DataFlowPoisson?

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

DataFlowPoisson?

Bob Brown

I think that the name of this class is actually a typo?

 

Instead of:

 

groovyx.gpars.dataflow.operator.DataFlowPoisson

 

Should this really be:

 

groovyx.gpars.dataflow.operator.DataFlowPoison

 

The purpose of this class is to “poison” a channel such that it can no longer operate, is it not?

 

Cheers,

 

BOB

 

---

Bob Brown

Director, Transentia Pty. Ltd.

Training, consulting, development at the leading edge of technology.

 

    73 Frasers Rd.                Mobile: 0416-078-813

    Ashgrove                      Email: [hidden email]

    Brisbane                      Web: www.transentia.com.au

    Queensland 4060

    Australia

 

NOTE: This email may NOT be taken to constitute an agreement to form a

"prior relationship" or to give "prior affirmative consent" for the

purposes of opt-in mass marketing.

 

---

 

"Electricity made us angels 

   and money made us fools, 

 but fear of future shackled us

   to gods who ride the backs of mules."

 -- Bill Nelson

 

Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Russel Winder
On Mon, 2011-05-16 at 19:04 +1000, Bob Brown wrote:
[ . . . ]
> groovyx.gpars.dataflow.operator.DataFlowPoison

> The purpose of this class is to “poison” a channel such that it can no
> longer operate, is it not?

or groovyx.gpars.dataflow.operator.DataflowPoison even -- the names of
Dataflow things has changed recently to fit with the spelling the rest
of the world uses, i.e. it is dataflow not data flow.

--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc (205 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Vaclav
Administrator
Thank you for the correction. Now, when we are about to change the name it I'm actually thinking of naming it "PoisonPill". Any opinions?

Vaclav



On Mon, May 16, 2011 at 11:24 AM, Russel Winder <[hidden email]> wrote:
On Mon, 2011-05-16 at 19:04 +1000, Bob Brown wrote:
[ . . . ]
> groovyx.gpars.dataflow.operator.DataFlowPoison

> The purpose of this class is to “poison” a channel such that it can no
> longer operate, is it not?

or groovyx.gpars.dataflow.operator.DataflowPoison even -- the names of
Dataflow things has changed recently to fit with the spelling the rest
of the world uses, i.e. it is dataflow not data flow.

--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: [hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder



--
E-mail: [hidden email]
Blog: http://www.jroller.com/vaclav
Linkedin page: http://www.linkedin.com/in/vaclavpech
Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Russel Winder
On Mon, 2011-05-16 at 12:34 +0200, Vaclav Pech wrote:
> Thank you for the correction. Now, when we are about to change the
> name it I'm actually thinking of naming it "PoisonPill". Any opinions?
>
> Vaclav

Poisoning is the act of changing the metadata/control-data associated
with a channel.  "poison" (as in "to poison", a verb) would therefore be
a sensible name of a method that undertakes the poisoning.

Poison (as a noun) is the substance that is used during poisoning, but
whilst this exists in the real world, it is not something normally
associated with channels.  Having an object of type DataflowPoison (or
whatever) implies that the act of poisoning the channel is achieved by
sending in a special data item (an instance of DataflowPoison) rather
than making a state-change method call on the channel itself.

I guess it depends on what happens at the reader end:  does the reader
see a closed channel, or do they get given the "poison" object and no
longer able to read from the channel after that?

I'm afraid we need some use cases to be decisive . . .

>
> On Mon, May 16, 2011 at 11:24 AM, Russel Winder <[hidden email]>
> wrote:
>         On Mon, 2011-05-16 at 19:04 +1000, Bob Brown wrote:
>         [ . . . ]
>         > groovyx.gpars.dataflow.operator.DataFlowPoison
>        
>         > The purpose of this class is to “poison” a channel such that
>         it can no
>         > longer operate, is it not?
>        
>        
>         or groovyx.gpars.dataflow.operator.DataflowPoison even -- the
>         names of
>         Dataflow things has changed recently to fit with the spelling
>         the rest
>         of the world uses, i.e. it is dataflow not data flow.
>        
>         --
>         Russel.
>         =============================================================================
>         Dr Russel Winder      t: +44 20 7585 2200   voip:
>         sip:[hidden email]
>         41 Buckmaster Road    m: +44 7770 465 077   xmpp:
>         [hidden email]
>         London SW11 1EN, UK   w: www.russel.org.uk  skype:
>         russel_winder
>
>
>
> --
> E-mail: [hidden email]
> Blog: http://www.jroller.com/vaclav
> Linkedin page: http://www.linkedin.com/in/vaclavpech
--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc (205 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Vaclav
Administrator
The act of poisoning has an effect on operators, rather than on channels, since operators are the elements that are active. When an operator receives the "poison pill", it stops its event loop and so it will not read any further messages from its input channels. Also, the "poison pill" then is sent on to all the operator's output channels for the operators down the stream to taste as well.


