'''(or: Radical Minimalism in Tcl)'''

   * [Tcl]: [everything is a string]
   * [Starkit]: everything's a metakit
   * [RS]: everything's a proc
   * [vfs]: everything's a file system

[CMCc] thinks there are a lot of radical minimalists among Tclers, and that's a good thing
(although one might ask why there needs to be more than one minimalist, or more than one
kind of minimalism :)
----
[RS]: First of all, I have neither invented [proc]s, nor am I the only one who writes them. 
And I'm still discovering in how many situation we don't even have to write a proc, when an ''[interp] alias'' can do it in a lightweight way...
For my thoughts on code minimalism ("Tcl'ing is fun. Less Tcl'ing can be more fun"), see [KISS].  [CMcC] it wasn't a criticism, I just remember [RS] as coming up with the most creative abuses of proc.
----

Some more ''Everything's a ...''

   * [CMCc]: everything's an array

I'm considering a mapping from array to directory, such that get/set are read/write files, array names is glob, and so on.  This is the dual of [vfs]'s '''everything's a file system'''

This would be quite nice for persisting config arrays, and reasonably efficient by combination with ''Everything's a file system''

([AK]: The OS [[Plan 9]] treats practically every resource a process can ask for as a filesystem or a file in such.)
----
28mar03 [jcw] - ''Oooh... everything's an array - yes, please!''
(then we could go beyond [Tequila], DKF's namespaces-as-ensembles, [yet another object system], and add seasoning to taste...)

(There's a vfs driver for namespaces, both commands and vars/arrays, the command variant is now part of the [tclvfs] project, see "tclprocvfs.tcl")

In all seriousness, though "everything is an array" has IMO a lot of fascinating implications, if they can handle both traces and persistence.  For example, with traces a write trace can be "call", while a read trace access the result value - that in turn is a way to introduce async processing, "futures", Abstract State Machines, ... eh, well, by now I admit that I haven't the faintest idea how much of this can work and be useful.

For a completely different "everything is a..." perspective, see some 2002'ish ramblings on Data Views [http://www.equi4.com/179] and Generalized Data Structures [http://www.equi4.com/39].