Re: [ns] Relation among Delaylink,Simplelink,LinkDelayClass

[Andrea Josi <mr_josi@yahoo.it>]

Sorry again,

i would like to know whether the next statement is
correct or not:

the link is composed of a set of connectors; for
example, the connector that implements the head of the
link is introduced by:

set head_ [new Connector]

Actually, the Connector class is only inserted in OTCL
class hierarchy at startup by the SplitObject
"register" method (due to the definition of a static
object of ConnectorClass): 
anyway the Connector class has no OTCL methods, apart
from those inherited by the SplitObject class; 
this is also the case of the "init" constructor:
i assume that "new Connector" triggers the invocation
of SplitObject "init", as no "init" method is defined
for the Connector class.

Thanks again :)

--- Haobo Yu <haoboy@packetdesign.com> ha scritto: > >
Anyway, this approach doesn't follow the
> "standard" NS
> > philosophy based on OTCl/C++ linkage;what is the
> use
> > of LinkDelayClass in this context then ?
> > (by the way,the documentation refers to a
> Delay/Link
> > class, which probably was part of an old NS
> version)
> 
> Every link in ns is a _composite_ object that
> consists of many "smaller"
> objects, e.g., ttl, queue. DelayLink is one
> components there for
> propagation delay.
> 
> c++/otcl linkage is for interfacing the otcl code
> with c++ code and
> conceptually has nothing to do with object
> composition.
> 
> You may want to read slides from the last two ns
> workshop, which, I think,
> explain in detail the ns internal.
> 
> - Haobo
> 


______________________________________________________________________
Do You Yahoo!?
Il tuo indirizzo gratis e per sempre @yahoo.it su http://mail.yahoo.it