Gesture problems (Sdk) - Awportals.com

Thread

Board Archives

	Bots
	Community
	General Discussion
	Sdk
	Wishlist
	Author Search

Site Features

	Citizens List
	Badges & Achievements
	Events Directory
	News and Articles
	Worlds List
	About Alphaworld
	Search Object Paths

Activeworlds Support

	Download Activeworlds
	Quickstart Tutorial

Historic Archives

	Truespace Archives
	Newsgroup Archives

Gesture problems (Sdk)

Gesture problems // Sdk

roland vilett

Dec 17, 1998, 6:11am

Sorry. It's just that I thought he was giving *me* a hard time about the
docs...I'm a bit over-extended here trying to keep everything going at once
and sometimes I can get defensive.

-Roland

[View Quote]

edward sumerfield

Dec 17, 1998, 11:34am

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<body bgcolor="#FFFFFF">
[View Quote] <blockquote TYPE=CITE>
[View Quote]

<blockquote TYPE=CITE>Andras Sarkozy wrote:
<blockquote TYPE=CITE>The same you posted some times ago and here. aw_wait(1000)
is there and nothing really else.
 The loop is here:
   for (int i=0;i<36000;i++)    // 10 hours lifetime
     {
      (AWWait)(1000); // 1 sec only
         if (reset_gesture &&
reset_gesture < GetTickCount ())
         {
    /* time to set our gesture back to 0 again */
       (AWIntSet) (AW_MY_GESTURE, 0);
       (AWStateChange) ();
       reset_gesture = 0;
         }
     }
  
Edward Sumerfield wrote:
<blockquote TYPE=CITE>Because Edward writes very inefficient code?
Which of my bots were you running?
Most of them have one second poll (aw_wait(1000)) so that I can use
timers that are accurate to one second. I have not looked at how much CPU
this takes up.
Andras Sarkozy wrote:
<blockquote TYPE=CITE>If we are at this topic :
 My test involved 9 instances of the bots by Edward and it consumed
%66 of my Dual 300 MHz P2 processor. Anyone know any explanation for it?
Roland Vilett wrote:
<blockquote TYPE=CITE> Again, yes, this is correct. 
But the beginning of this thread I believe was about the case of someone
wanting an avatar to repeat the same gesture over and over again in response
to chat events.  In this case changing AW_MY_GESTURE specifically
to zero is required in order to ?reset? the gesture.  Of course you
could set it to some other gesture instead of zero, but then you would
have more than one gesture going on. -Roland
<blockquote
style="BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: 5px">Edward
Sumerfield<ESUMERFD at POBOXES.COM> wrote in message ?<a href="mailto:3676DF62.65421F6D at poboxes.com">3676DF62.65421F6D at poboxes.com</a>?...You
De Man Andras.
So this is a documentation point to Roland then to clarify that that
it is not the change to zero but the CHANGE of the gesture that is important.
Andras Sarkozy wrote:
<blockquote TYPE = CITE>I tested this case - to change to a different than
zero gesture will initiate properly the new gesture, while resetting to
zero will not cause change in the avatar movement. It is needed only for
the SAME gesture than the last one..
  
