Rob: Late this week I continued work on the production version of GistView, the Java data viewer portion of the project. At Tiger's request we've moved back some dates in our schedule which in my opinion uses up most if not all of our available slip time. Development on this portion at this point is highly informal, which may work since only Tiger and myself are involved.
Problems encountered include a probably too great relaxation on following a real devepment process requirement on my part which I believe results from a low buy-in by the remainder of the group, and a severe lack of time on my part. My work this week was largely confined to Saturday. Tiger and I have 8 days to get the next iteration of the process finished. Since the coding can take advantage of the work done on the prototype to a very large degree, I think this remains attainable, but I see the potential for problems if the reused code is assumed correct/good/adequate/etc without further verification.
Tiger: Working on java codes, but run into some problems, trying to fix it now.
Tim: I wrote the easy part of the libgist this week: the
library makefile, the
Gist.h file that a user needs to
include, a consistent means of handling the necessary global variables
gist_define_*() calls, opening the logfile, writing
the logfile header, etc. Basically all the stuff I could write
without writing any parallel code.
While working on another project, I wrote some code for dealing with endianness issues with binary files. The code could probably be used almost as-is in libgist, resulting in much smaller log files, and faster reading and writing. The client has shown some interest in this, but it is not a primary requirement.
Xiao: during this week, I researched some on Windows NT performance measures; but so far we did not find very powerful and useful system calls to get some counters information. And also I started coding the Gist library and focused on implementing the Gist_log call.