Garrett Goebel <[EMAIL PROTECTED]> wrote:

> I gather that Rocco Caputo would prefer a textual notation
> that looks like Perl. Whereas torvald is recommending XMI. 

Considering what Torvalds thinks for an object modelling language
seems strange.  Is the kernel OO?  I really dislike the man's coding
style as well.  Documentation far too yin.

The approach I am working on is modelling structure with a web based
front-end;

   http://sam.vilain.net/tangram/schema1.png
   http://sam.vilain.net/tangram/schema2.png
   http://sam.vilain.net/tangram/class.png 

....and storing the actual class information in a data structure (which
may or may not come from a database), then generating .pm files from
that structure.  The last screenshot is browsing in-memory symbol
tables, not the official object structure.

These .pm files are classes that represent UML Classes, Attributes and
Methods:

    http://sam.vilain.net/tangram/Class.pm
    http://sam.vilain.net/tangram/Attribute.pm
    http://sam.vilain.net/tangram/Method.pm

Once you have your class structure in this form you can call
Class->as_string to convert to a .pm file.  For the majority of simple
classes (ie, mostly data + some methods) this is perfect.

At any point, you can decide that a module should not be stored in the
database and is just a normal .pm file by setting a flag.

Flat files are an awful way for humans to examine and edit data
structures.  Don't restrict yourself to it.
--
   Sam Vilain, [EMAIL PROTECTED]     WWW: http://sam.vilain.net/
    7D74 2A09 B2D3 C30F F78E      GPG: http://sam.vilain.net/sam.asc
    278A A425 30A9 05B5 2F13

Real computer scientists love the concept of users.  Users are always
real impressed by the stuff computer scientists are talking about; it
sure sounds better than the stuff they are being forced to use now.

Reply via email to