Re: [GNU-COBOL] GNU-COBOL list reborn

["Justin C. Ferguson" <jferg@lusArs.net>]

When last we left our heros, they were speaking of:
> Being, by nature, a back-end type as well, I'd like to see what you have
> in mind.  Without intending to give offense, I'm not much interested in a
> term project or proof of principle exercise, but I find the idea of
> creating a production quality compiler with the expectation of having
> real users pretty damned interesting.  In my most recent efforts, my team

	I think, without a doubt, that that is the goal of all involved
	at this point.  I think that since the project started as a term
	project for Laura and Chad, the misconception that that's what it
	still is may be there, but that's just that, a misconception...in
	fact, Chad and myself are no longer even in school, and are still
	with the project because we want to be.  
		That said, however, I think the work Laura and Chad did 
	in proof-of-concept is a valuable addition to the project, and should
	not be discounted just because it's not "production-class".  It 
	will definitely provide a valuable foundation for getting the 
	production-quality version finished.  
		I also hope that it goes without saying that, _above_all_, 
	we should ensure that our compiler does the Right Thing[tm] 
	according to the standard, and implements all required portions of
	the standard.  More importantly, though, the "optional" portions of
	the standard should be looked at whenever implementing the required 
	portions, because the impression I got from Laura was that many of 
	the more arcane extensions to COBOL were where the traditional 
	parser generators were failing...so the last thing we want to do is
	create a huge base of code, and then realize that we can't implement
	something major with the tool we choose.  

> I'm more in favor of opening the bazaar, Chad.  When I see the first
> technical food fight on this list, I'll know we're on our way.

		While I'm sure you're going to incur the wrath of Chad ;-> for
	mentioning CatB, I, in a sense, agree with you.  I think attempting
	a too "Cathedral-esque" arrangment is what has been holding us back;
	however, I also think a complete Bazaar is going to prove somewhat 
	chaotic.  My thoughts are as follows... (and yes, most of this was 
	thought of as I was trying to fall asleep last night, so it may just 
	be insane rambling).  
		It seems to me as if the creation of the compiler is going to
	have four basic sections: a.) grammar/parser  b.)  runtime libraries
	c.) documentation  d.) testing.   My thought is that one person 
	should be "coordinator" for each of these separate sections (bazaars), 
	much as Linus is for the Linux kernel.  This will help prevent some 
	of the chaos, and the project will thus become a bazaar of bazaars.
	(Man, that word looks weird when you type it repeatedly.)  I also
	think that there should be one person, a "project secretary", as 
	Chad put it last night when I was bouncing some of these ideas off of 
	him, who is responsible for a.) making sure the 4 groups don't go off 
	in completely divergent directions, b.) keeping communication going
	between he groups, c.) directing traffic, and d.) delegating in the 
	case of any grey areas between the groups.

		That said, this may just be my anal-retentive personality 
	coming out, and this level of organization may not be necessary.  
	Feel free to rip it to shreds if you like.  I'm completely open to
	other ideas.  

	Back to writing Java.

								JF

-- 
Justin Ferguson - Tech. Analyst/Geek of All Trades   jferg@lusars.net
http://www.lusars.net/~jferg/                            jferg@geeks.com
"If you've never stared off into the distance                jferg@acm.org
then your life is a shame..." - Counting Crows

--
This message was sent through the gnu-cobol mailing list.  To remove yourself
from this mailing list, send a message to majordomo@lusars.net with the
words "unsubscribe gnu-cobol" in the message body.  For more information on
the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.