AltME groups: search

When we added validation for the license field, we limited it to 
licenses that had actually been used by people contributing scripts:
http://www.rebol.org/cgi-bin/cgiwrap/rebol/license-help.r


I reckon there are way too many license variants already in the world 
-- it's the bane of the open source movement. (We had a list of over 
40 licenses that people in theory said they'd like to use when contributing 
scripts)

But,. as the note on the page says:

f you'd like to use a license for a contributed script, and you can't 
see it listed above, please send us Feedback.

And note that if you extract the code from a running rebol, you won't 
get the license.  I got the above from the new SDK.
So, that message needs to be included with your script.

Wasn't Carl thinking about a new license at one stage ?

Now the desktop source has a different license from the rest of rebol 
code.

There might be a standard copyright somewhere that says you can't 
reverse engineer ...

I guess it should be ... there are a lot of patches to the code on 
rebol.org, which maybe we should check the licenses on them.

Yes, Sunanda, I don't care much either, but this isn't my code, at 
least not completely. And though I'd think think that license hassles 
are just a superflouous burden ...

I guess in these times we should pay at least basic attention to 
it.

I presume AGG makes it easier to distort text for a visual Captcha 
...

Yep, we have our  own sort of flood detection running. I call it 
RID (rampaging intruder detection)


Problem with a captcha tyope solution is that we may need View to 
generate the random images. But our CGIs currently can only run under 
Core. Plus it disadvantages the visually disabled.

Each time I see an attack like this, it gives some more ideas for 
tightening up. Most of the time (like tis time) were one step ahead 
of the vandals.

Thanks -- we've been lucky so far.

REBOL isn't really a target  for hackers......Though it may become 
so if when it grows in popularity.

Here's an example of another Captcha - get the user to evaluate a 
rebol expression :)

They'd have to know REBOL first, though.

There are some very simple things we can do that would throw most 
of the stupid bots -- like a checkbox that has no other function 
other than to be checked.

The same bot  came back yesterday, via an anonymous proxy.

We are not alone -- this outfit is probing flaws in HTML forms on 
many (non-REBOL sites).

It would seem a reasonable security test on many websites to not 
allow sign-ups or feedback messages (etc) from anonymous proxies.

Neither rebol itself, nor me are able to find a rebol header in the 
file.

Thanks -- looks like there is a problem with the content-length header 
 (I thought there might be as I remember problems in testing when 
trying to get a value that all browsers were happy with. I think 
I've made IE happy while bemusing Firefox).


I've re-uploaded the script. It now doesn't set a content-length 
header. So, if you try again, it should work. But you may not get 
a download progress bar (because your browser isn't told the length 
in advance)

My apologies for the problem!

I tried doing a search for "probe system" in the mailing list because 
I know it is a topic that has come up a number of times and has been 
discussed.  After spending 30 minutes searching for relevant conversations 
but came up with essentially nothing even though I tried alternate 
search engines such as Google.  I'd like the ability to specifically 
search for code snippets.  I tried different methods of entering 
the search value but in every case I ended up with {Searched for: 
["probe" "system"]}  That's not what I want.  I want "probe system" 
How do I get it?

(there is a comment inside obj-browser.r mentioning Ammon might use 
for RIDE IDE...and it might have some methods to do what you need 
to do.)

With google, put a search phrase in quotes:
     "probe system"
that will seek the precise phrase.

To restrict google to a specific site, the syntax is:
     "probe system" site:www.rebol.org


To search only some URLs within a site, give it a characteristic 
part of the URL:
     "probe system" site:www.rebol.org inurl:thread


Note that Google has not indexed all the thread -- which is why we 
added the in-site search.

I don't want to "probe system" after a while, if you wait long enough, 
then it will print out but I thought that there were some optimizations 
made to make "probe system" start printing sooner or something to 
that effect.  It's not "probe system" that I'm interested in.  What 
I'm interested in is finding the conversations that I participated 
in several years ago on the REBOL ML concerning the problems with 
"probe system"  but none of the search engines I used could produce 
them.

Sunanda, "probe system" site:www.rebol.org inurl:thread did not return 
what I was hoping for.  By dropping the inurl: then I was able to 
come up with one relevant thread and the rest of the results were 
talking about probing in the system object like system/script/args. 
 I can't seem to get rid of the /whatever after system with Google. 
 I was hoping that the ability use quotes to search for a specific 
phrase instead of breaking it apart and searching for multiple values 
could be added to the rebol.org search engine.

