[cells-devel] Fwd: web applications with Cells

[Ken Tilton]

---------- Forwarded message ----------
From: Ken Tilton <kentilton at gmail.com>
Date: Jul 15, 2006 1:43 PM
Subject: Re: web applications with Cells
To: "Andrew K. Wolven" <awolven at yahoo.com>


On 7/14/06, Andrew K. Wolven <awolven at yahoo.com> wrote:
> Do you use allegro?

yes.

> I think it would be useful to have persistent Cells
> objects,

yes, I have done that in the past with Allegrostore, may do it someday
with Rucksack.

>where the slots are indexable.  I believe that Cells flat updates
> contents of all dependent slots on trigger.  So you can index the slots.
>
> Anyway,
> I have made my own javascript-windows-in-a-browser (that look like M$), and
> then I found a much better javascript object-oriented window system which
> can do Mac OS X looking windows among other things.

Nice, what is that system?

>
> When the window jumps up, it keeps open a persistent connection (and I'm
> reading a webpage on how to do this), the server can send information to a
> listening javascript object.

Sounds like one could deliver a serious web app with that.

>
> So lets say you have three windows open, two are of the exaclty the same
> object and one different one.  From the different one, you drag and drop the
> contents of a slot into a slot of one of the two same objects.  The second
> same object needs to know how to update, right?  [this is just a *simple*
> example]  Javascript does an HTTP request to the server to tell it that it
> has been updated, and how.  Cells then changes the real object in persistent
> clos/cells, the dependent objects then serve *back* to the javascript object
> in the browser to update.

Sounds right.

>
> It could be cool.  I don't see how to do it with lazy evaluation
> efficiently.

Why do you want it to be lazy? And your sentence is confusing. "Lazy"
would not be inefficient, it would make things harder to program.
Puzzled.

kt