1  ...  6  7  8  9  10  11  ...  36  |   Order by default

Apr 25, 2004, 5:46pm

This is a graphic of what is happening:

http://www.xs4all.nl/~xelag/sdk_uni_error.gif


Alex

Apr 28, 2004, 12:25pm

This is the newest xelagot topic: Ghosts
http://www.imatowns.com/xelagot/xlgtopicghosts.html
It applies to all bots and browsers, so I make it generally available.

For particulars about xelagot, see What's New
http://www.imatowns.com/xelagot/xlgwhatsnew.html

Alex

Apr 28, 2004, 8:50pm

I'm glad Linux is getting some attention. But your reply does not
address the point I made :)

Alex

[View Quote] >Maybe I should spill the beans... But during some chat with Will (999),
>he mentioned a new linux world server that was a total rewrite of the
>first, including support for MySQL. I'm not sure about any of the
>details, or any estimated release date. He just said he had been working
>on rewriting the server and that it supported MySQL.
>
>-Jeremy
>
>
>
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>Jeremy Booker - Owner / Webmaster
>JTech Web Systems
>www.JTechWebSystems.com
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>"Therefore do not worry about tomorrow, for tomorrow will worry about
>itself. Each day has enough trouble of its own." -Mathew 5:34
>
>
>
>
[View Quote]

Apr 29, 2004, 8:21am

A rather rude reply, jerme.

It was the wrong thread, that's all. Linux is a separate problem,
there's a lot written here about it and it deserves its own thread :)

Alex

[View Quote] >dude... ease up a little... take a step down off the pedestal.
>
>I had relevant information about the server...
>
>It's being rewritten. Maybe they'll fix your ghost bug in the rewrite??
>
>-J
>
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>Jeremy Booker - Owner / Webmaster
>JTech Web Systems
>www.JTechWebSystems.com
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>"Therefore do not worry about tomorrow, for tomorrow will worry about
>itself. Each day has enough trouble of its own." -Mathew 5:34
>
>
>
>
[View Quote]

Apr 30, 2004, 1:58am

I did not chose to leave them out of my report, I mention bots
explicitely. Bots do not log in with a citizen number, they actually
log in with citizen privs. If you check world logs, you will see the
following when an entity logs in:

for tourists, a cit 0 and privs 0
for cits, a cit x and a priv y
for bots, a cit 0 and privs x

notice also that bots log in with the priv password.

In the SDK events, because AW did not at first mention privs, avatar
adds events for bots use the cit field for its privs, and of course so
do the bot user interfaces. It is inconsequent, but presents no
problems, as one can distinguish a cit from a bot by its name (the
famous [...] thingies)

I tried to keep it readable in the pages... for technical and
especially non-technical people. If you have a suggestion on how I
can rephrase it, welcome: xelag at 3dee.nl

Alex

On 29 Apr 2004 22:28:29 -0400, "tony m"
[View Quote] >I'm not sure if you thought of it before, but bots are the only other entity which can be represented by your citizen number concurrently while you're logged in. So your second case does not consider any existing bots the citizen may be running in that world.
>
>Just a thought -- maybe you do know about it, but chose to leave it out of your report.
>
[View Quote]

Apr 30, 2004, 2:12am

On 29 Apr 2004 20:23:38 -0400, "ep0ch" <deltawolf at deltawolf.com>
[View Quote] >Thanks Alex, I will work on adding a "ghost checking" system to my bot.
>Shouldn't take more than 5 minutes with the information you've provided.
>
>Cheers,
>Ep0ch

Oh it will... the fact is, unless this is solved at world server
level, we will not get a watertight solution. If you start analising
the case, you will see a few problems:

case 1: how do you distinguish the live from the ghost? how do you
make sure that when you get an avatar delete, you delete the ghost and
not the live (if both are present)? In fact, you need to have a
counter, counting up and down the number of adds and deletes per
session number... In these cases, the SDK also gets confused, and chat
does not propagate.

case 2 and 3: which session number is valid? the last one received,
or the previous one? Bots are mobile, and entities can come in and
out of reach. Since for the world server, both the ghost and the live
are real, you can get avatar add and avatar delete for both session
numbers.

