Info on continuations?

[vasudevram]

Michael wrote:
> vasudevram wrote:
>
> >
> > Hi,
> >
> > I am Googling and will do more, found some stuff, but interested to get
> > viewpoints of list members on:
> >
> > Continuations in Python.
> >
> > Saw a few URLs which had some info, some of which I understood. But
> > like I said, personal viewpoints are good to have.
>
> Python doesn't really support continuations. Generators (and co-routines in
> general) have similar properties to continuations, however they're not
> continuations. Closures are also sometimes considered cousins to
> continuations, and python's support for those is pretty good IMO. (again
> however, closures are not continuations).
>
> Since it looks like you're also looking for what continuations are, I think
> the following is the simplest way of explaining them. If you've ever
> programmed in BASIC, *in a way* if you think of methods/functions as a
> named GOSUB, then continuation (in a way) are a way of giving a name to a
> goto (it's more subtle than that though since a continuation os often
> defined to "remember" state in a similar way to a closure). Like a function
> you can pass them round as objects. Like a generator/closure they remember
> the state they were in.
>
> *Personally* , I think python NOT supporting full continuations is a GOOD
> thing, since full continuations, whilst powerful, are also a great source
> of confusion for people. (That said, I have a usecase I'd find them useful
> for - I think they'd be useful for plugin architectures - but IMO that
> doesn't outweigh the risk of code obfuscation :-)
>
> One particular usecase that people seem to like continuations for,
> specifically how they're used in seaside, is actually a subset of
> functionality that python *can* support (to a large extent).
>
> Essentially the idea is to be able to take a web application and make it
> look linear. CherryFlow allows you for example to write this:
>
> @expose
> @flow
> def example_flow():
>     yield view.first_page()
>     if request.args["choice"] == "a":
>         yield view.choice_a_page()
>     else:
>         yield view.choice_b_page()
>     yield view.last_page()
>
> (example from: http://tinyurl.com/qzpqu )
>
> This causes the user's browser to show a page where they have a choice of
> "a" or "b". Depending on what they choose, they then either are presented
> with choice_a_page or choice_b_page. Finally, no matter which option they
> chose, you are presented with the last page.
>
> Something similar can be done using Kamaelia, though at present only the
> mechanism exists there, without any sugar (http://tinyurl.com/n3bh7 -
> specifically websiteSessionExampleComponent). The reason I mention it is to
> say that it's relatively simple to do using python :-)
>
> I have to stress though, this usage of continuations in Seaside, as
> emulate-able in python is a subset of the full power of continuations,
> which isn't available in python at present. (Though it's possible hacking
> greenlets could result in something). (In fact the way seaside uses them as
> far as I can tell is more to implement co-routine like behaviour than
> anything else (!))
>
> Anyway, hope that's interesting/useful - looking at your other comments,
> you're just looking for information and usecases at the moment :-)
>
> Regards,
>
>
> Michael.

Yes, that's quite right - only looking for info and use cases at
present - and thanks for the info, it was interesting and useful :-)
Need to check all the inputs out ...

Vasudev