[Seaside] Questions about seaside and pharo

[Ramon Leon]

> With respect, having personally lived through all of it, there is no 
> "wheel" of technology.  

With respect, so have I, and we'll agree to disagree. :)

> Given the limits of processors (and the speed of light), short of  
> quantum computers (which are happening), the best scaling choice we have 
> is multi-processing.

Agree.

> Although running multiple processes and a load balancer does work, the 
> world is not moving in that direction.

I was only answering what Pharo does, not making a statement about where 
it should go.  I'd be great if the VM could access native threads, but 
it doesn't.

I currently run something like 224 Pharo images across 15 machines for 
my Seaside application. About twice what I need but that's for 
deployment reasons rather than load reasons.

> That's disappointing but probably accurate.

Meh, other frameworks to choose from, Seaside is great for component 
based applications where you want to do everything in Smalltalk and 
avoid all the complexity the web can drown you in, but that's not what 
you want.

It's especially great for building an object oriented web applications 
where any screen in the app could be one of many specialized subclasses 
because unlike any other framework I'm aware of, I can factor my HTML 
and subclass my views to override very specific render methods for all 
the various "customizations" clients demand.  I don't think I could do 
what I do in any of the modern client side frameworks.

That teapot framework reminds me a lot of Sinatra, looks perfect for a 
RESTful server framework.

At the end of the day, what's important to me is using Smalltalk, with 
whatever framework.

-- 
Ramon Leon