From: ptkwt@...1.aracnet.com (Phil Tomson)
Date: 2003-01-14T16:53:45+09:00
Subject: Re: UML tool for Linux?

In article <200301141542.16206.tim@bates.id.au>,
Tim Bates  <tim@bates.id.au> wrote:
>On Tue, 14 Jan 2003 03:23 pm, Phil Tomson wrote:
>> What good is UML?  I mean, it seems to me that especially with languages
>> like Ruby you could just code your OO system and then refactor as needed.
>
>I haven't actually used UML before, which is why I'm only just looking for 
>software for it. The reason is, I come from a strong engineering background 
>(father was an engineer, I'm now studying engineering myself) and as such I 
>find it helpful (sometimes even necessary) to get my ideas on paper, draw 
>diagrams etc. I'm currently trying to get my head around the relationships 
>between a significant number of classes and modules. Being a fairly visual 
>person, I want diagrams, but rather than try and do it with conventional 
>pen-and-paper methods, and end up as I usually do with lots of lines and 
>scribble and crossings-out, I thought I'd give UML a shot. It's designed for 
>this sort of thing and has the advantage that I can check it into CVS along 
>with the code... It may work, it may not.
>

I was originally more of a hardware engineer than a software engineer, now 
I'm into software for creating hardware, but I digress... Again, years ago 
when I started out on the hardware side of things schematics (pictures 
that represented a circuit and it's connections) were used heavily in 
hardware design.  Nowadays schematics, for digital design anyway, seem 
rather quaint.  The problem with schematics was that it took so long to 
draw them and for complex systems you needed lots and lots of schematics.  
In the late 80's and early 90's hardware descriptions languages burst on 
the scene... as it turns out a picture wasn't worth a thousand words.  
HDL's have proven to be a much better way of representing hardware 
designs.  Now HDL's are moving to a higher level of expressiveness (like 
programming languages in general) and are being based on C++ (or Ruby - 
see RHDL :).

So it would seem that both on the software side (with flowcharts) and the 
hardware side (with schematics) there has been a move away from the 
pictorial representation of designs because of the inherent limitations of 
those representations for representing complexity and their 
'unwieldlyness'.  Given these trends I have to conclude that UML will 
disappear eventually as well... I guess that's why I've never bothered to 
learn UML.

Phil 

-- 
"Or perhaps the truth is less interesting than the facts?" 
Amy Weiss (accusing theregister.co.uk of engaging in 'tabloid journalism')
Senior VP, Communications
Recording Industry Association of America