AltME groups: search

I will collect all questions from here and ask Carl. The mailing 
list is ok to send questions too. Anything that can reach me is ok 
actually.

btw - what about restricting this channel to only Questions and answers? 
I mean - let's create even other group - RT QA chat

Q. Given that  window transparency is OS specific, will there be 
a dialect that covers both Windows, Linux and 40+ other OS?  In other 
words, does RT plan on continued support of so many languages, or 
are we entering a new era of specific OS support?

I hope it is still valid that cooperation with RT is possible. I 
mean - last few weeks I play with some Win32 functions (thanks to 
Gregg) and I would like we would have proper app behavior in multi-monitor/multi-desktop 
environments .... so I wonder if any SIGs will be created, some ppl 
will be invited to participate, comment etc., or if RT is gonna cook 
it all themselves?

Hi .... with recent Rebcode releases, we can see that internally 
new Core is marked as 2.7 and View is marked as 1.4 Is it just working 
"title" or will those products be marked as that? And if so, can 
we know, what other changes will go for 1.4 View release target? 
Will there be any AGG fixes/additions (to support SVG RebGUI progress), 
or even VID changes? I still think, that VID is missing few fine 
styles as tab, group-box, better list as was introduced on IOS Developer's 
server, (eventually tree, menu), to allow novices to start using 
VID/View more productively. Any chance RT can tell us, what is the 
plan for 1.4 release?

(Copied from rebcode group) Could you add struct! support to /Core?

I keep on having situations that would be made much easier by struct! 
when I don't need libraries. For instance, conversions from external 
binary data encodings to internal REBOL values, say for file formats, 
network protocols and so on. Now rebcode has added other forms of 
strong typing like the type-specific opcodes and the vectors. Having 
structs with their constrained field types, their specific data layouts, 
would be a perfect match for the low level operations of rebcode. 
They would be helpful later when implementing your own data types 
as well.

Q: What does the world on Nov-15-2005 look like?


A: Our main goal is to get REBOL into the hands of more users, not 
just programmers and techies.... by the millions over time.  By doing 
that, we create a market for not only handy free REBOL apps, but 
also for commercial apps and entire businesses that are related to 
REBOL.



Q: Given that  window transparency is OS specific, will there be 
a dialect that covers both Windows, Linux and 40+ other OS?  In other 
words, does RT plan on continued support of so many languages, or 
are we entering a new era of specific OS support?


A: Our plan is to make that a window option that is part of the face/options 
for a window.  If an OS does not support this mode, then the option 
will be ignored, but the application will still be fully functional.



Q: I hope it is still valid that cooperation with RT is possible. 
I mean - last few weeks I play with some Win32 functions (thanks 
to Gregg) and I would like we would have proper app behavior in multi-monitor/multi-desktop 
environments .... so I wonder if any SIGs will be created, some ppl 
will be invited to participate, comment etc., or if RT is gonna cook 
it all themselves?


A: Yes, there are many such special interest projects currently going 
on. (Most of them are occurring via private projects in AltME and 
IOS.)  These days 90% of REBOL changes are done in cooperation with 
the REBOL community.



Q: Hi .... with recent Rebcode releases, we can see that internally 
new Core is marked as 2.7 and View is marked as 1.4 Is it just working 
"title" or will those products be marked as that? And if so, can 
we know, what other changes will go for 1.4 View release target? 
Will there be any AGG fixes/additions (to support SVG RebGUI progress), 
or even VID changes? I still think, that VID is missing few fine 
styles as tab, group-box, better list as was introduced on IOS Developer's 
server, (eventually tree, menu), to allow novices to start using 
VID/View more productively. Any chance RT can tell us, what is the 
plan for 1.4 release?


A: Regarding 2.7 and 1.4 question: we change the revision numbers 
(the second number) whenever there is a major change in REBOL that 
may be unstable.  The /core 2.7 kernel (that is in /view 1.4 as well) 
adds new datatypes to REBOL, and they are the first datatypes added 
in several years, so we consider this to be a major change, and marked 
it that way.
Yes, we do plan to be making a few AGG fixes very soon.

Oh, and regarding VID: we plan to be making very big changes there. 
More to come soon.


