[plt-scheme] Continuations

From: nish (nish2575 at yahoo.com)
Date: Tue Nov 21 22:08:23 EST 2006

My naiveness might even bring a fresh perspective to this discussion. I
entered application development by going from a little c++ and soem
comp sci and reading sicp -> coldfusion at work -> a lot more comp sci,
scheme reading  -> c# & delphi at work-> some more programming language
reading -> back to php now at work. 

So i've built and seen a number of the traditional apps. using the
branch on a keyword within a server page. and various workflows strung
across server pages. coming from the simple client perspective, i've
resisted ajax a bit, seeing it as a potential for unnecesarily
complicating some clients and introducing new possible threading bugs,e

but anyways, i just recently sepnt a little time just developing scheme
servlets (and a scheme server on bytmark!!) with plt web. i read the
continuation papers around when they came up and they've been in the
back of my mind. i've always kept an eye out for when any of my apps
started to feel like they were turning into wizard type apps. 

> > > The world hasn't moved anywhere[1]!
> >
> > I disagree.  Two points:
> >
> >  - AJAX apps benefit from a very different model to CGI
> > style apps
> ... or to put it more generally - interactive applications. I fail to
> see how putting more logic and interactivity at the client side has
> shown weakness of existing models in stronger ways.

well. i think many web apps have limited how many features they've
developped because of the limts of the client. as we push back towards
teh fuller client, people will have to write more and more
servlets/server side code to handle all of the new abilities exposed
through the client. from the large corporate web apps i interact with
lately in 2006, they've haven't really hit that limit where they are
writing so much complicated server side code that they are desperte for
more abstraction. they seem to be getting by with the old paradigm. 

> Sincerely
> - Akhilesh

> > I think they will solve a number of the problems (but not
> > bookmarkability, though you can solve this in other ways),
> The problem solves itself if we assume continuations are perpetual.
> That also means that may be persisted continuations need to be made
> "continu-able" across runtime invocations... and that probably leads
> us to a servlet model where servlets are instantiated in a fixed
> known
> initial environment.

Excuse my ignorance here, but can you detail out a little more what
you're envisioning please. are you thinking of each web app is a
(multithreaded) program that gets started once, and frozen on and off
by various browser clients, and garbage collected with timeouts..

Overall, as i've had continations on the back of my mind while folowing
whatever standard operating procedures wehrever i'm working, i'm always
looking for excuses to use them. and i have coded a few hand-rolled
versions of continuations along the way. but as much as i think they
are cool and interesting, a lot of times i come bakc to this thought: 

in the beginning was teh commmand line ... modal interface -> WIMPs and
implicit event handler loop with the user more in control with a few
wizard/modal apps on teh side -> web apps that usually have some sort
of menu and conform mostly to the idea that the user is still in
contorl of creating the applicaiton sequences. 

and most features or new servlets i add to an app, still tend to be a
single request that is handled in the top request loop. i think of a
contination as needed when a request from the top request loop needs a
few nested steps before completeting. i haven't come across enough
times where i need that that nested / wizard flow in an app yet. 


Sponsored Link

Compare mortgage rates for today. 
Get up to 5 free quotes. www2.nextag.com

Posted on the users mailing list.