On Mon, May 16, 2011 at 3:13 PM, Russel Winder <[hidden email]> wrote:
On Mon, 2011-05-16 at 12:34 +0200, Vaclav Pech wrote:
> Thank you for the correction. Now, when we are about to change the
> name it I'm actually thinking of naming it "PoisonPill". Any opinions?
>
> Vaclav

Poisoning is the act of changing the metadata/control-data associated
with a channel.  "poison" (as in "to poison", a verb) would therefore be
a sensible name of a method that undertakes the poisoning.

Poison (as a noun) is the substance that is used during poisoning, but
whilst this exists in the real world, it is not something normally
associated with channels.  Having an object of type DataflowPoison (or
whatever) implies that the act of poisoning the channel is achieved by
sending in a special data item (an instance of DataflowPoison) rather
than making a state-change method call on the channel itself.

I guess it depends on what happens at the reader end:  does the reader
see a closed channel, or do they get given the "poison" object and no
longer able to read from the channel after that?

I'm afraid we need some use cases to be decisive . . .

>
> On Mon, May 16, 2011 at 11:24 AM, Russel Winder <[hidden email]>
> wrote:
>         On Mon, 2011-05-16 at 19:04 +1000, Bob Brown wrote:
>         [ . . . ]
>         > groovyx.gpars.dataflow.operator.DataFlowPoison
>
>         > The purpose of this class is to “poison” a channel such that
>         it can no
>         > longer operate, is it not?
>
>
>         or groovyx.gpars.dataflow.operator.DataflowPoison even -- the
>         names of
>         Dataflow things has changed recently to fit with the spelling
>         the rest
>         of the world uses, i.e. it is dataflow not data flow.
>
>         --
>         Russel.
>         =============================================================================
>         Dr Russel Winder      t: +44 20 7585 2200   voip:
>         [hidden email]
>         41 Buckmaster Road    m: +44 7770 465 077   xmpp:
>         [hidden email]
>         London SW11 1EN, UK   w: www.russel.org.uk  skype:
>         russel_winder
>
>
>
> --
> E-mail: [hidden email]
> Blog: http://www.jroller.com/vaclav
> Linkedin page: http://www.linkedin.com/in/vaclavpech

--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: [hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder



--
E-mail: [hidden email]
Blog: http://www.jroller.com/vaclav
Linkedin page: http://www.linkedin.com/in/vaclavpech
Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Matt Passell
In reply to this post by Bob Brown
On Mon, May 16, 2011 at 5:04 AM, Bob Brown <[hidden email]> wrote:

I think that the name of this class is actually a typo?

 

Instead of:

 

groovyx.gpars.dataflow.operator.DataFlowPoisson

 

Should this really be:

 

groovyx.gpars.dataflow.operator.DataFlowPoison


I actually quite like the idea of a data flow fish - makes me think of salmon swimming upstream via several different tributaries. :)  Ask Guillaume if you have no idea what I'm talking about.

--Matt
Reply | Threaded
Open this post in threaded view
|

Re: DataFlowPoisson?

Russel Winder
On Mon, 2011-05-16 at 09:41 -0400, Matt Passell wrote:
[ . . . ]
>        
>         groovyx.gpars.dataflow.operator.DataFlowPoisson
[ . . . ]
>
> I actually quite like the idea of a data flow fish - makes me think of
> salmon swimming upstream via several different tributaries. :)  Ask
> Guillaume if you have no idea what I'm talking about.

:-)

I actually got sent along the lines of statistics and queueing theory:
Poisson distributions.  Very pertinent to the flow of things in other
things!

--
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[hidden email]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [hidden email]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

signature.asc (205 bytes) Download Attachment