Q: Could you add struct! support to /Core?

I keep on having situations that would be made much easier by struct! 
when I don't need libraries. For instance, conversions from external 
binary data encodings to internal REBOL values, say for file formats, 
network protocols and so on. Now rebcode has added other forms of 
strong typing like the type-specific opcodes and the vectors. Having 
structs with their constrained field types, their specific data layouts, 
would be a perfect match for the low level operations of rebcode. 
They would be helpful later when implementing your own data types 
as well.


A: On structs: yes, we will enable this feature on core, but it should 
only be used for lower level code.  Objects are more powerful.


Q: Could you add an APPLY opcode to rebcode?

    apply: ["Apply function or path to arguments, save result" word! 
    word! | path! block!]

In rebcode:
    apply x f [arg1 arg2 ...]
Is equivalent to this in REBOL:
    x: do f arg1 arg2 ...


The advantage to doing function calls this way is that the arity 
of the opcode is fixed, even if the arity of the function called 
can't be known ahead of time. The value assigned to the function 
word could be either a function or a path, or for efficiency you 
could have a seperate opcode APPLYP for path values (I'd prefer just 
one opcode for generality but it's your call).


A: I'm not sure what is meant by the path for it. You mean for refinements?
That may actually slow down the apply interface.

Thanks Gabriele and Carl.

How about web publishing the answers somewhere, so the millions without 
Altme can see them?
Perhaps as a weekly blog?

Any reason why you can't be more specific here "Oh, and regarding 
VID: we plan to be making very big changes there. More to come soon." 
?

Web-public already ---- True, but only for a month or three, depending 
on the Q&A rate.
http://www.rebol.net/altweb/rebol3/chat390.html

shows only the most recent 300 messages.....Many web-public groups 
have "lost" a lot of valuable discussions that way. A more permanent 
and Google-friendly home is needed for much of this.

thank you very much Gabriele and RT  - this aproach is SO MUCH appreciated, 
that you will probably not believe it!

The Q for next weeks round is:

(note - my view may be influenced by insufficient knowledge in the 
area given) - last weeks I played with wrapping some Win32 functions. 
I started discussion on dll.so channel, to ask developers, if they 
would enhance interfacing to C libraries in some way, and there was 
few ideas appearing. We currently have also rather strange callbacks 
support (limited to 16) and I would like to ask, taking into account 
that DLL interface in Rebol was not changed/enhanced since it appeared 
long time ago, if RT sees any area in which it could be made more 
robust, developers friendly etc.?

A: I'm not sure what is meant by the path for it. You mean for refinements?

Yes, that is what I meant. If it is a major slowdown to check for 
this, add the APPLYP opcode I mentioned for the more general case, 
or have APPLY for the general case and APPLYF to apply functions 
only. Or you could just apply functions and have the refinements 
case be put in a do block or a wrapper function. You can handle refinements 
in rebcode, right?

Anyway, thanks for the struct! support and the new datatypes!

Nevermind on the paths for refinements: I just saw rebcode test 10 
and your apply, putting the refinements in the argument block - it 
will do nicely :)

Some time ago I talked with Carl here on AltME and mentioned PhotoThreads. 
Carl wanted me to keep an eye on it. There is newer version now. 
What is PhotoThreads can be read here: http://www.sics.se/~adam/pt/about.html
.... maybe we are lucky enough to get concurency with Rebol 3.x family 
:-)

Gabriele, I definitely would like to be informed and involved when 
VID changes are happening. You will let us know won't you ?

To rebcode I mean, and the flag would be the condition flag.

Gabriele or Carl, have you checked the conversation in "Rebol Enhancements" 
where we were hashing out enhancements to and comments on rebcode? 
Carl did request this in his blog...

I just fear one thing which tended to happen in the past - RT was 
silent before the release, release was done and then it was late 
for suggestions. Gabriele, please, try to avoid that :-) Tell Carl, 
please, that folks do need his attention here. You have some input 
in Rebol Enhancements group. In fact we don't know where to post, 
so some of us submit even to RAMBO. Everybody feels we are near freezing 
for 1.4/2.7, so please let's not cut some fine late enhancements.

hello, where I could  find facts and values about Rebol vs Business 
world ? with reboltof (christophe Coussement, we have the project 
to write a white paper about Rebol and Business (target date, on 
March 17, 2005, for the french Rebol Day, Paris). How many developpers 
works with Rebol ? how many dowloads of Rebol/core and View ? How 
many licenses for SDK ? Examples of Business with Rebol, tools, softwares. 
Real-world softwares.... etc.

