Re[2]: Version issues

Jim Whitehead (ejw@ics.uci.edu)
Thu, 1 Apr 1999 17:12:17 -0800 

From: Jim Whitehead <ejw@ics.uci.edu>
To: Versioning <ietf-dav-versioning@w3.org>
Date: Thu, 1 Apr 1999 17:12:17 -0800
Message-ID: <005c01be7ca5$d997b9c0$d115c380@ics.uci.edu>
Subject: Re[2]: Version issues

Just caught this one -- it's probably out of sequence with the rest :-(

-----Original Message-----
From: Bradley.Sergeant@merant.com [mailto:Bradley.Sergeant@merant.com]
Sent: Sunday, March 14, 1999 10:03 AM
To: 'Bruce Cragun'; jamsden@us.ibm.com; Chris Kaler (Exchange)
Cc: gclemm@atria.com; dgd@cs.bu.edu; ejw@ics.uci.edu
Subject: Re[2]: Version issues


     I would support a three level model such as Chris is suggesting.  From
     my experiance it matches a rough segmentation of the market: Low,
     Medium, High.  I can understand resistance to N levels, especially if
     they are perceived as arbitrary.  My opinion is that three levels are
     just about right to capture the various market needs.

     On the other hand, the benefits of going to a three level model
     diminish if it does not discourage mixed-level implementations.

     --Sarge
______________________________ Reply Separator
_________________________________
Subject: RE: Version issues
Author:  "Chris Kaler (Exchange)" <ckaler@Exchange.Microsoft.com> at
SMTPPOST
Date:    3/11/99 1:30 PM


2. The volume of email being generated is overwhelming.
[CK] I think the level of mail is great -- we need to get
     exposure.  However, we need to get it onto the alias
     and not amongst ourselves.  I know it is difficult to
     get all this mail, but the alternative is to have
     too many isolated conversations.

Now, on to my issues with the whole discussion thread.

1. I am against including parallel development in a simple
implementation (Level 1).  The overhead that brings with it is
cumbersome, and the situation of having multiple checkouts on the same
revision is not what I would consider essential for a simple versioning
system.  One checkout at a time ONLY.  If a system requires parallel
development, let it implement it but don't require it for Level 1.
[CK] I agree with you.

2. I believe it *is* worthwhile to discuss simple-vs-advanced at this
point.  I see where you are coming from, Jim, in wanting to postpone
this discussion, but please realize there is also value in making sure
*now* that the levels can indeed be defined.  The issues can be resolved
in parellel rather than having to wait a few months.
[CK] I agree here too.

3. Chris, you seem to have a very solid understanding of the
simple-versioning needs.  Thanks for your support!  About the only thing
we differ on now seems to be the parallel development issue.
[CK] Thanks -- I'm a DM vendor too you know :-).  I don't think
     that parallel development belongs in level 1.  However, I
     strongly believe that basic parallel development is important
     without having to take on all of the level 2 features.  I've
     been thinking that we should consider 3 levels if we can get
     the elements to build on each other: 1) basic versioning, 2) basic
     parallel development, 3) configurations, advanced versioning, and
     advanced parallel development.