Bots in global mode should be able to solve these problems easier, but
it's not a 5 minutes work, I've been at it for months :)

Alex

Apr 30, 2004, 2:36am

What I forgot to mention is that, in the very begining of the SDK, we
did not even get the citizen number in the avatar add and delete event
(and no privs of course) so you could not even know who owned a bot.
And you could not even know what citnum XelaG or E N Z O had, except
by querying the universe server for the citnum of a name.

This was for me the reason to introduce the xelagot greet. At least
xelagots could identify themselves and their owners to other xelagots.

It took a while for AW to realise the cit was important. Botmakers
from the start pointed out to AW that cit, privs and IP (for bots with
eject rights) should be sent to bots directly with the events
concerning entities entering a world... after a while, first cit and
then much later privs were added, IP is still a matter of querying the
world server. Seems that what AW itself does not need for its own
projects is of no concern: no vision in this sense, and certainly they
do not consult other programmers who do have some sort of experience
about all this.

Alex

[View Quote] >I did not chose to leave them out of my report, I mention bots
>explicitely. Bots do not log in with a citizen number, they actually
>log in with citizen privs. If you check world logs, you will see the
>following when an entity logs in:
>
>for tourists, a cit 0 and privs 0
>for cits, a cit x and a priv y
>for bots, a cit 0 and privs x
>
>notice also that bots log in with the priv password.
>
>In the SDK events, because AW did not at first mention privs, avatar
>adds events for bots use the cit field for its privs, and of course so
>do the bot user interfaces. It is inconsequent, but presents no
>problems, as one can distinguish a cit from a bot by its name (the
>famous [...] thingies)
>
>I tried to keep it readable in the pages... for technical and
>especially non-technical people. If you have a suggestion on how I
>can rephrase it, welcome: xelag at 3dee.nl
>
>Alex
>
>On 29 Apr 2004 22:28:29 -0400, "tony m"
[View Quote]

Jun 7, 2004, 8:36pm

Collision is detected frame by frame by the rendering engine of each
individual browser and is cpu intensive - it is all done client side.
This means, there is absolutely no internet traffic involved (except
for the downloading of objects and updating of avatar and object
positions). It would be absurd to have all this activity broadcasted
to all, even updates of avatar position are sent at fixed intervals,
skipping older ones :)

The difference is that object click/select or avatar click require a
conscious action from a user, whereas collision is a frame by frame
calculation. When you bump into something, it is your browser that
decides you have bumped., and acts accordingly on you. The
consequences are broadcasted as your new position, avatar state, etc,
at due intervals (15 per sec is maximum if the world allows it, 1 per
secong is default). But the actual bump, which depends on your fps
and involves multiple objects, can not be broadcasted.

Alex

On 6 Jun 2004 19:34:10 -0400, "r i c h a r d"
[View Quote] >An event raised when ever someone collides with an object with a name or
>with out a name.
>
>The event would then send to the sdk object number and basic properties as
>well as the avatar session name and location.
>

Jul 4, 2004, 9:46pm

C++ 3.6 SDK build 38 has now been lying on AWI's computer for 3 weeks.
When I was aware of this, I waited one week to upload my 3.6 bot, that
had added SQL functionality. The 3.6 SDK never came, so I uploaded my
3.6 bot without the 3.6 code.

When Roland Vilett was in charge of the programming team, we always
received a beta SDK for the coming version on time, while the beta
browsers/servers were being tested. Bot programmers had the
opportunity to test and send their comments. I always prepared
special new features for xelagot with each major upgrade, this was a
great way of coordinating enhancements to the AWI system.

This has changed drastically. Since Shamus, Will and others are in
charge, we receive nothing: 3.5 and 3.6 follow the same pattern. This
is therefore not only a problem for VB programmers (responsible: Will
= 999) but for all programmers (responsible = Will, who does the SDK
upgrades). The responsability may not lie totally by Will, it is in
the end the provider's responsability: AWI.

I'm very sorry to have to say so, but as long as AWI team denies
communication to its external programmers and users, we can not
reliably work to enhance their product. They don't even reply to my
emails.