Really? I thought the probe system hang was fixed a long, long time 
ago, no stack overflow for me here. Besides it is better to just 
use 
save %system.txt system

< I can't seem to get rid of the /whatever after system with Google.>

Yes -- Google doesn't understand that system/words (say) is a single 
item. While the REBOL.org seacth does. Ditto, the REBOL.org search 
recognises ! and ? as part of a word raher than discardable syntax 
-- so you can search for object? or datatype!

On the other hand, Google better handles phrases like "probe system" 
because its indexes include proximity information..
No one search is ideal.

So, Sunanda,  we don't have any proximity infomation and hence can't 
search for a particular phrase including space(s) like, "probe system"???

Correct Ammon. The REBOL.org search indexes threads (not individual 
messages) and does not include proximiy in its indexes.

So a search for [probe system] returns all threads that include either 
word.

Sunanda showed how I can use existing rebol.org technology (packages) 
to store and synchronize the comlib project files. But I am probably 
going to get a domain with Yeksoon.

Unfortunately, the library apparently doesn't have an index that 
would be condusive to such a search.  Oh well.  Wishful thinking...

Did you try the topic index -- it's intended to categorise what a 
thread is about rather than what words it contains:
http://www.rebol.org/cgi-bin/cgiwrap/rebol/ml-topic-index.r?i=p

How about a fulltext-search on rebol.org, as option?

You can download Grahams' ML archive application (or. at least you 
could until he started having trouble with 1and1's hosting).

Assuming it's still available, you get 70Meg of ML messages and it 
has a search that runs on your ow desktop.

well, I finally made contact with 1and1 tech support.  They said 
it was a billing issue and I have to contact accounts on Monday. 
 Now, what billing issue could there be with a 3 year free account 
that has 1 year to run ?

Sunanda,  I didn't try that.  I had assumed that search would use 
that index as a key.  Is there a way to search the topic index other 
than by clicking on a letter in the alphebetical list and relying 
on the browser's find functionality?

I can search for scripts using {in author "Volker"} and get what 
I'm looking for.  Perhaps adding a {in topic "blah"} capability to 
the ML search engine would be a good idea?

(It will take a long while, at least on dialup)

You could download all messages via the LDS interface...But I wouldn't 
recommend that: you need over 43,000 separate calls to do that.

REBOL.org doesn't offer a single packaged download of the ML archive. 
Maybe it should....

....Bui remember, that service is available from Graham; though it 
is off line until 1and1 sort out thier billing problem (scroll up 
to see Graham's message)

<< Is there a way to search the topic index other than by clicking 
on a letter>>
In theory:

http://www.rebol.org/cgi-bin/cgiwrap/rebol/ml-topic-index.r?i=probe

takes you straight to the entry (if any) for 'probe.  But that looks 
broken right now....I'll look into it.


Other than that, no.  Only about 80% of the threads are indexed in 
the topic index. When we get closer to 100%. we'll enhance the existing 
search to take account of the topic index.....so results will include 
and be prioritised by the topic index. 

Well, you could try, in Google:
probe site:www.rebol.org inurl:topic

But that (as with any search anywhere) is dependent on Google having 
the page indexed.

Back in February (!), Anton wrote:

<<By the way, something that I never really liked is the short rebol 
description at the top of rebol.org.>>

OLD TEXT: <<It's a messaging language. 'rebel' not 'ree-ball'.

 code is data and data is code. CGI? Yes. 40+ platforms. only 250K-500K.

 console or GUI. Yes, just a single file. commercial versions = same 
 core + ODBC, SSL, and more -

 shallow learning curve, amazing depth. keep IT simple. if you liked 
 "The Matrix"...>>

We may be slow, but we get there:   http://www.rebol.org

The banner text has been updated to ready pretty much what Anton 
suggested.

Thanks Anton!

Now, a quick challenge.  I'd like several banner texts that we can 
rotate to help keep the thing fresh, and to help unveil other aspects 
of REBOL.

Can we all chip in, please?
Either --- a 75 word (or shorter) snapshot of what REBOL is

or --- some bullet points that we can join to others to help paint 
a picture.

Thanks!

Maybe one way is for each banner to highlight a feature or strength 
of Rebol.
eg

Dialects are Rebol's greatest power. Dialects allows one to focus 
on expressing the domain specific language.


something along that line..but preferably something that SHOWS the 
benefit and not just tells it.

You are almost one step ahead of me there, Yeksoon.

The code is in place to make each of the banner statements clickable. 
That will take you to a page with an extended explanation.

