[REBOL] Re: Rebol.org plugin question
From: SunandaDH::aol::com at: 2-Sep-2004 4:47
Tom:
> there is the whole only works on XP with IE or latest firefox with
> ActivX plugin that would have to be spelled out
Good point, thanks. If/when we provide a "run this script" link it'd have to
go to a page that explains that nothing may happen, and what to do if it does.
Jason:
> My understanding was that the Rebol plug-in already sandboxes Rebol scripts
> in the browser from doing damage, just as flash does -
> i.e. locks one out of the file system.
That's exactly the sort of question I'd like the answer to too. Is the
sandbox completely overriden when someone clicks "yes to all" to the security
message -- as it is in unplugged REBOL?
> I suppose a script could parse though looking for 'show view layout'
> etc and try to run them in the browser.
> A fast hack to explore scope of the existing possibilities and hope to
> encourage future ones.
Another possibility for a fast track is to enable plugin execution
only for named Library members. That way you and a few other brave REBOL
skirmishers could explore the issues while only risking your own machines. If
anyone wants to be part of a plugin beta team at REBOL.org, please let me
know your REBOL.org user-name.
> Does anyone have stats on browser platform access to Rebol.org
> for example?
> How many people are in fact using XP : IE Firefox ?
I got stats by the bucketload. But turning them into facts is not so easy.
Browsers spoof their identifies and there is no 100% way of knowing what they
really are.
Having said that, here's some stats for the whole of the month just gone,
August.
We executed 129,081 CGIs. Plus we served an unknown (because I haven't looked
it up) number of static pages (not that many: not much of the site is static).
Of those 129,081 CGIs, there were 1566 different user-agent identification
strings. From those we have to work out what the browsers might be. Here are
some actual UA id strings from August:
Mozilla/5.0 (compatible; Konqueror/3.1; FreeBSD)
Mozilla/5.0 (Windows NT 5.1; U) Opera 7.53 [en]
Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.7) Gecko/20040628
Firefox/0.9.1
Mozilla/5.0 (Macintosh; U; PPC; en-US; rv:1.2.1) Gecko/20021130
Mozilla/5.0 (X11; U; Linux ppc; en-US; rv:1.6) Gecko/20040414 Epiphany/1.2.6
Mozilla/5.0 (X11; U; Linux i686; en-GB; rv:1.4) Gecko/20030630 Galeon/1.3.8
Note that the major identity in all the examples is "Mozilla/5.0" while the
various sub-strings suggest a range of platforms and browsers.
474 12819 9.9% gecko
This means 474 different browsers identified themselves as "gecko" (i.e. had
gecko
somewhere in their UA id string. We executed 12,819 CGIs for them.
That's 9.9& of our total CGIs. Similarly:
777 26002 20.1% msie
163 6715 5.2% firefox
213 2492 1.9% linux
66 2760 2.1% mac os
18 1063 0.8% rebol (mainly automated services for downloading
packages and scripts)
13 47 0.0% amiga
(Note there is some double counting: "Mozilla/5.0 (X11; U; Linux i686; en-GB;
rv:1.4) Gecko/20030630 Galeon/1.3.8" is both "linux" and "gecko"
It is also not easy to turn these figures into counts of *human* visitors.
Approximately half the total visits are from bots, so that "9.9% Gecko" is
probably closer to "20% human visitors using Gecko" .
One conclusion you can draw is that REBOL.org attracts an above-average
number of non-IE users. That's probably not a surprise.
If anyone wants the August UA list to do their own stats, please let me know
-- it's a 12K file.
Hallvard:
> But if rebol.org should ever want to implement such a feature,
> I'll be happy to share my rebol- and html-code.
Thanks we might take you up on that.......It's also another fast track way
for anyone who wants to start experimenting with running random scripts as
plugins.
Sunanda