Alex

On 4 Jul 2004 17:02:23 -0400, "ksg" <kingsmallguy at my.activeworlds.com>
[View Quote] >Simple answer probly not, ive waited and waited for a while and no new
>versions came
[View Quote]

Jul 13, 2004, 1:39pm

The xelagot man doesn't know really, but since it is a string, I guess
you can add the currency symbol:

€ 9,99

(with the internationally used comma :)

Alex (XelaG)
Xelagot Man

On 12 Jul 2004 14:48:16 -0400, "ksg"
[View Quote] >erm that Xelagot guy will know, sorry ive forgot his name
[View Quote]

Jul 13, 2004, 10:20pm

Wrong.

I was brought up in Uruguay and Argentina in the 50s-60s. We use the
comma there. When I came to live in Europe (Netherlands) in the 70s
up to now, the comma we use here too. When I was in the USSR in the
70s, they used the comma there too. The strange Europeans you mention
are probably strange to you because of recent events: no more french
fries in the USA!

I personally prefer the decimal point. But what the heck :)

Alex

[View Quote] >comma isn't internationally used >_< only them strange europeans use it (by
>strange, i mean the germans and french!) eveyrone else uses .s! :)
>
> -SWE
>
[View Quote]

Jul 14, 2004, 9:03am

Nice :)

[View Quote] >btw this is with no offense to Alex, :) ur great, but hey sometimes yaw
>gotta know when to suck it up and let sum1 else be right (even if their
>completely wrong)
[View Quote]

Jul 14, 2004, 10:57am

http://united-states.asinah.net/american-encyclopedia/wikipedia/s/si/si.html

[View Quote] >There is, of course, also an ISO norm for this, it is the 31-0 as of
>1992. It applies to all technical documents (when you write privately,
>you can, naturally, as well use eg. the assyrian stone carving symbols
>or Roman digits):
>
>The decimal seperator is the COMMA and has to be used in all ISO
>documents. However, in _english_language_ texts, it is also ALLOWED to
>use the POINT (unless they are english ISO documents).
>It is, BTW, NOT allowed to start a digit with the decimal character, so
>values like .5 or ,5 are incorrect (correct is 0.5 or 0,5).
>
>The thousands seperator, however, is neither comma nor point nor upper
>comma (as in Switzerland for example), but a SPACE, eg. as in 123 456
>789.
>
>In this regard, threemillion fourhundredthousand twohundredtwentyseven
>and 87/100 would be written as:
>
>ISO international = 3 400 227,87
>non-ISO english language = 3 400 227.87
>
>FOUR digit years, page references and such shall be written without
>seperators, eg. 2004-07-14 or Page 1234.
>
>:-)
>
>
>
[View Quote]

Jul 14, 2004, 11:37am

The problem becomes accute really when parsing strings. For xelagot
and AW related applications, I decided to adopt the UK/US convention
of the dot as decimal separator, because it is the standard used by
AW, also for text SEQ files and RWX files. This caused, in the
begining, problems for users in many countries, where the computer
locale uses the comma as decimal separator: numbers in files were
being read wrongly and numbers in strings were being parsed wrongly.
Luckily, Delphi has a solution: your application can define it's own
decimal separator, so all my programs start with

Application.UpdateFormatSettings := False;
DecimalSeparator := '.';

This solved the problem for me :)

I do not use the thousands separator at all in my programs. Using a
space would be very difficult, as it is used mostly as a normal word
separator. The same goes for the comma.

Alex

[View Quote] >There is, of course, also an ISO norm for this, it is the 31-0 as of
>1992. It applies to all technical documents (when you write privately,
>you can, naturally, as well use eg. the assyrian stone carving symbols
>or Roman digits):
>
>The decimal seperator is the COMMA and has to be used in all ISO
>documents. However, in _english_language_ texts, it is also ALLOWED to
>use the POINT (unless they are english ISO documents).
>It is, BTW, NOT allowed to start a digit with the decimal character, so
>values like .5 or ,5 are incorrect (correct is 0.5 or 0,5).
>
>The thousands seperator, however, is neither comma nor point nor upper
>comma (as in Switzerland for example), but a SPACE, eg. as in 123 456
>789.
>
>In this regard, threemillion fourhundredthousand twohundredtwentyseven
>and 87/100 would be written as:
>
>ISO international = 3 400 227,87
>non-ISO english language = 3 400 227.87
>
>FOUR digit years, page references and such shall be written without
>seperators, eg. 2004-07-14 or Page 1234.
>
>:-)
>
>
>
[View Quote]

Jul 14, 2004, 11:53am

Another problem with parsing numbers in strings is the E, as for
example used in AW coordinates. In scientific notation, the E is part
of a number, for example, 1.2345e-10. So what I do when parsing
coordinates is test for the presence of E and e at the end of a number
before extracting the number.

