On 09/06/14 07:39, Alf Birger Rustad wrote:
Hi James,
I believe most of your questions have been (at least attempted) answered on the
contributions
link of the webpage, and the front page of the wiki.
http://opm-project.org/contrib.php
http://www.opm-project.org/wiki/index.php/Main_Page
Hello Alf,
I do believe that OPM has the trappings to become an outstanding
open-source project. Before I can contribute. I need a gentoo based
working platform; something that I and another are working on. I
strongly prefer gentoo for a variety of technical reason. But first and
foremost is the "structured" solution to what is often characterized as
"dependency hell"; the myriad of softwares that have to have the proper
options (compile and runtime) correctly set for pristine
interoperability between the various codes of a complex system.
But, for me and many other technically astute coders, there is no finer
linux distro than gentoo. It is geared towards a very stable and user
selectable codes that can be compiled and run with thousands of
differing options. Yet the latest git-released codes can be formed into
a temporary package, known as an overlay, for quick installation on any
gentoo system. Gentoo runs on a very wide variety of hardware.
The ebuilds are shell scripts that auto-magically handle the compile and
runtime issues and serve up a variety of "flag" options. Flags allow
codes to be compile and run in differing fashions, but most importantly
to properly inter-operate with other codes. Gentoo supports little down
time between code releases and distro availability. I do greatly
appreciate the efforts of the OPM team and when my development platform
is ready, contributing to OPM. Naturally, I support folks using a
variety of open source distros as best that the project can support
effectively. OPM Needs to run on a variety of servers to ferret out run
time issues. Can you list for the servers now running OPM?
My prediction is that OPM is going to need a robust system to handle
the eventual myriad of compile time and runtime options that emerge from
the development efforts. This is something that a opm-dev group needs to
discuss on a wide basis, imho.
Like most open projects opm is best described as a meritocracy. You will find a
strong
correlationbetween dominant developers and maintainers on each repository.
OK, I agree with this sentiment with no issues! But, are you saying that
the collection of separate module bug trackers are a good idea?
Certainly, it is the easiest way to start a project, particularly when
the relevant codes are wildly separate. As these modules evolve into an
over-arching solution, there will be bugs and issues related to the path
forward, that will be best facilitated with a single bug tracking system.
I do believe all issues of importanceto the project are brought into open
discussion. We
do have a "patches are welcome" policy. However,our community today is
developer heavy,
so we do miss a larger user community. I hope that willchange as the code base
becomes
more relevant for users.
Opm-user as a group would greatly encourage users to test the offerings
and offer feedback and ideas and overall "inspire" many about the
project. An active (and encouraged) user community is often the
life_blood of an opensource project. Also, there is a strong movement,
particularly among the smaller and mid-sized (100MM to 1B) organizations
to standardize much within the simulation, modeling, Big Data,
Geological, Regulatory, and data (set) management aspects of subsurface
physics. These forces can all be positive influences for OPM. OPM could
easily become the "Blender" of subsurface physics, imho.
Ah yes, my other suggestion opm-dev. Forming a group to discuss code
development issues (ppm-dev) and a separate group for users (opm-user)
is warranted? Many folks could read what the devs discuss and gain a
deeper understanding to the codes, the overall interaction between the
codes and where development muscle is needed. You have many college age
kids in fields such as Geology, Petroleum and computer science that
could discover a plethora of interesting physics (math) from a very
practical exposure to the OPM develop, discussions, and decision trees.
In any case, opm is an open project,and I do not believe decision processes are
more
closed here than in other open projects. Typicallykey developers will make
design
decisions simply by contributing code (like the current situationwith multiple
grid interfaces).
Now I like what I'm hearing (surprised?) The best coders always win in
open-source. If for no other reasons they become better (stronger)
coders and deeper subject matter experts. But, you miss the point. A
DISCUSSION, by all willing to "throw_down" on what is being discussed,
is quintessentially important to arrive at the most robust, often
competitive, solutions. In gentoo, we have choices where sharp minds
disagree. Via what we call the "flag" option, one could choose, at
compilation time, whether or not to use a particular mathematical
library A or B or C; which would be coded up by those with different ideas.
Like it or not for OPM to be the "King of porous medium codes", it's
going to have ferment the choices of the various mathematical
approaches, data handling and heuristics on how to render the resulting
graphics. I see no discussion on the framework to allow (encourage) the
various coders to discuss, disagree and develop competing frameworks
for their vision of the best solution.
Ultimate, who decides the best solution? Simple, it's your USER
community. So opm-dev open-user and a singular, robust bug tracking
system are requisites. I have no problem with fiefdoms particularly
those coveted by developers. My whole life, besides being intellectually
aggressive, has been spent in competitive sports. Competitive sports has
an over_arching theme that is supremely important. Teams (solutions)
often look very strong and superior. But, you must "play the game" in
order to find the King. Run the competing codes, display the graphical
results and gauge the time line and accuracy of the result. Then a
tournament King will be (temporarily?) crowned.
So let us discuss and restructure the tournament, so as to attract a
myriad of brilliant minds to openly compete? Or is OPM like middle
Europe back in the age of castles, queens and sparse_cannons?
We are still a young project though, so we will hopefully get moreof the
expected
infrastructure of an open project in place over time. Make no mistake though, we
aim to run opm as an open project.
Alf. Buddy? Were talking about installing Bugzilla and forming a second
discussion group and renaming the discussion groups to clearly delineate
function. Surely with the current brilliant minds dominating OPM, these
merely trivial tasks do not need time, but action?
Surely the 'old guard' at OPM is not afraid of a wee bit of competition?
Cheers,
Alf
I apologize in advance if my ideas or words offend anyone. But let's get
this project kicking, open and showing some simulations on the net?
Separate bug trackers are idiotic, at best. Defend your fiefdom(s) as
the surfs are rising up? (love that old euro baggage .....I wanna be a
Viking someday).......
curiously,
James
_______________________________________________
Opm mailing list
[email protected]
http://www.opm-project.org/mailman/listinfo/opm