While reviewing the action! functions, I noticed the path action. 
The doc comment says "Path selection.". The parameters aren't typed. 
Does anyone know what this action does, and how to use it? Or whether 
it can be or should be called directly at all?

Gabriele - could you please pass my latest comments/links to Carl? 
I looked into Io slide-show and found they do use libevent library. 
I posted links into TechNews group. There is interesting reading 
in-there, also hidden is some links - liboop and The C10K problem 
..... Maybe we will find another libraries for further rebol improvements, 
like Cyphre did find agg.

I've got a question for RT about rebcode. Is rebcode going to support 
paths and/or some kind of binding? Thanks

What do you think about  http://mail.rebol.net/maillist/msgs/39493.html
? Why not say a word in your blog, if you think that it's interessant 
for rebol developpment, and if you want to contact them ?

14.11: only RichText is late, actually. And, you have seen it at 
DevCon, so you know it's real. I agree with you an official update 
would be great, though.

Gabriele - I know about rich-text, but from Cyphre I also know, that 
at DevCon, IIRC, it was part of an external library, so hopefully 
there is no problem with further integration, all goes well, and 
we will hear some updated status info soon ... would be nice.

Q about encryiption within rebservices: my *guess* is if you are 
using core or view, then rebservices can' t use rsa encryption method 
but must use encloack/decloack method. And there is no plans to upgrade 
for rsa support in the next releases of  core and view . I'm rigth?

I worked on an STunnel installer and portal for SURFNet. Not sure 
what parts of it are public. Maybe Maarten will chime in.

Some questions for tommorows Q&A wednesday: 1. What is fixed/added 
in 2.6.2/1.3.2 (change-log, please) ? 2. What is planned for 1.4.0 
(rebcode, rebservices, rich-text, RIF, and last but not least, fixed 
sound ...) ? 3. When can we expect 1.4.0 ? Thanks.

In Carl's blog is written "More than 50 RAMBO tickets have been implemented 
or fixed." so there's at least ten more fixed bugs and convolve effect 
has been added, so there are some new features (or at least one :)

some are labeled "fixed in Core 1.3.2" and other names... don't know 
if it searches single words or exact phrases

Q: (note - my view may be influenced by insufficient knowledge in 
the area given) - last weeks I played with wrapping some Win32 functions. 
I started discussion on dll.so channel, to ask developers, if they 
would enhance interfacing to C libraries in some way, and there was 
few ideas appearing. We currently have also rather strange callbacks 
support (limited to 16) and I would like to ask, taking into account 
that DLL interface in Rebol was not changed/enhanced since it appeared 
long time ago, if RT sees any area in which it could be made more 
robust, developers friendly etc.?


A: We are planning to do a lot more on DLLs. In fact, future versions 
of REBOL will expand on the way DLLs are used in REBOL.   For example, 
I would like to see DLL support for media loaders and savers, so 
if we do not directly support a specific type of media file (say, 
TIFF) then an external DLL can be provided to load it.  There are 
a few other DLL related features down the road, but it is still a 
bit early to talk about them.



Q: I realize that the open sourcing of the viewtop wasn't that successful, 
but do you still intend to keep releasing newer versions of it? AFAIK 
the current release is over a year old. I've experienced a lot of 
obvious bugs in the viewtop editor, which I think can easily be solved 
by people outside RT.


A: yes we will continue to release newer versions.  View 1.3.2 fixed 
a number of bugs in the Viewtop editor that were listed in on RAMBO. 
Any fixes and enhancements from the community are greatly appreciated 
(by everyone, not only RT!)  You can post them to RAMBO, and we will 
review and include them (if they look good).



Q: While reviewing the action! functions, I noticed the path action. 
The doc comment says "Path selection.". The parameters aren't typed. 
Does anyone know what this action does, and how to use it? Or whether 
it can be or should be called directly at all?


