On Wed, 5 Mar 2003, Neil Gorsuch wrote:
> Jeff and I talked off-line, and will continue the discussions. We'll try
> to come to some kind of a compromise proposal before the next meeting.
I think the e-mails so far have clarified a lot of what we're agreeing and
disagreeing on:
+ We agree that the installer needs to be re-written. It's horrible
as it is now. It would take just as much work to fix it as it would
to re-write it, so we might as well make a good plan and re-write it
Right (hah).
- We disagree on what it needs to be re-written to be. Jeremy has
pointed out some interesting tools that we might want to look into.
Jeremy's a better web searcher than I am. :-) Tom Naughton also has
some interesting information in this area.
+ We agree that we need an installer GUI. As was pointed out on the
call, we all want a GUI. This is obvious. It is likely to be the
entry point and main installer for 99.99% of those who use OSCAR for
the first time. Hence, it has to be functional, professional, etc.
- We disagree on the relative importance of the GUI. Specifically, we
disagree on definitions of "professional" and "pretty" as applied to
GUI's. At the core of this, IMHO, is the defition of our target
users. I think we can agree that OSCAR has users ranging from
ametuer high school students to advanced research center sysadmins.
I think we can also agree that OSCAR is not software for your
gramma.
There's more to say about all of these, of course, but are those at
least generally right?
I think the following issues are not fully sorted out yet, and these
statements may not be fully true:
- We disagree on how much of a role a GUI should play in the role of
OSCAR design.
- We disagree on the importance of a CLUI (and perhaps even on what
the definition of a CLUI is).
- We disagree on the importance of having a unified I/O interface.
There's *lots* more to say about these three bullets. I'll talk with Neil
and the NCSA crew about these and all the other issues before the call
next week. Comments on the mailing list are good, too.
> And we're really not picking on Jeff, it just looks that way 8-).
Yeah, RIGHT! ;-)
Honestly, the disagreement is good -- differing opinions are good. We'll
eventually reach consensus, never fear. This disagreement is *why* we
have a group consisting of members from different organizations, after
all, and why OSCAR is not just one person's pet research project. This is
not a flame war, even though it may look like it to outsiders.
That being said, I agree with selected parts of the posts that have come
out of the NCSA group. Sean Dague has said from the very beginning (I've
been trying to parrot his words, but apparently unsuccessfully), and Terry
mentioned the same thing: we need to create a detailed flow chart of the
installer steps first, and then design the interface.
I even took a stab at this once (on a plane back from SC), and although I
didn't finish it, I used that as the basis for the MM design.
Unfortunately, I must go back to working on a proposal that is due
tomorrow. Ugh!
--
{+} Jeff Squyres
{+} [EMAIL PROTECTED]
{+} http://www.lam-mpi.org/
-------------------------------------------------------
This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger
for complex code. Debugging C/C++ programs can leave you feeling lost and
disoriented. TotalView can help you find your way. Available on major UNIX
and Linux platforms. Try it free. www.etnus.com
_______________________________________________
Oscar-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/oscar-devel
