History in the Squeak image (was: [BUG] Problem with renaming cla
chrisn at Kronos.com
Wed Oct 9 18:59:46 UTC 2002
On Wednesday, October 09, 2002 (2:24 PM), Ken Causey wrote: "Guess I need to
decide how important history in the image vs. history in CVS is to me."
This comment jiggered a memory of a question I asked on this list some weeks
ago regarding developer initials and timestamps on methods in the Squeak
Image. At the time, Dan Ingalls and Scott Wallace pointed me towards some
code and lent me their knowledge on the physical structure of methods in the
image -- thank you Dan & Scott for your assistance! :-)
The intent, or goal, I originally had was very much related to history.
Browsing through the latest image with annotation panes turned on, you will
quickly note that many methods have no historical information. These
methods do not have authors initials or timestamps. This bothers me on at
least two levels.
1. Is this method old or new?
2. Who deserves credit for this work?
Another problem I see, is the fact that we don't know who wrote a method
before it was changed by, say, me. Given all of the recent news about
digital rights, I think these may be important questions; it's always nice
to give credit where credit is due.
So, last night I went onto the web and found a Squeak 1.2 image, thinking
that I could (somehow) compare the code from that image with the code in the
latest image. Even if the 1.2 image did not have author initials, I might
at least be able to label the "undocumented" methods with something like:
>From Squeak 1.2, 06/29/1997 4:42 PM. Unfortunately, the .sources file
appears to be missing from the various sites that have a 1.2 image.
What would be nice, in my opinion, is some way of labeling the
"undocumented" methods in the 3.2 & 3.3 with the date where the method first
appeared in its current form. Another nice to have would be a simple
revision history of the people who changed the method since its creation.
I'm not suggesting that we retain all of the old code (that would be
unnecessary bloat), but it would be nice to know if a particular method was
written by, say, Dan Ingalls, back in 1972. :-)
An added benefit of a history mechanism, would be a list of developers who
were, at one time, familiar with the code that is currently kicking you in
the head. <grin>
More information about the Squeak-dev