl interpretor.
My conclusion is that there may be a way to have your cake and eat it
too. I know little about plug-ins, but from a usability perspective,
Tcl's implementation seems to work as I would expect it to work.
> Apparently, the Netscape wizards never imagined that someone
> would try to do both. Their API does not address the issue,
> which could be solved by simply allowing our plug-in to NOT
> cause the page forward operation... or to allow it to tell
> the browser to immediately go back a page. Otherwise you
> have this blank page problem. But, we don't see that anywhere
> in the API docs (on Netscape site).
When I looked last night, I didn't see it either.
> We're open to ideas. We are experts in REBOL, not in browser
> internals. This has been problematic. We need to be careful
> not to set up a legacy issue for REBOL. If we do this wrong,
> it could come back to bite us later when we want REBOL to display
> its content directly within the browser window.
I think we all would agree with not creating legacy problems. But that
is why the REBOL plug-in is in Alpha - no promises have been made.
Right?
> Any browser experts out there?
Unfortunately, I am not. But I have downloaded the Tcl plugin source and
will peruse it later for practical pearls that I would be happy to pass
on.
--Scott Jones
