Re: [Fresco-devel] Questions

On Thu, May 27, 2004 at 10:31:46AM +1000, michael wrote:
> Hello all,
> 
> This is my first post so I hope I don't annoy to many people.
> 
> I have a couple of question about the architecture of fresco. I have 
> download the milestone (unfortunately lacking for a chance to browse it 
> just yet). I read the critique by Graydon in the last message from the 
> list and have to disagree with some of the points made - well not 
> essentially disagree yes these ideas are very hard to successfully take 
> from idea to implementation. Any system in use today is aspiring to meet 
> these features and any future systems will (most probably) not even be 
> considered if it doesn't.

Well, this is just my opinion; I'm not entirely sure how much the
other developers agree.  But I think that Graydon's critique is
actually very sound; traditionally the people designing Fresco (and
this includes me) have been much more worried about making a system
that looks good on paper, has all sorts of nice feel-good buzzwords,
etc., and not really worried about how it will all come together in
the nuts and bolts.  Like I said in my commentary to Graydon, there
are a few really important, core ideas that Fresco needs to keep if
its to be worth the trouble -- but there's also a lot of junk that
just gets in the way of getting work done.  I've been arguing for
a bit now that a fairly radical redesign is needed -- see my posts
with "ReFresco" in the title.

> Now that I've thrown in my 2 cents worth, my questions are,
> 1. Rather than taking the traditional GUI approach of providing simple 
> GUI Toolkits has a Component oriented approach been considered? i.e. 
> providing a component architecture (and wiring standards) in which 
> applications can be developed?

Interesting question.  Can you actually explain what a
Component-oriented approach is?  I've never been able to figure it
out.  Any time I try reading something about components, I end up
bogged down is this great morass of mechanisms, with very little
explanation of principles or purpose.  And sometimes I wonder if it's
another bit of the design snake-oil so common in programming.  But
maybe not -- _do_ you know of any good explanation of the benefits,
and how it achieves them?  Or even a simple description of what it
actually does?

> 2. a) What version of CORBA are we talking about ?

None, if I have my way ;-).  Currently, though, it's, umm, I forget
exactly.  Whatever OmniORB supports, basically.

>    b) If Fresco is already going to be Component oriented why adopt the 
> CORBA Component Model (CCM)?

Not sure what you're asking.  See above, though.

> 3. Have you consider working Mono (www.go-mono.com) to provide support 
> for language neutrality i.e. supporting the .NET development framework?  
> (this would open Fresco up to a large number of applications in due time 
> i.e. the next generation of Gnome applications could potentially all run 
> under Fresco without modification).

"language neutrality" is a funny phrase; it doesn't mean anything
alone.  Certainly I'd be happy to see someone wants to write a client
library for Mono, but that's because I'd be happy to see someone write
any client library :-).  And the server itself is only one program;
"language neutrality" for it is pretty pointless.

Besides which, Mono doesn't really provide language neutrality; it
just provides support for a few new .NET languages that not many
people use yet.  It would be just as easy to support the _current_
generation of Gnome applications without modification by supporting C.

Of course, it won't actually be that easy; just because two programs
are in the same language doesn't mean they'll seamlessly interoperate.

> I think thats enough questions. The primary issue I can see on the 
> surface is that Interviews/Berlin/Fresco which ever name you choose has 
> never been market oriented (yes I know this is an open source project), 
> hence why it has taken so long to reach it current level of development. 

Actually, that's part of it.  But I think a bigger reason is that the
code is so very hard to work with, for the reasons Graydon outlined in
his message.  And if you read ReFresco 1, you'll see my argument that
CORBA is one of the major sources of this.

> By market I mean who will the users be and how can we entice them to 
> transition to using this well needed replacment for the X-windowing 
> system. If we create this "market" then more development support will 
> arrive (and the task won't seem so big) because people will see a use 
> for Fresco (in my own vision is a newer unix/linux based system that is 
> very capalbe of going toe-to-toe with the dominant GUI operating 
> system(s) in the market place today).
> 
> I think these statements are a bit brash for a new comer but I make no 
> apologies, correct me were I am wrong. Fresco is a the type of project 
> that 
> has alot of pontential to make a real impact on the software industry 
> and that excites me.

No problem.  We need iconclasts around to shake us up :-).  (We all
have to be iconclasts ourselves to be working on Fresco, after all...)

-- Nathaniel

-- 
"Of course, the entire effort is to put oneself
 Outside the ordinary range
 Of what are called statistics."
  -- Stephan Spender