Edward Sumerfield wrote:
<blockquote TYPE = CITE>Just one more question for you Roland.
The browser is deciding on the gesture to initiate if the number from
the world server is different from the gesture it just performed?
We have said that we need to reset the gesture to zero 5 seconds after
the initial gesture is sent. Could that rule be rephrased as ?to initiate
a sequence of gestures, each successive gesture must be DIFFERENT from
the last.?
The difference between these statements is of coarse that instead of
zero we could set gesture 4, then 5 seconds later, set gesture 3, then
5 seconds later back to gesture 4 again. Would this sequence cause the
browser to correctly initiate each sequence?
Roland Vilett wrote:
<blockquote TYPE = CITE> Whether
or not their would be increased lag due to an increased number of users
would depend on the outgoing bandwidth of the server.  Servers with
a high user load and low bandwidth would experience poor response time
for all transactions, position updates included, as you would expect. 
If the server has adequate bandwidth available, though, I don't think you
would ever notice a difference in propagation delay whether there was 1
user or 100 users in the world. I
chose 5 seconds because that's what the AW browser does...5 seconds after
signaling a gesture, it returns to 0. I'm actually not sure whether a shorter
timeout would cause the gesture to terminate prematurely or not...my suspicion
is not, since gestures can overlap (e.g. you can start a wave before completing
the angry sequence, and the animation code will attempt to do both at the
same time.) To know for sure I'd have to dive deep into the avatar animation
code, which is not something I particularly relish (it isn't my code.) 
A simple experiment with a bot though would say for sure...-Roland
<blockquote
style="BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: 5px">Edward
Sumerfield<ESUMERFD at POBOXES.COM> wrote in message ?<a href="mailto:367576AC.18445D02 at poboxes.com">367576AC.18445D02 at poboxes.com</a>?...I
see. Its a perspective change from me. I keep making the mistake of see
actions as changes in the server but really they are changes in other browsers
propagated by a 1 second polling server.
You know this means I am going to have to rethink my class library infrastructure
again. Doh.
A performance question for you. You say that the server updates the
browsers every second. Is there any kind of propagation delay noticed as
the number of browsers increases or does the max logins per world keep
this to a minimum.
You recommend resetting the gesture after 5 seconds. A gesture sequence
takes a certain amount of time to run in a browser. If I set to gesture
to 5 and reset to 0 in two seconds will it stop the currently running sequence?
Is there some timing advantage to noting long gesture sequence compared
to short sequences and varying the reset time interval appropriately?
Roland Vilett wrote:
<blockquote TYPE = CITE> No, this
won't work.  In order for the gesture to be reset, the other browsers
have to see the change.  Remember that multiple calls to aw_state_change()
are not propagated out to other clients more than once per second. 
Your code will simply change your state in the server, and then immediately
change it back to the way it was before.  To everyone else in the
area, you will appear to have maintained gesture #5 all along.The
correct way to do this is to add code to your bot to always change the
gesture back to 0 again a few seconds after it was changed to 5. 
The easiest way to do this is to note the current time when you set the
gesture to 5 (the Windows system call GetTickCount() returns a handy time
value in milliseconds), and then in your main event loop periodically check
this saved value against the current time.  You will need to modify
your current call to aw_wait() since your program now needs to take action
in addition to simply responding to events. The
relevant code changes could look something like this: #define
FIVE_SECONDS    5000 unsigned
long reset_gesture; void avatar_chat() {
   int rc;
   printf(?chatting\n?);
   aw_int_set (AW_MY_GESTURE, 5);
   if (rc = aw_state_change ()) {
     printf (?Unable to change state (reason %d)\n?,
rc);
     exit (1);
   }  reset_gesture = GetTickCount() + FIVE_SECONDS;} /*
replace aw_wait(-1) with this: */ while (!aw_wait
(1000))  if (reset_gesture ?? reset_gesture ? GetTickCount ()) {  
/* time to set our gesture back to 0 again */   aw_int_set (AW_MY_GESTURE,
0);   aw_state_change ();   reset_gesture = 0; 
}  Edward Sumerfield<ESUMERFD at POBOXES.COM> wrote in message
?<a href="mailto:36756086.CBC209DB at poboxes.com">36756086.CBC209DB at poboxes.com</a>?...
<blockquote
style="BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: 5px">Cool.
So If I do
    aw_int_set(AW_MY_GESTURE, 0);
     aw_state_change();
     aw_int_set(AW_MY_GESTURE, 5);
     aw_state_change();
It will work every time?
Walter Knupe wrote:
<blockquote TYPE = CITE>You have to reset the gesture to 0 and then back
to 5 to have the others
 pick it up again. The resetting to 0 is done automatically by the awbrowser
 after you clicked on a gesture begin
Walter
Edward Sumerfield schrieb in Nachricht ?36748FD4.84162BA5 at poboxes.com?...
 ?The following source is supposed to initiate a gesture every time
it hears
 ?some chat message. However, it only does it once after the program
starts
 and
 ?never again. What am I doing wrong?
?void avatar_chat() {
 ?  int rc;
 ?  printf(?chatting\n?);
 ?  aw_int_set (AW_MY_GESTURE, 5);
 ?  if (rc = aw_state_change ()) {
 ?    printf (?Unable to change state (reason %d)\n?,
rc);
 ?    exit (1);
 ?  }
 ?}
 ?</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
</blockquote>