A: the PATH action is what the interpreter uses to evaluate VALUE/selector 
expressions for each datatype. It is an internal action and has no 
external purpose in programs. These kinds of words often appear as 
a sort of "side-effect" from how REBOL is structured.  Datatypes 
are implemented as a sort of object class, where the interpreter 
"sends messages" to the class to evaluate expressions. The PATH action 
is a message that tells the datatype to perform a pick-like or poke-like 
internal function.


Q: Is rebcode going to support paths and/or some kind of binding?


A: Certain rebcode can support anything we feel is important to put 
into it, but note: many things we add could slow it down, by a lot. 
 For example, if we were to allow paths as variables, I estimate 
that rebcode would be about two times slower than it is now.  Perhaps 
one way to solve this issue is for you to use COMPOSE prior to specifying 
your rebcode body.  Within the compose, you can use IN object 'word 
to "pre-compute" the context references for words. For example:

    add.i (in object 'num) 10


Your question about binding is not clear to me. Rebcode already supports 
binding. Your rebcode can be part of an object context, and rebcode 
function words are bound to the code context.  (Perhaps you are referring 
to an older bug that has since been fixed?)



Q: What do you think about  http://mail.rebol.net/maillist/msgs/39493.html
? Why not say a word in your blog, if you think that it's interessant 
for rebol developpment, and if you want to contact them ?


A: Recently, I had the chance to sit down and talk with one of the 
main people from the One Laptop Per Child project (he is a friend 
of mine from Apple Computer days).  The project has an interesting 
goal, but there are also many difficult issues around it (not just 
in the technical side, but also on the social and cultural sides). 
My current understanding is that the target software is Smalltalk 
based. Yes, it would be very interesting to allow REBOL on that system, 
but if you look at the list of principals for the project, you will 
see that such a revolution is unlikely.  Is it possible that perhaps 
REBOL could provide some additional capability in the future? I think 
so. We have some special plans that I think will bring REBOL to platforms 
like that in the future. But, this is too early to say more.



Q: 1. What is fixed/added in 2.6.2/1.3.2 (change-log, please) ? 2. 
What is planned for 1.4.0 (rebcode, rebservices, rich-text, RIF, 
and last but not least, fixed sound ...) ? 3. When can we expect 
1.4.0 ? Thanks.


A: 1. Gregg is preparing a summary. The document should be available 
this week. 2. We are evaluating a large variety of changes in REBOL, 
more than even the 1.4 release that we've talked about.  I hope to 
be able to say more about these plans soon.

Q: "Is rebcode going to support paths and/or some kind of binding?"

Its about passing objects as parameters. Then binding is not known 
at compile-time, and changing anyway. It can be an ugly extra instruction.

Q: "One Laptop Per Child"

A: "My current understanding is that the target software is Smalltalk 
based"

Why not integrate rebol into smalltalk? You proposed a rebservice-access-library 
anyway, which - ups - is accidentally a real rebol under the hood. 
I expect that smalltalk is squeak, as i saw Alan Kay within the supporters, 
and Child and latops is his long-running topic. If, its multi-platform 
too, maybe we could even use squeaks multimedia in our lacking areas.

In response to the answer to my question about the path action (thanks!) 
I wonder, why do internal functions like path and native need to 
be assigned to global words at all? Can't they be assigned to some 
internal object fields, or in some way be hidden? I know that encapsulation 
isn't really REBOL's style, but littering the public interface with 
undocumented natives that can't even be used by mezzanine code seems 
a little silly.

As for the question about rebcode binding, let me make it clearer:


It would solve all of the object field retrieval problems in rebcode 
to have an opcode added that would be the equivalent to the IN native, 
or maybe BIND. You wouldn't need path evaluation to be added - you 
could just retrieve an object field word and use setw/getw with it. 
Yes, you can use apply, but it is very slow in comparison with opcodes. 
Working with object fields is a pretty basic operation that doesn't 
have any direct support in rebcode, and some have come to miss it, 
particularly those that use objects in data structures. I often use 
blocks for data structures when I can because of the context overhead 
of objects, but some are more used to thinking in object-oriented 
terms. Plus, there's all of those built-in object-based structures. 
Still, I suspect that many of these questions would go away if apply 
was faster :(

Q: What will be the future plan with the browser plugin, say within 
the next 3-6 months? How much more do we need before it can be called 
production ready? I have a few large scale projects that require 
a production ready plugin (so users won't have to download and install 
an executable).

i think that ladislav checked on this... and said that the current 
SORT was ok

that was Gabriele who asked about sorting and if it needs unicode 
... we need unicode for rebdb though ...

