UML cooperation

[vio]

* Alejandro Imass <address@hidden> [001023 19:46]:
> Hey.
> 
> I have revised my stuff, and will be working on an
> official model soon. I think we should work on it
> toghether.

Hello Alejandro,

Yes, I'd like that.



Regarding cooperation, perhaps this would be a good time also for me to say 
where I stand in the GNUe effort. Basically, I am somewhere on the border (as 
opposed to the core). I'm with the core GNUe on principle, and partially on 
implementation. Yes, I've started a "somewhat" parallel project named myERP, 
but mainly because I don't want to waste energy convincing the core GNUe team 
to adapt my own views. You have your own views on how to do things, and I 
respect them. I just am more comfortable in implementing my own views (and not 
having to convince anybody that we should do this and that instead, like 
"asking permission" to do something), and being a project leader is the path of 
least resistance (hence myERP's inception). But besides that, I'm with you 
guys, and hope we cooperate as much as possible. GNUe and myERP are NOT 
competitors (that's a stupid thought!!). Plus you have code out, myERP has 
nothing yet. Just wanted to point that out.

What I've been doing regarding the open source ERP effort, since I've "joined" 
it: lots of reading and studying. Well, I've learned Python, then I've 
"translated" an SQL script into Python (written a Py script to convert SQL 
table definitions into Python class definitions, to give me an enterprise's 
skeleton - since the original SQL did that ... but in SQL; now as things are 
evolving, this code probably won't get used further, though - read on). Then 
I've read "ERP: tools, techniques and apps for integrating the supply chain" (I 
believe this should be mandatory reading for this project - this book pasically 
points out pusiness processes (design or "re-engineering") as the most 
important element in ERP, which should get the most attention, time, effort, 
Nobel Prize). And last week started reading in parallel "UML with RationalRose" 
and the excellent "Practical C Programming" (to refresh my memory in 
preparation for "C++: the core language" - from the time O'Reilly did really 
excellent books). "XML specification guide" sits also on my desk, for dessert. 
Note to self: must find something good on CORBA. Oh yes, I've also learned how 
to start a project on sourceforge.

Well, you probably see where I'm getting at (my "modus operandi" or something) 
from these readings: I'm going towards modeling the principles of "ERP: tools, 
techniques and apps" (reinforced by the 2k edition of "Operations Management", 
a truly excellent book for this project, filled with useful formulas ... 900 
pages though) in RationalRose (RR), then letting RR generate C++ code from that 
model (could spit out Java also, but why ?), and apply only the finishing 
touches manually (hence refreshing my C/C++ region of the brain). I am 
therefore moving away from Python for 3 reasons: because it's not that much 
additional effort to write C++ code instead, the code will be 100 times faster, 
and mostly because RR will do most of the work.

If myERP (my personal version of GNUe, if you prefer) will become a monumental 
flop, at least I've increased my valuation (hopefully) in the marketplace. All 
this being said, I definitely believe, Alejandro, that we should cooperate at 
the modeling level, even though afterwards GNUe will go the Python path, while 
myERP will go C++. So what tool are you modeling on ? How can we exchange 
models (compatible file formats would be nice) ? Have I mentioned that I just 
started learning UML (can only do "use case" diagrams at this precise moment in 
the time/space continuum). More diagrams coming soon.

So see you,
Vio