</body>
</html>

edward sumerfield

Dec 17, 1998, 12:06pm

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
You are doing a great job Roland. Keep it up.
[View Quote]

Roland Vilett wrote:
<blockquote TYPE=CITE>Sorry.  It's just that I thought he was giving
*me* a hard time about the
 docs...I'm a bit over-extended here trying to keep everything going
at once
 and sometimes I can get defensive.
-Roland
Edward Sumerfield wrote in message <3678399A.C087FB16 at poboxes.com>...
 >Don't be hard on the guy Roland. He is only trying to help with specific
 >phrasing of your pending documentation. He doesn't deserve to be
 reprimanded. As
 >we work through this "work in progress" we have all had to deal with
the,
 >expected, inaccuracies. This is not a problem except that we are all
 interested
 >in improving things for the future.
 >
 >Roland Violet wrote:
 >
 >> So, ummm, what exactly is your point, Canopus? That the documentation
is
 >> incomplete? Thanks, but I already knew that. Like the SDK itself,
and
 Active
 >> Worlds in general, the documentation is a work in progress, and
it will
 >> improve with time
 >>
 >> -Roland
 >>
 >> Canopus wrote in message <3677002D.6D2E7673 at ix.netcom.com>...
 >> >Thanks, Roland. I hope that all your clear and helpful remarks
from this
 >> thread
 >> >will soon make their way into the documentation!
 >> >
 >> >Roland Vilett wrote:
 >> >
 >> >> This isn't quite correct.  Every time you call aw_state_change(),
as
 with
 >> >> all API methods that use attributes, the values of all attributes
are
 >> sent
 >> >> to the server regardless of whether your code sets them or not. 
If
 you
 >> have
 >> >> never set an attribute, its value will be zero for integers,
false for
 >> >> booleans, and empty for strings (note this doesn't apply to attributes
 >> that
 >> >> are set automatically by the SDK  in response to events
and requests.)
 >> If
 >> >> you have set it before, it will continue to retain that value
until
 you
 >> set
 >> >> it again.
 >> >>
 >> >> This is why it is okay to only set AW_MY_GESTURE and call
 >> aw_state_change(),
 >> >> without setting all other AW_MY_ attributes as well, assuming
you have
 >> >> already called aw_state_change() before to locate your avatar
in the
 >> world
 >> >> somewhere.  The SDK will simply send again all the same
values for
 >> AW_MY_X,
 >> >> AW_MY_Y, etc. as before.  Note that this is different from
the concept
 of
 >> >> the server assigning default values for unspecified attributes.
 >> >>
 >> >> -Roland
 >> >>
 >> >> Canopus wrote in message <3676B2E8.4E3A30AC at ix.netcom.com>...
 >> >> >Also the documentation for aw_state_change might mention that
if you
 >> leave
 >> >> >out an attribute for AW_MY, the server gives you a default.
If you
 leave
 >> >> >them all out on your first aw_state_change,  your bot will
be a
 Tourist
 >> >> >located at GZ doing nothing special (all zeroes).
 >> >> >
 >> >> >Canopus wrote:
 >> >> >
 >> >> >> The need to separate state_changes by a timed lapse of 1-5
seconds
 >> >> >> should be in the documentation. It came up earlier in connection
 with
 >> >> >> bot movements (MY_X, MY_Y, MY_Z, MY_YAW); now it comes up
again in
 >> >> >> connection with bot gestures (MY_GESTURE). The example given
would
 be
 >> >> >> appropriate, since it also shows that there are (still
 undocumented)
 >> two
 >> >> >> state_changes needed in certain cases, such as gestures. New
 >> programmers
 >> >> >> are reluctant to read through over 1000 newsgroup messages
looking
 for
 >> >> >> solutions.
 >> >> >>
 >> >> >> Roland Vilett wrote:
 >> >> >>
 >> >> >> >    Part 1.1      
Type: Plain Text (text/plain)
 >> >> >> >              
Encoding: quoted-printable
 >> >> >
 >> >> >
 >> >> >
 >> >
 >> >
 >> >
 ></blockquote>
</html>

canopus

Dec 17, 1998, 7:40pm

Thanks for all the updates in the documents, Roland. I just went through the
methods, attributes, and events and downloaded over a dozen ones that sounded
new to me, including aw_state_change. (Hmmm, maybe it was there for weeks
already.)