Gabriele: I checked the support of capital letters and found out, 
that all Czech capital letters found in ISO Latin 2 were supported, 
i.e. case conversions were supported. This is somewhat simpler than 
Czech sort.

pity we don't know RT's priority list even a bit. 1.3.2, that is 
the cool thing. However, async kernel retired no-one is able to know 
for how long, rebcode is retired no-one knows what and when happens 
to it (e.g. will there be bitmap opcodes as discussed implemented?), 
unicode, rif (where even RT admits it is important for Altme2.0/coop 
project, rebdb is stagnatting because of that), rebin, we wait for 
what-is-going-to-happen-with-VID since devcon, being said we will 
know "soon" ... etc ... So, what is next, guys, and when? :-)

It is not about "we tell you and you quote us on that" thing, but 
about bit of a planning for others ... we surely don't want 14.11. 
exact dates :-) Just months, average ....:-)

I have rebol scripts for converting chars to unicode, the problem 
is, that Rebol cannot print unicode strings and the socond thing 
is, that you need conversion tables. There must be some way how to 
distribute these tables from network, because I'm sure you don't 
want to make Rebol about 2MB bigger than now you to support some 
charsets.

Thanks, Peter and Pekr. But of course I am a little disappointed 
as I badly need unicode support right now.

I have two needs: (1) I need to display Greek, English, and Indonesia 
in the same file. (2) I need to be able to convert uppercase Greek 
characters with accent and breathing marks to lowercase without losing 
the accent and breathing marks.

Displaying is probably the most difficult, specially for indonesia 
language. Conversion should be that hard. I guess you are receiving 
your data in unicode format. So you will need a parser to scan the 
bytes in the unicode buffer (a simple string!) and maybe generate 
the display by using the glyphs that encode each char.

Indonesian is no problem. Greek with accent and breathing marks is 
the problem.

YekSoon: Presuming the "Indonesian Language" is Bahasa Indonesia 
and not one of the hundreds of native languages, the difficulty will 
be if it is written in Jawi script. (Which I'm sure you know is Arabic 
script with a couple of added letters).

I was away for a week plus few days, and I thought that maybe I will 
be suprised by some new releases of Rebol. With RT announcing removal 
of alpha releases it seems to me, that they also managed to remove 
steam out of the community once again ... ? At least ppl had something 
new to test, post new request, ideas, etc. Now we once again don't 
know, when things like Rebcode will return back, or how Rebcode will 
be further developed, as we know, it misses security things ....

Is there any schedule for anything actually? Rebcode, rich-text, 
rif? All we know for past two months is OS-X is being worked on/updated, 
and some unconfirmed rumours of Rebol 3.0 being considered to be 
developed. Any other info?

question: I own a Linksys WAP54G access point which runs on a MIPS 
processor with a small Linux server on it. I tried loading REBOL/Core 
for MIPS onto it, but it couldn't run.


The thing is, there is quite a lot of embedded hardware that runs 
such small linux servers. It would be easy to make control software 
via REBOL, connected to a PC running an encapped REBOL/View application. 
This would allow for rich realtime control software, rather than 
using the normal (slow and non-realtime) built-in webserver. I already 
managed to get realtime readouts on signal strength, by polling the 
access point through telnet and displaying a simple meter in a REBOL/View 
script, something not normally possible. But  you could do much, 
much more, if you could run /Core on it directly. I think there is 
a lot of unused potential here.


Would RT consider such ports of REBOL/Core to various embedded hardware 
products and provide a list of embedded hardware products that can 
run /Core?

addition: I mean embedded hardware products that run Linux and is 
fairly easy to access for experienced users

Thats what why-wire has done with openbsd. Running on smaller embedded 
hw requires careful crafting of memory footprint and cpu use.

core by itself consumes 2MB of RAM and depending on your program 
it can take more than 10% of the CPU. BEER was designed with the 
restrictions of embedded hw in mind, so it keeps very tight memory 
use and cpu use is the main reason for BEER to be async.

