Vision for Squeak/Smalltalk web development

What I think would be wonderful for the community to work towards
overtime and what I think would also help Seaside in garnering a
potentially larger share of distributed applications. Would be a nice,
documented, deployable, end-to-end Squeak stack for distributed
applications.
(Yes no "Web" in that expression. :)

It would be nice if on the server side we have Squeak/Smalltalk/Seaside
and on the client a nice, easily installable, well tested, Squeak
browser plugin for delivering rich apps on the client side. No need for
flash, flex, silverlight, etc. A completely open source end-to-end
solution for rich distributed applications.

Have good documentation for all the standard deployment solutions.

All Squeak/Smalltalk stack.

Seaside - Squeak-plugin

Seaside
Persistence options
   in-image
   and if desired or necessary -
     Magma
     Glorp-Gemstone, etc.
Squeak-plugin
   Degrading to Ajax
   Degrading to Non-Ajax

Nobody offers a complete stack like that in the open source world. You
could have the possibility of rich distributed apps and all in Squeak.
You can't do that in Python, Ruby, Perl, etc.  (Okay maybe Java, but
yuck. I don't like sites which use Java in the client. And thats my
opinion as a consumer, not developer.)

Most of the tools are already there. We just need to clean up the stack,
polish it, document it better.

Now, I don't really know much about the Squeak-plugin. And I don't know
how ready for primetime it is and if it is ready yet to compete with
Flash/Flex. But it would be nice for that day to come. I also don't know
what the implications would be to using the plugin for a Seaside
application.

It would be absolutely fantastic if one day we can deploy and the web
browser (if being used to access the site) would simply open up the app
in an independent squeak process and not in an in-browser-plugin. That
way the browser won't take down the Squeak app if it decides to crash.

Maybe this could be something to think about. Maybe Seaside 3.0 :)

What do we want included in a Squeak/Smalltalk/Seaside stack?

Can we develop a road map to get there?

I really believe that if done right, that we can go places the other
guys can't easily go. We can build a stack that they can't easily build.
A stack that they don't have a vision for and no means to implement.

This is an exploitable situation. And can be promoted totally in a
positive manner without mentioning or demeaning any other technology in
its promotion.

We can play just as nice and better than most with any current and
future browser, html, http, technologies. And if we can have a quality
Squeak on the client, either via plugin, Sophie, ???. And an image
appropriately designed, shrunk, ??? for client-side deployment. Then our
deployment options are much greater than most others.

I just wanted to toss this ball out there and see how it goes.

Thanks.

Jimmie
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

On Fri, Jul 13, 2007 at 11:24:02AM -0500, Jimmie Houchin wrote:
> What I think would be wonderful for the community to work towards
> overtime and what I think would also help Seaside in garnering a
> potentially larger share of distributed applications. Would be a nice,
> documented, deployable, end-to-end Squeak stack for distributed
> applications.
> (Yes no "Web" in that expression. :)

I really don't think Seaside is the right tool for distributed
applications; it is quite fundamentally a server-oriented
architecture. You may find this interesting, however:

http://www.vpri.org/pdf/tinLizzie_TR-2007-001.pdf

--
Matthew Fulmer -- http://mtfulmer.wordpress.com/
Help improve Squeak Documentation: http://wiki.squeak.org/squeak/808
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Matthew Fulmer wrote:
Well, maybe I didn't express that accurately enough. I don't mean
distributed as in many-to-many, but rather simply I am doing Desktopish
type stuff from a server based app. I think a full Squeak stack can do
better than Google Office or the current supply of server-based
Ajaxified apps.

I think we can do those to if we really wanted to, but.

What if offered a Squeak alternative to Google Gears?
http://gears.google.com/

Google Gears is providing an opportunity to do a desktop type app with
client-side Javascript and SQLite. We can do it better and easier, I think.

I may not have expressed myself well in avoiding using "Web" in
describing the ideas. Basically I didn't wish to include "Web" because I
think we can do better than what the "Web" means to most people. And if
you go from Squeak server (Seaside) to Squeak-plugin/Squeak-client then,
is it really "Web", or simply something richer being served over http,
or somesuch protocol?

Most people are seemingly attempting to bridge the problem with
server-side-technology-of-choice (Rails or such) and Ajax/Flash. In
which case you have to master multiple technologies, languages, and such.

Question is, is this anything the community wants? How do we motivate
ourselves to get there? Can we as a community achieve such? Is the
community to small? Lack time? Do we need more people? If it is
something we want, how do we overcome the obstacles?

It seems to me that if we want a end-to-end Squeak/Smalltalk stack, that
as we solidify, polish and document each piece of the stack, allowing
people to use that portion of the stack easier and better, that we will
attract others. And hopefully we will gain sufficient people to complete
the stack with quality code, deployable solutions and good documentation.

Well, I don't if I've cleared anything up or not. Or if I muddied the
waters further. We'll see. :)

Any way, I haven't read the article about TinLizzie yet. But I have it
downloaded, printed and in my to read stack. Soon!

Thanks.

Jimmie
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

I like the idea of the homogeneous stack, and I feel like software is
moving this way.  I will have a blog post about it soon enough, but the
basic idea is that things in CS always go back and forth.  I don't know
how many times the industry has switched from homogeneous to
heterogeneous environments, but we are very hetro right now and the
pendulum is swinging back the other way in my opinion.

Jimmie Houchin wrote:
_______________________________________________
Seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside