Isjtar

i'm experimenting with association networks, nn, the sort of thing where the basic unit is simple, but the network the strength.

scripting is ideal for making this kind of thing, but i get nothing but crashes.

generating the nodes is no problem, it's the connections where it goes wrong.

it's an overflow thing, but i'm not quite sure why. i try to cut down on uzi's, but you need at least one for getting things done in acceptable speed.

would JS be more stable? i wouldn't think it makes a difference and i feel safer in patchland.

so, any tips or solutions you gys could give me?

here's the patch, it will normally crash, so watch out!

------------------------------------------------------

i'm experimenting with association networks, nn, the sort of thing where the basic unit is simple, but the network the strength.
scripting is ideal for making this kind of thing, but i get nothing but crashes.

here's the patch, it will normally crash, so watch out!
------------------------------------------------------
save this dummy as "n"

scripting-woes

Hi isjtar , i think triggering the uzi *too early* causes the main crash, i tried storing the node number before firing the uzi. it worked fine once , but when i tried to connect a new node Max kept crashing so it seems that deleting old connections before creating new ones solves the problem . although i dont know what you are exactly aiming for .

Hi isjtar , i think triggering the uzi *too early* causes the main crash, i tried storing the node number before firing the uzi. it worked fine once , but when i tried to connect a new node  Max kept crashing  so it seems that deleting old connections before creating new ones solves the problem . although i dont know what you are exactly aiming for .

Quote: karl-otto von wrote on Wed, 28 June 2006 19:25

----------------------------------------------------

> Hi isjtar , i think triggering the uzi *too early* causes the main crash, i tried storing the node number before firing the uzi. it worked fine once , but when i tried to connect a new node Max kept crashing so it seems that deleting old connections before creating new ones solves the problem . although i dont know what you are exactly aiming for .

sorry, should've explained better. i noticed that deleting some connections solves the problem, but that's the point, i'm trying to connect everything to everything all in one step.

it's not really a problem, i work around it with s/r, but it's just strange that scripting is so unreliable compared to the rest of max.

Quote: karl-otto von wrote on Wed, 28 June 2006 19:25
----------------------------------------------------
> Hi isjtar , i think triggering the uzi *too early* causes the main crash, i tried storing the node number before firing the uzi. it worked fine once , but when i tried to connect a new node  Max kept crashing  so it seems that deleting old connections before creating new ones solves the problem . although i dont know what you are exactly aiming for .
----------------------------------------------------
sorry, should've explained better.  i noticed that deleting some connections solves the problem, but that's the point, i'm trying to connect everything to everything all in one step.
it's not really a problem, i work around it with s/r, but it's just strange that scripting is so unreliable compared to the rest of max.


> but that's the point, i'm trying to connect everything

> to everything all in one step. it's not really a problem, i work

> around it with s/r, but it's just strange that scripting is so

> unreliable compared to the rest of max.

For me scripting is not unreliable, if you read my previous post its 

even expected behaviour if you create feedbacks which would create a 

stack overflow in the second you feed any information into it...

(Which you would definitely do by connecting "everything to everything")

junior wrote:
> but that's the point, i'm trying to connect everything
> to everything all in one step. it's not really a problem, i work
> around it with s/r, but it's just strange that scripting is so
> unreliable compared to the rest of max.

For me scripting is not unreliable, if you read my previous post its 
even expected behaviour if you create feedbacks which would create a 
stack overflow in the second you feed any information into it...
(Which you would definitely do by connecting "everything to everything")

-- 
Stefan Tiedje------------x-------
--_____-----------|--------------
--(_|_ ----|-----|-----()-------
-- _|_)----|-----()--------------
----------()--------www.ccmix.com


Quote: Stefan Tiedje wrote on Wed, 05 July 2006 10:36

> For me scripting is not unreliable, if you read my previous post its 

> even expected behaviour if you create feedbacks which would create a 

> stack overflow in the second you feed any information into it...

> (Which you would definitely do by connecting "everything to everything")

connecting "everything to everything" isn't necessarily a cause of stack overflow, as the inputs and outputs of the subpatches aren't even connected yet.

maybe there's a mistake in my patch which causes double connection messages, i can imagine that this causes problems.

Quote: Stefan Tiedje wrote on Wed, 05 July 2006 10:36
----------------------------------------------------
> For me scripting is not unreliable, if you read my previous post its 
> even expected behaviour if you create feedbacks which would create a 
> stack overflow in the second you feed any information into it...
> (Which you would definitely do by connecting "everything to everything")
>
> Stefan
----------------------------------------------------
hi stefan,

connecting "everything to everything" isn't necessarily a cause of stack overflow, as the inputs and outputs of the subpatches aren't even connected yet.
maybe there's a mistake in my patch which causes double connection messages, i can imagine that this causes problems.

scripting woes