What's the current state and future plans of Rebol/Services? It's 
been a long time since there have been any news ...

Is there any news on the DIFF function as written about in the blog 
here: http://www.rebol.net/article/0126.htmland http://www.rebol.net/article/0128.html
?

very cool, thank you Gabriele ... that is exactly what was missing. 
IMO ppl can wait, even longer then roadmap check-point state, but 
there should really be some roadmap and some more-often releases 
of alpha stuff - at least I found it vital, as ppl were throwing 
in some ideas, bug fixes, etc.

Why is the View 1.3.2 release for Linux not on the main RT View download 
page? On the test build page it has release status, and I'm getting 
people who are unable to find the Linux download

problem is that it does'n work out of the box in all distros (you 
may need symlinking/installing libs), and there's the freetype issue.

Problems installing Rebol/View latest official release 1.3.2 onto 
Windows 2000 Professional with all the latest Service packs...


Rebol was installed using the default parameters, with the exception 
of the path, it was changed to C:\Program Files instead of C:\Program 
Files\Rebol\View.  It was noticed immediately that the Microsoft 
Quick Launch toolbar disappeared.  After an uninstall  we found that 
the entire IE\Quick Launch folder was erased, but don't know if this 
was done by the Install or Uninstall.  Afterwards, we rebooted the 
PC and found that the desktop image (which was added through IE) 
was also gone.

RT may want to try and replicate this install/uninstall inorder to 
confirm whether there is a problem with the installer.

It's not going over very well with my five web-developer colleagues... 
they are saying he has been infected by the Rebolla virus and getting 
a very good chuckle out of it!!!

Further investigation shows that the files that have gone missing 
from his system are all in the c:\Documents and Settings\<User Account>\Application 
Data\Microsoft\   Internet Explorer and MSN Messenger folders... 
the files impacted seem to be all of the customizations he has made 
to these two apps.

when he went to change from C:\ to d:\ and hit the "Change" button 
on the installer, we didn't catch that it had removed the "Rebol\View" 
portion and defaulted to the drive he launched the installer from.

so note that 1) it does not delete anything without asking confirmation 
and 2) only deletes the data dir which should be appdata\view\

no, it only seemed to delete registry settings and the appdata/view 
folder.  Nothing from his program files folder was erased, which 
is a good thing for him.

So one thing I would like to suggest is that the "Change" button 
in the installer defaults to the folder as indicated in the default 
path provided when you start the installer.  I think this is the 
root of the problem as the text looked like it was fully provided 
but the path field was short and we didn't confirm the full path. 
 In my opinion it should have included \Rebol\View, but believe it 
did not.

Gabriele - but maybe it would be worth to publish some blog article, 
about the state of Rebol 3. I know that every single project is delayed, 
but it starts to look strange once again - alpha in may, beta in 
June. It is end of June, according to another blog Carl is supposed 
to spend several weeks in France at that time, and no mention of 
when aproximately alpha will be out?

so it's not in the source for the view installer, and not in the 
normal mezz code. has to be a different component.

don't tell me, that for each instance of Flash e.g., it uses separate 
threads? And what embedded OSes, which do not have threads? I wonder 
if it limits number of flash banners on the site displayed?

Thats the part called thread-safe. Such things can be done, but need 
some organisation. For example (speculation): rebol does something 
like
  forever[ do-next-step ]

And now that has to be changed insomething which returns in between. 
And save all theglobals currently scattered everywhere in the sourcecode.

I know we talked about it a while ago, but those discussions are 
long gone from REBOL and from the web......I apologize, I should 
have archived them. Can you repost your thoughts on the default security 
model, in *concise* posts please? thank you!

I'm sorry, those discussions are long gone from AltME and the web 
archive, that's what I meant to say.

Extra Security, some thoughts:
- 'secure for ips, eg: secure [net ask tcp://rebol.com allow]
- don't share sandbox-folders between hosts.
- if possible memory-restriction, hd, cpu?
- clipboard-restriction somehow?