Example: one of the soundbites might be "includes cross-platform 
assembler for 20x code speed-up". If clicked will take you to a paragraph 
on rebcode with links to more documentation.

But first, we need the 20 best soundbites :-)
Any offers?

The library list talks about a new project where there is a patched 
version of View that can make library calls.  I haven't downloaded 
it .. and hope that it doesn't violate RT's rights.

Graham, do you mean the dll called by the plugin? It is a (slightly) 
patched version of View which can be called as a library (the other 
way around). Regular View/Pro can make library calls.

Whether the plugin can do what it is said that it can has been a 
matter of much debate. How it does it is similar to how View does 
what it does, with little variation.

http://www.rebol.org/cgi-bin/cgiwrap/rebol/view-script.r?script=rebcall.r
seems to work through a port. I guess he runs some native app to 
do the real call.

I think the "dll" is an external programm. He opens a port and does 
everything through read-io/write-io

It really is a dll.

seems he adds a service to windows-network, or something. dummy open 
starts network, then he can connect.

Yeah, but that it'still runs in-process. He just hacks the network 
API so that inserts on that port are redirected to the dll. It seems 
a little sloppy to me.

His addin code is in a real dll. This dll runs in the memory space 
of the rebol.exe process so that no marshalling is needed. The invocation 
of the code in this dll is done through hacking the network access 
api. I don't know how this affects real network code - I am a little 
afraid to run this patch on my system, given this implementation 
strategy.

I would not do it. As i understand it, anything with tcp-access can 
then run code, even if sandboxed. Flash from a browser?

Banner text:

The banner text at REBOL.org is now clickable, and leads to a page 
intended to justify the various claims we make about REBOL:
http://www.rebol.org/cgi-bin/cgiwrap/rebol/expand-banner-text.r


Corrections and suggestions for added bullet points are most welcome.

The new clickable banner text concept is a great idea. I much prefer 
it to the original.


One comment on the rebcode description - "decimal order of magnitude" 
is probably accurate but seems to read like a "point something" (10%?) 
speed increase. It maybe better to just to say "order of magnitude".

As always, good work for implementing it Sunanda.

Does anyone know if you can embed the binary of a .dll into a rebol 
script, and load it?  or does it need to be a file?

You can extract the dll to a file on first run and then link to it, 
but you can't link to an embedded dll directly.

You could also rewrite the link loader to load a dll directly from 
a resource - see the source to BackOrifice for this amazing trick. 
This would require some rewrites to the native-code portions of REBOL 
though.

What are you trying to do? Can you get away with downloading the 
dll as a seperate file on first use?

Im just trying to hide the dll to keep things squeaky clean.. what 
i've done is embed the binary, and I'll write it to a file, when 
I'm finished, Ill just remove it.

When I deliver the app, it will have a single .exe

Nice idea -- someone is workng on it, but it may not happen for quite 
a while.

Sunanda, Can it be a manually maintained xml file until it can be 
automated?  I am doing that for my internal blog until I add the 
automation code to blog. r (that I expect Carl already has on his 
version).  I have a trigger for when a blog article is added to use 
"editor ftp://...../rss.xml"to make whatever additions that I want 
to expose via RSS.  It's suboptimal but I don't have any complaints 
from the people that they have to visit the pages to see What's New. 
 And since they weren't visiting regularly to poll for What's New 
anyway, if the RSS feed it updated a few hours later it is still 
an improvement.

Mike, I'd rather wait until the volunteer who is doing the RSS has 
added the code, rather than bypass their efforts.

Meanwhile, you have several ways of finding out what  scripts are 
new or changed on REBOL.org:
-- we'll sen doyu a tailored email

    http://www.rebol.org/cgi-bin/cgiwrap/rebol/cpt-update-profile.r
    See Section 4, Notices
-- you can peek at your unsent notices to get advanced warning:

     http://www.rebol.org/cgi-bin/cgiwrap/rebol/uns-display-notices.r?ml=yes&my=yes
-- You can use the LDS API:
       do http://www.rebol.org/library/public/lds-local.r

      probe lds/send-server 'list-updated-scripts [5]   ;; what's changed 
      in the last 5 days?
-- Just eyeball the script library home page:
     http://www.rebol.org/cgi-bin/cgiwrap/rebol/script-index.r

Sunada has access.  I put it there since the 4 scripts work together, 
its a very simple way to have a known working (tested) directory 
shared with others.  I don't know how to do this on the Library, 
and not wanting to mess up what is there already (in case someone 
sees a problem), so I stuck them in a simple place for review.