Alex

Jul 14, 2004, 12:47pm

The fact, SWE, is that there are two main systems in use for the
decimal separator. As programmer, you need to take this into account,
or your programs may only work on some locales and not on others. I
shared my experience with Delphi, maybe VB and C programmers can now
tell us how they solve this?

Alex


[View Quote] >well, i lived in england, then moved to libya, and they use commas in both
>places! all arab countries use commas! :) plus, the chineese and indians
>(actually, i'm not sure about the chineese and indians) so, that's more then
>half the world, hahaahaha!
>
> -SWE

Jul 18, 2004, 8:57am

On 18 Jul 2004 01:28:25 -0400, "tony m"
[View Quote] >No, bots can't detect if someone is a caretaker.

Two particular exceptions are:

1) any bot can detect #1 citizen (citnum or privs), #1 is CT
everywhere.

2) a CT bot knows its own citnum (which is technically a priv), so it
can know that its login owner is CT. Also that anyone wearing this
priv is CT (citizens and bots).

>
[View Quote]

Jul 22, 2004, 3:39am

The terrain specifications are not modifiable locally by a browser.
The only thing the browser can do is download the terrain (ascii) file
and render it. Call yourself lucky if that succeeds :)

In fact, AWI has made a big blunder in its hole code (no pun implied).
Originally, it reserved four of the 256 byte encodings for special
effects. One of them was to become the hole in terrain. The net
effect was that one would have textures from 0 to 62 available in four
rotations (terrain0 to 62 = 63 terrain textures x 4 possible rotations
N W S E, 252 combinations), the non used "terrain63" was the reserved
space with 4 possibilities for special effects, one of them to be used
for hole. Due to bad programming, another byte encoding was used for
the hole, not using one of the 4 reserved encodings, but the one
corresponding to texture62 (a legal texture) fourth rotation pointing
east (try it and you get a hole). I pointed out this mistake to
Shamus, and he called me bluff in the beta ng (by lack of decent
answer), his ignorance and utter arrogance, not mine, any programmer
with a bit of mind can check this. A good programmer will acknowledge
her/his mistakes. The original 3.3 code (and mistake) was from Roland
Vilett, understandable because of his inminent demise, but it was
still relatively easy to correct when discovered early in 3.4 beta, a
few months after the introduction of 3.3. It is now virtually
impossible to do so, because terrain with holes has been widely
implemented and a change in the hole code would be difficult to
correct. So we are stuck now with one texture less: 62 instead of 63
textures for terrain.

The world still goes round though, even if the summer is crappy in
Amsterdam :)

Alex


On 21 Jul 2004 23:30:54 -0400, "ubermonkey"
[View Quote] >Unfortunately there's just no way to make a terrain cell a hole for only one
>user; the change would impact anyone in the world. While a bot could be made
>to create this effect, it would be totally broken if there were multiple
>users in the world, assuming they didn't all intend to move as a group. If
>you only intend one user to see it at a time, then a relatively simple bot
>could be made to handle this, however there's no decent way for said bot to
>detect the user's current visibility setting (unless you set min vis to 100
>and just assume 100m vis for everyone). If a min vis of 200 was possible,
>then there would be no confusion at all here, though it might cause some PCs
>to die.
>
>Okay, hold on -- it is actually possible, it's just something you couldn't
>accomplish unless you were a universe owner or had way too much money to
>waste (err, I guess those are effectively the same thing).
>
>The only way I can see this working for a group of people in a world at the
>moment is complex and bizarre; it requires you to have as many identical
>worlds as you intend to have users. Each user is sent to one of the empty
>clone worlds (up until none of them are available, at which point they are
>automatically ejected), where the terrain effect in question functions with
>no issues. The strange part is that we then use additional bots to mimic the
>movements and chat of the users in the clone worlds; the effect should be
>totally seamless (assuming all worlds and bots run from the same PC, or at
>least the same LAN), aside from the fact that everyone you talk to will have
>[these] around their names. Writing the bots would actually be the easy part
>here, as the biggest concern is obtaining all those clone worlds and bot
>rights.
>

Aug 20, 2004, 5:07pm

I sent you a zip by email.