- check for memory-access, specially disable struct! . IMO real hackers 
will figure out how to inject code by poke. 

- reblets can store executable code by naming the file *.exe. Does 
not run immediate, but script can open folder in explorer by browse, 
and one wrong click runs it. (or is windows smarter now? Maybe you 
could add an own extension always, and maybe store everything as 
64#{} ?

- Make sure untrusted reblets don't run invisible, can snoop clipboard, 
or at least users online-times. I guess creatives can find other 
uses.

- Maybe some kind of log about starts/stops, with urls? To have a 
little chance of tracking. Some kind of global console.
- Running out of thoughts for now.

A couple of quick thoughts:

[*] Don't allow reading/writing outside of a local sandbox......That 
includes not allowing access to URLs elsewhere on the web.  permitting 
wider local access and permitting wider web access should be separate 
security settings
[*] Disallow send by default

Yes of course with dialog. And some way to set prferences. I like 
how noscript does that.

Maybe change it and open the users mail-client with the prepared 
email?

Thinking about it, yes. Maybe the standard things native, browser, 
email, editor. With an option to use mail and editor inbuild. Maybe 
these could run by 'launch, so they are rebol, but reblets cant touch 
the real code.

Maybe a native 'field, which is not accessible from the script until 
"enter", and allowing pasting there? Rebol3, how much protection 
can modules give?

Yes. You should try disabling third party cookies in your browsers 
and see how much stuff is leaked to  through that., easy enough to 
steal from a form a user just filled out ;-)

Looks like Volker covered the security issues I'd note.  About Rebol 
as a COM server process--I would think that would be the way to go. 
 Pretty sure that is how Acrobat runs, too.  Basically, the first 
time you run into a PDF on the web Acrobat32 starts, and handles 
all instances.

hmm, blocking port 25? What if I am a hacker and I run smtp on purpose 
on different port? ;-)

Be carefull with restricting Rebol. I like Rebol, because I can simply 
do things, which I cannot do in other apps. I cannot believe, that 
Maxim wants to remove send even from Rebol and replace it with some 
application. Why? Do I need xxMB large aplication just to send message? 
What will be better in Rebol than in Flash, if you remove the best 
parts of Rebol?

Because you run in the browser now and spy/adware/spammers will love 
rebol without such things.

With exe people run rebol intentionally (although /desktop goes in 
the other direction). With plugin user goes to a web-page and webdesigner 
turns that into a rebol-page without notice.

I ask once again - isn't it possible to send email using javascript 
and getting into browser settings?

Security extension, yes, removal of something - hehe, how uneducated 
imo :-) Is smtp so difficult to build? Having tcp socket is dangerous 
already, as I can build my custom smtp in script, and have server 
at the other end of the country, which listens on 8080 and doing 
smtp ....

I agree oldes, and we cant "remember" specific directories!

And a good sandbox lets you connect only to your homeserver, where 
the reblet comes from.

And I'm sure, you will not be able send emails from my browser as 
I don't remember that I allowed to use such a port to any application

Volker: And what's bad on connection to home?

I want both. Settings are in %user.r, by secure. And %user.r is modified 
by the panel. As it is currently with 'set-net and /desktop.

I do my own app, on puprose, and browser mailer pop-ups? Uh, that 
should be optional at least ....

And mail is critical IMHO. Its on account of the sender in the eyes 
of most people, even if one just fakes the from. If you can that 
from users machine, you have even the right headers.

And hopefully that control-panel is more verbose than the current 
requester. And offers good informations about the effects.

The plugin *needs* to be highly restricted by default. Please scroll 
up to the top of this group where BrianH and others made some fine 
points about security.

but system dialogs are half-way solutions - 1) they can't be translated 
2) they are ugly and do not copy design principles of your apps .... 
stating that - is there a secure way of how to overcome this? Could 
you provide your own UI and supply it for the internal security system? 
Probably not, as I could ask user completly different question :-(

I like that ugly and different. Tells me i am not working inside 
the app. Because inside the app, if it asks me "Do you like [x] please?" 
i click yes, whatever [x] is. Its in a sandbox, no?