Yeah, I think I found that too, antimatter15 - since Bouncy worked for
my human account at the time I tried it on the same wave I was testing
my robot on.

The blip content regex solution won't help robots that don't remove
the triggering text content: e.g. text management robots, notifying,
etc.

But I have thought of another idea, have a whitelist of gadgets that
the robot specifies it wants to listen to in the capabilities.xml.

Are there any situations when a robot would want to listen to unknown
gadgets?  The white-list (rather than black-list) would then allow a
similar method to the regex content filtering on what gadgets get
their events sent to your robot.

This could either be a list of gadget URL's or just a single regex for
the gadget URL's to match against before sending the gadget events to
the robot.


On Dec 4, 8:06 pm, antimatter15 <[email protected]> wrote:
> I think I recall that Bouncy doesn't respond to posts made by robots
> (It can be added by robots though).
>
> I would like to think the solution would be to have blip content regex
> matching in compatibilities.xml
>
> On Dec 4, 9:54 am, Daniel Faust <[email protected]> wrote:
>
> > > I tried adding code to call Bouncy in to assist my robot in escape :D
> > > Sadly, it came to the wave at my robot's call, but didn't respond
> > > to the robot's cry of distress!
>
> > Sounds like a good idea. If i'm guessing correctly, bouncy needs to be
> > added to the wave and then receive configuration messages via messages
> > posted to the wave, right? In that case, posting that extra message to
> > configure bouncy to drop the bot out of the wave seems like an
> > expensive operation.
>
> > Imagine the situation if bouncy could be added to a wave, and then,
> > via a cheap and direct urlopen() call from the bot to bouncy, be
> > configured behind the scenes to remove the bot from the wave. Could
> > that be possible? The only issue I see here is that that urlopen call
> > won't hand over bouncy any context which it will probably need to
> > remove the bot from the wave. On the other hand, bouncy could queue up
> > the request removal and execute it during the next incoming message
> > coming from the problematic wave.
>
> > one issue is that if bouncy is also receiving all that incoming data,
> > then it might be the case that it also hits the quota limit, which
> > might be the reason why bouncy didn't respond to your call. So maybe
> > having google remove the incoming bandwith quota limit on bouncy would
> > be an good thing to do. Or, if that won't happen, built a network of
> > bouncies, which take care of each other so that no more than one is in
> > a given wave, and where, if one dies, others could jump in. or the
> > affected bot would first query the system to see which bouncy bot is
> > still up and running.
>
> > There would also be the need of taking protective measures against
> > direct urlopen() configuration requests from sources which might have
> > bad intentions.
>
> > Probably doing something with the capabilities.xml would be a better
> > option. Or add a dynamic blacklist.xml file.
>
> > ----------
>
> > The issue with gadgets requiring user interaction before being allowed
> > to submit data seems to be a good thing to desire. Because, as it
> > looks now, it's only necessary for someone to take a look at a wave,
> > and, as soon as a gadget gets loaded and submits data, the user gets
> > inserted as a participant. At least it looks that way, since some
> > waves have a large amount of participants whose inclusion couldn't be
> > explained by other means.
>
>

--

You received this message because you are subscribed to the Google Groups 
"Google Wave API" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-wave-api?hl=en.


Reply via email to