Alex

On 16 Aug 2004 15:23:41 -0400, "strike rapier"
[View Quote] >Hi Everyone,
>
>Would anyone be able to post a list of the different (constantly changing)
>attributes for the different SDK builds for what resolves to what, such as
>World Ratings, Object Limits etc, and what they appertain to? I admit that I
>have no idea what these have been in the last few builds and I am hoping
>someone has been keeping a closer eye on them than I.
>
>- MR
>

Aug 20, 2004, 4:42pm

The user list, if enabled (it is always enabled for cit #1 bots) must
be refreshed at least once every 30 minutes if the bot calls it. The
reason is that departed users are kept in the universe database for
only 30 minutes. If the refresh interval is longer than that,
duplicate entries will appear for users that log off and back in, as
the user "delete" (AW_EVENT_USER_INFO with aw_userlist_state = 0) will
not be sent to the SDK because the universe database has removed this
user from its "departed" database table.


Thanks, Chris, for the info.
Alex

Aug 21, 2004, 11:22am

The SDK docs have now been updated with this information, thanks Chris
:)

http://www.activeworlds.com/sdk/aw_user_list.htm

Alex

[View Quote] >The user list, if enabled (it is always enabled for cit #1 bots) must
>be refreshed at least once every 30 minutes if the bot calls it. The
>reason is that departed users are kept in the universe database for
>only 30 minutes. If the refresh interval is longer than that,
>duplicate entries will appear for users that log off and back in, as
>the user "delete" (AW_EVENT_USER_INFO with aw_userlist_state = 0) will
>not be sent to the SDK because the universe database has removed this
>user from its "departed" database table.
>
>
>Thanks, Chris, for the info.
>Alex

Dec 4, 2004, 3:46am

It's in my xelagot help files: Ghosts

http://www.imatowns.com/xelagot/xlgtopicghosts.html

Alex

[View Quote] [View Quote]

Dec 25, 2004, 11:49am

World disconnects do not cause a change of session number, only
universe disconnects do that. Yes, I have encountered this sort of
things many times in the past, I still don't have a clue if this has
been (partially) fixed. See
http://www.imatowns.com/xelagot/xlgtopicghosts.html
written about a year ago.

Alex

Jan 26, 2005, 2:43pm

Thank, Andras. Can the http://www.activeworlds.com/sdk/ pages be
updated too?

Alex

[View Quote] >Hi folks,
>
>SDK build 40 can be downloaded from http://www.activeworlds.com/sdk/
>The Linux counterpart is available at http://www.andras.net/linuxsdk.html
>
>Change history can be seen at http://www.andras.net/linuxsdk.html

Jan 29, 2005, 6:46pm

Hehe :) Well it's probably not Andras' task to update the pages to
build 40, but it should be coordinated someway.

Update needed:

http://www.activeworlds.com/sdk/whatsnew.htm
http://www.activeworlds.com/sdk/aw_data.htm
http://www.activeworlds.com/sdk/aw_data_set.htm

Has been updated to build 40::

http://www.activeworlds.com/sdk/download.htm

No idea if other pages need updating. But Andras' page mentions build
41 for Linux, probably not applicable to Windows
http://www.andras.net/linuxsdk.html

I will wait in any case for the pages updates before updating my
Xelagots, I don't use aw_data. If nothing else has changed, seems
unnecessary to hurry and update of my bots.

Alex


[View Quote] >That would go against the master plan of the SDK pages. They plot to
>keep developers in the dark, it consumes them.
>
>-Joe

Apr 17, 2001, 10:00pm

hehe, you'll see how many idiots (well not really) will telegram instead of
emailing... its still an enormous struggle to get ppl to email. I wish AW
would implement a telegram stop, like they have for join.