Thanks....IOS/Developer is working finw for me, and I can see your 
changes.

IOS/Developer is a bit of a forgotten Community gem (responses in 
the IOS group, if it still exists

Graham in the All group asked:

<<Is it not true though that you are limited in what you can do with 
rebol.org ? >>
Yes and no.

It's a shared server with the usual sort of quote limits, so that 
stops some of the wilder plans.

On the other hand, the limits are fairly generous. Right now, we 
do these sorts of numbers:
* 150,000 CGI executions per month
* 2.5 gigabytes of data served per month

* 70 megabytes of hard disk used (mainly for the Mailing List archive)


Of course some sort of bandwidth quota would exist with a dedicated 
server too -- that one comes from the ISP, not the box.
---- 

We've had the same problem most of us have had with servers at one 
time or another: REBOL going feral and failing to close tasks. That 
brought down the server once, and got us suspended for about 30 minutes 
until I beseeched forgiveness. Of course, that could happen with 
a dedicated server (no one quite knows why it happens).


Similarly, we get some leak-through from what are supposed to be 
the chinese walls; and that has caused the odd problem: mainly file 
permissions going askew.
---

As Peter said, the main limitation is manpower. There are several 
things under development at rebol.org right now; there is room for 
many other improvement projects too.

Regarding Carl's blog about hiding rebol scripts in web pages, I've 
always wanted a way to directly execute a rebol script from the library 
from my rebol console.  I used this trick of embedding a rebol script 
between [ .. ] in my offline mail reader, and wonder if it can be 
used in the library as well.

Use the download URL, eg:

do http://www.rebol.org/cgi-bin/cgiwrap/rebol/download-a-script.r?script-name=ascii-chart.r

well, o7o went one better with a small function with..  run °the 
script°

Just an idea ... how about subscribing the mailing list to a weekly 
library update?

This way one would see, that at least something is happening in rebol 
land >:-)

Sununda: Can you add a search box back to the front page, shouldn't 
need to go to a second page for a common action.

Allen -- problem is we have *three* separate searches:
-- scripts
-- mailing list
-- articles
Not easy to decide which should be the one on the home page

If you pick a section, then the right search appears on the LHS menu.

That's a possibility -- but you are only ever click away from the 
search page.

....which is exactly the same distance as having to click a radio 
box.

have the single search  as a wrapper around the other threethan on 
the search list return page show the hits in the different catagories

Sunanda, same number of mouse clicks: Yes. Same distance: No, you're 
a whole page away from what you're looking for. And if you don't 
know or don't remember that, you'll have to find the search before 
searching (pun intended ;-)

Convert the radio boxes into buttons and you've saved a mouse click. 
;)

Ingo, what's needed is some sort of REBOLbot that trawls the main 
REBOL websites, takes all REBOL-related RSS feeds (and perhaps chats 
to RIXBot), and produces a regular (hourly?) summary of what's new.....That 
way people can easily go where the action is.


If someone writes such a bot, then its summaries could be published 
anywhere. REBOL.org is one possibility; though REBOLCentral was announced 
to do that sort of publicising.


But first, someone needs to write the automated REBOL newshound. 
 Any offers?

Anton, you could set 
http://www.rebol.org/cgi-bin/cgiwrap/rebol/script-index.r

as your bookmarked page: then you'll always see first a summary of 
new scripts and have an on screen script search box.

:-)  I did wait a looong time to publish the Reversi script here 
(first version was made for /View 0.10) - It's strange to have a 
script who plays a game better than its author...

Thanks for helping. Posting a dot as the first post in the group 
(and sometimes even multiple posts w/ a single dot) is actually, 
quite an embarrassing experience, but at the moment it's the only 
way for me to sync to some of the groups.

Just seen Vince's reply as a result of Thor prodding the group to 
resync it. Thanks Thor.

<It's strange to have a script who plays a game better than its author>

Straneg, yes......But most of our scripts do things better or faster 
than we could ever manage. That's why we write them rather than employ 
slaves.

Graham -- nice idea, but I can see two problems:

[1] [web-public] is not a fine enough flag to control whether a whole 
group should be published publicly.

web-public needs a start date [or ranges of dates] to indicate which 
messages were posted while the group was not private.

Without that, we'd be republishing messages from a "secure and private 
world" (Altme's own description) without explicit permission from 
all the contributers.

Either that, or the publisher will need a list of which forums are 
public, and the date range(s) for which messages can be published.
That initial list would need to be constructed manually.

Thereafter, they could keep it up to date by watching for "[web-public]" 
to appear or disappear in group titles.

