Dear All,

I'm so sorry!

The mail below is a *prematurely* sended *unfinished* one, which I just
wanted to store as a draft, but I'm - obviously - too tired now...
And now there is no chance to call it back ;-)

Please ignore it!


Greetings
Stephan

Stephan Rudlof wrote:
> Dear All!
> 
> I want to provide some feedback.
> 
> To me: I've heavily worked with Squeak a few years ago, made some pause
> for different reasons and have now come back somewhat.
> 
> Note: the term 'package' means here an arbitrary amount of code
> - building a semantic unit,
> - separated from other code,
> - in an arbitrary format,
> - accessible by some URL.
> 
> 
> Tools for Administrating and Providing Squeak Packages
> ------------------------------------------------------
> 
> To get an overview you have to look into many things.
> There are tools for administrating and especially loading packages:
> PackageLoader, MonticelloBrowser.
> There are tools for generating packages: Monticello, ChangeSetBrowser,
> other Browsers.
> There are repositories:
> - SqueakSource (http://kilana.unibe.ch:8888) for Monticello repositories,
> - SqueakMap (http://map1.squeakfoundation.org/sm/) for packages in
> arbitrary formats.
> There is a web site for documentation of all this stuff:
> - Squeak Swiki (http://minnow.cc.gatech.edu/squeak/).
> 
> 
> SqueakMap
> ---------
> SM is great!
> 
> 
> Monticello and SqueakMap
> ------------------------
> 
> It's difficult to see the best way for publishing a Monticello package:
> - Directly at the SqueakMap server?
> - Or at SqueakSource?
> - Or at your own domain?
> I assume that for collaborate work SqueakSource is the way to go.
> 
> The SM packages are shown in PackageLoader, but Monticello packages may
> have some dependencies just shown in the MonticelloBrowser.
> 
> 
> Dependency System
> -----------------
> A SM dependency mechanism is lacking.
> There is some Monticello depency mechanism.
> If SM gets one soon the ...
> 
> 
> Namespaces
> ----------
> I remember the efforts to install Environments for separating
> namespaces, which habe been abandoned after a first trial to work with them.
> But without namespaces adding methods to classes more than one package
> depends on is potentially dangerous since it can break things.

-- 
Stephan Rudlof (sr at evolgo.de)
   "Genius doesn't work on an assembly line basis.
    You can't simply say, 'Today I will be brilliant.'"
    -- Kirk, "The Ultimate Computer", stardate 4731.3