[View Quote]

Jun 2, 2005, 9:07am

What happens is, say you have sequence number 25 for a given sector of
8x8 cells, and the world has 27 as latest sequence number in that
sector. Since the time you (hmm... your bot) got the # 25, two cells
have changed (or one cell twice). When you query and send sequence
number 25, the world will then send you the full updated information
only for the cells that have changed, so you only need to delete the
cell object info (in CELL_BEGIN) of the cell being updated, and refill
it with the new info in each CELL_OBJECT. But if you send sequence
number 0, the whole sector will be sent again. That is what xelagot
uses, it gives a choice of "use cache", loading its cache for the
sector and only updating changed cells, or not using cache and
updating the whole sector.

[View Quote] [View Quote]

Jul 5, 2005, 2:51pm

Mark, I use the object number method (combined with X Z coordinates)
to identify the object in question, and don't rely on the FIFO method
for this. For object_add and object_load, the object number should
suffice becaused it has been issued locally by the SDK. For
object_delete, it is safer to to use three parameters: object number,
X and Z.

For some other callbacks (citizen_by_number for example), I wait for
the callback before issuing another query, so FIFO does not apply
either.

For some operations you MUST wait for the callback before re-issuing,
for example, cell_next. In this case, the information is sent
serially, you rely on the data in the callback to decide whether to
re-issue cell_next or not, etc.

FIFO may be valid for some same-operation calls, seems logical, but
this is not documented nor garanteed.

Alex

On 4 Jul 2005 13:38:01 -0400, "Strike Rapier" <markyr at gmail.com>
[View Quote] >I am making a queue system for my object events, that uses function pointers
>to inform different sections about things.
>
>I am just using FIFO at the moment, i may have to change over to the object
>number method.

Aug 6, 2005, 10:16am

You could encode the filename using for example the encoding technique
for URL strings. That's what xelagot does when saving files for
individual worlds.

unchanged: "a" to "z", "A" to "Z", underscore "_", numbers "0" to "9"
spaces are replaced by "+" symbol (or by %20)
all other characters are formed by a "%" followed by two characters,
standing for the hexadecimal value of the character. For example, "%"
is encoded as %25.

To read back, apply the converse system: (1) change all "+" to spaces,
(2) then any time the "%" char appears, check if the following two
chars form a hex number, convert it to the corresponding character.

Alex.

On 4 Aug 2005 14:43:03 -0400, "Flashbstudios" <staff at auburnflame.com>
[View Quote] >I am wondering how this is possible since "s arnt allowed with filenames
>on my computer can anyone tell me how (I am using the VB Wrapper Build
>41 btw...)

Aug 6, 2005, 10:25am

I use these two function (in Delphi ppascal). Note that in Delphi,
string chars start at 1, I think in C they start at 0.

function URLEncode(u: string): string;
var
i: Integer;
const
okchars =
'1234567890_ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz';
begin
Result := '';
for i := 1 to Length(u) do
begin
if Pos(u[i], okchars) > 0
then Result := Result + u[i]
else if u[i] = ' ' then Result := Result + '+'
else Result := Result + '%' + IntToHex(Byte(u[i]), 2);
end;
end;

function URLDecode(u: string): string;
var
i: Integer;
begin
Result := '';
i := 1;
while i <= Length(u) do
begin
if u[i] = '%' then
begin
try
Result := Result + Char(StrToInt('$' + u[i + 1] + u[i + 2]));
except
end;
inc(i);
inc(i);
end
else if u[i] = '+' then Result := Result + ' '
else Result := Result + u[i];
inc(i);
end;
end;


[View Quote] >You could encode the filename using for example the encoding technique
>for URL strings. That's what xelagot does when saving files for
>individual worlds.
>
>unchanged: "a" to "z", "A" to "Z", underscore "_", numbers "0" to "9"
>spaces are replaced by "+" symbol (or by %20)
>all other characters are formed by a "%" followed by two characters,
>standing for the hexadecimal value of the character. For example, "%"
>is encoded as %25.
>
>To read back, apply the converse system: (1) change all "+" to spaces,
>(2) then any time the "%" char appears, check if the following two
>chars form a hex number, convert it to the corresponding character.
>
>Alex.
>
>On 4 Aug 2005 14:43:03 -0400, "Flashbstudios" <staff at auburnflame.com>
[View Quote]

1  ...  6  7  8  9  10  11  ...  36  |   Order by default