There is one other issue about archiving AltME chat - the conversations 
aren't threaded. You'd never know if a search took you to the beginning, 
middle or end of the information for which you were hunting.

As a result, it would be pretty tricky to index the archive.


The ideal would be for "useful threads" to be selected, tagged and 
archived. I can't see anybody being able to automate that at the 
moment.

What could possibly work is if there was an AltME feature that allowed 
any user to mark a few entries, give them a title and post them to 
the ML as an email.

Or perhaps somebody could come up with a standalone View script to 
do so.

I once tried to create a threading extension to the 'link chat. You 
had to click on a message, and then your message was tagged as an 
answer to the original.

And nothing is *ever* going to help to make public the nuggets in 
the original REBOL world and its successor REBOL2.

That's a great pity -- all that wisdom and enthusiasm hidden on the 
local hard drives of the people who happened to be members of those 
Altme world.

REBOLCentral is now nearly five months old as a project, and it was 
initiated partially to break the REBOL community's addiction to the 
"dark net"

I guess they have some exciting stuff about to emerge to address 
these issues. Yes?

REBOLCentral is now nearly five months old as a project,

  it would be best to not connect the date someone begins talking about 
  something with when the "project began."


Separate than this specific example, I notice this comes up a lot. 
 I guess phrases like "Bob has been talking about fixing that for 
years now" just seem very odd to me.  Unless Bob himself set the 
end date….or start date.


Anyway…funny enough, we were just talking this week about what off 
the shelf tools we plan to use for RebolCentral.  Over the past 5 
months, we built a non-profit wonderfully successful website for 
another cause that has generated great usage and created a sense 
of community.  Now that it is pretty stable, we will be moving on 
to RC.

there were some others.  I contacted TGD to get one that saves...hope 
they have done more with it...but there were some...gonna search 
my harddrive.  course, what I want to do can just be a one-line replace...just 
wanna get a good tool for hex crafted as I need.  I started on one 
a while back...

Was this a problem with makedoc?  Or does the library convert < and 
> entities automatically?

It's a glitch between REBOL.org's escaping of codes and makedoc's 
escping of codes.

Someone has to do it to prevent XSS attacks via contributed documentation, 
but not both of uss.

Part of the problem  is that almost all Dialect==>HTML converters 
assume they are there to produce browser-ready HTML. In tne real 
world that isn't always true: they are a step in a pipeline.
I'll sort it out soon. Thanks for reporting the problem.

Yeah, we are still fixing a bunch of these type of errors with our 
MakeDoc and QML to HTML converter.  It is very funny to make self 
referencing system.

It'd fidddly too as REBOL.org post-processes the output from Makedocpro, 
Makedoc2, eText, NicomDoc and a couple of other D2H converters.
And they all have different quirks.

And not all codes need to be escaped -- those within <pre> tags for 
example; though you can't always be sure you've isolated them if 
the generated HTML is not well formed.
And so on.
I'm surprised at times that there are so few glitches.

That should be okay now.

As Reichart notes, the fact you can embed HTML into a makedoc2 source 
is a great freedom for the end use, but can be a pain for anyone 
wanting to post process the generated HTML that it emits.

I was 2/10ths of a second ( or whatever ) faster :)

Petr -- somethng else you might want to look at:--- publishing a 
personal page on REBOL.org
You could use it to showcase whatever is important to you.
Npt many library menbers do as yet:
http://www.rebol.org/cgi-bin/cgiwrap/rebol/lmp-index.r

well, why not to discuss it here - I found curious Carl's request 
to follow reboltak.com. It is interesting aspect - with various products, 
I am really used to on-line phorums (most of them require registration 
anyway). I am less and less willing to subscribe to some kind of 
MLs. OTOH, the interface is slow. Currently we've got ajax as a cure 
.... to get back where AltME is for few years already ....

I would welcome more open AltME design - it is a good app. But maybe 
on rebol.org we could list on-line rebol related worlds, and provide 
free accounts, and maybe later on we can come up with mechanism, 
how to feed some info into db .... to be searchable

As Reichart says somewhere, Altme REBOL3 is great for quick problem 
solving. But you have to find it first. Like most of the REBOLsphere 
it is close to invisible

(eg eFishAnt's recent serach for a REBOL hex editor. They exist: 
but can you find them?)


REBOL.org is just about the only exception to the invisibleness of 
the REBOLsphere -- try looking in Google for
   script library
as an example.

I recall someone had a program that indexed all the rebol scripts 
on the internet.