[View Quote] > Sorry. It's just that I thought he was giving *me* a hard time about the
> docs...I'm a bit over-extended here trying to keep everything going at once
> and sometimes I can get defensive.
>
> -Roland
>
[View Quote]

roland vilett

Dec 18, 1998, 7:31pm

This is a multi-part message in MIME format.

------=_NextPart_000_0023_01BE2A8A.B0AF8960
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Thanks :)

-Roland
[View Quote] Sorry. It's just that I thought he was giving *me* a hard time =
about the=20
docs...I'm a bit over-extended here trying to keep everything =
going at once=20
and sometimes I can get defensive.=20
-Roland=20

[View Quote] remarks from this=20
aw_state_change(), as=20
with=20
attributes are=20
not. If=20
you=20
integers, false for=20
to attributes=20
and requests.)=20
value until=20
you=20
assuming you have=20
avatar in the=20
values for=20
the concept=20
of=20
attributes.=20
<3676B2E8.4E3A30AC at ix.netcom.com>...=20
that if you=20
default. If you=20
leave=20
will be a=20
Tourist=20
1-5 seconds=20
connection=20
with=20
up again in=20
given would=20
be=20
undocumented)=20
gestures. New=20
messages looking=20
for=20
>

------=_NextPart_000_0023_01BE2A8A.B0AF8960
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN">
<HTML>
<HEAD>

<META content=3Dtext/html;charset=3Diso-8859-1 =
http-equiv=3DContent-Type><!doctype html public "-//w3c//dtd html 4.0 =
transitional//en">
<META content=3D'"MSHTML 4.72.2106.6"' name=3DGENERATOR>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV>Thanks :)</DIV>
<DIV> </DIV>
<DIV>-Roland</DIV>
<BLOCKQUOTE=20
style=3D"BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: =
5px">
[View Quote] =
href=3D"mailto:36790FFB.C385FCD at poboxes.com">36790FFB.C385FCD at poboxes.com=
</A>>...</DIV>You=20
are doing a great job Roland. Keep it up.=20
[View Quote] general, the documentation is a work in progress, and it will=20
 >> improve with time >> >> -Roland =

[View Quote] >> thread >> >will soon make their way =
into the=20
documentation! >> > >> >Roland Vilett =
[View Quote] >> >> are set automatically by the SDK  in =
response=20
to events and requests.) >> If >> >> =
you have=20
set it before, it will continue to retain that value until =
 you=20
 >> set >> >> it again. >> =
>>=20
 >> >> This is why it is okay to only set =
AW_MY_GESTURE=20
and call >> aw_state_change(), >> >> =
without=20
setting all other AW_MY_ attributes as well, assuming you have=20
 >> >> already called aw_state_change() before to =
locate=20
your avatar in the >> world >> >>=20
somewhere.  The SDK will simply send again all the same =
values for=20
 >> AW_MY_X, >> >> AW_MY_Y, etc. as=20
before.  Note that this is different from the concept =
 of=20
 >> >> the server assigning default values for=20
unspecified attributes. >> >> >> =
>>=20
-Roland >> >> >> >> Canopus =
wrote in=20
message <3676B2E8.4E3A30AC at ix.netcom.com>... >> =
>>=20
>Also the documentation for aw_state_change might mention =
that if you=20
 >> leave >> >> >out an attribute =
for AW_MY,=20
the server gives you a default. If you leave >> =
>>=20
>them all out on your first aw_state_change,  your bot =
will be a=20
 Tourist >> >> >located at GZ doing =
nothing=20
special (all zeroes). >> >> > >> =
>>=20
[View Quote] ------=_NextPart_000_0023_01BE2A8A.B0AF8960--

netropolis

Dec 19, 1998, 2:46am

[View Quote] <snip>

>
>
> That was my very first SDK apps to write (modify) so I used the
> example I found on my HD. I consider to write some more - so I'll use
> the latest wrapper - if I can find it for Cbuilder3 :o)

Andras, contact me, I can send you the .lib that works with BCB 3.

-Netro

andras sarkozy

Dec 19, 1998, 7:35am

The trick is simple - use more user accounts - $20 for 3 additional bots :o))

[View Quote] [View Quote]

Awportals.com · ProLibraries Live · Twitter · LinkedIn