On the web page, it says:

"Maintenance releases are more frequent and are comprised of bug fixes
and small refinements, and should not affect source functionality."

If there is a bug fix, or a refinement, wouldn't that improve source
functionality?

Would it be right if I say something like...the functionality of the
source code base is not affected? (Even that doesn't make sense to
me.)

Obviously, I'm stuck on nomenclature.  I think what is confusing me is
that I'm not sure what you mean by "source".

Again,  thanks for your patience with me;

The source code is just another name for the computer program. FDS is
a very large fortran program.

When we fix a bug, we are ensuring that the program is doing what it
was designed to do. For example, the user's guide might explain how to
output a certain quantity, but when the user tries to do it, that
quantity might show up as zero, or -9999, or whatever. This is clearly
a bug where the program contains an error that somehow over-writes the
computed value of the quantity. We fix it, and then release a
maintenance version. We then encourage everyone to upgrade to this
because we don't want anyone to think that the output quantity is zero
when it is not.

A change in functionality might be where we decide that a particular
default value should be changed to something else, or where we decide
to use a different algorithm for, say, smoke detector activation.
These types of changes require a minor release, because the model user
ought to be aware that the mathematical model has changed slightly.
Here, you might want to not upgrade if you are right in the middle of
something. However, you don't want to stay too long at a particular
version because we do not make bug fixes to older versions.

Hope this helps.

On Feb 4, 11:37 am, Gail <gwestle...@gmail.com> wrote:

Major and Minor Releases change, add or remove program functionality,
while maintenance releases repair or restore the intended
functionality to the other release types.

-Bryan

On Feb 4, 11:37 am, Gail <gwestle...@gmail.com> wrote:

Are you trying to determine when it is you would need to re-execute
the V&V plan to meet NQA-1?

If that is the case then it may be dependent upon 1) what the bug was
that was fixed, and 2) how you define your acceptance criteria in the
V&V plan.  A number of bugs that we fix are fatal errors that occur
during input processing or during runtime when a user tries to do
something we haven't anticipated.  For a bug fix like that, it isn't
likely to impact the results of any V&V case, as those cases ran
successfully.  Some bugs that are fixed are errors within an
algorithm, and those could result in changes in the specific numerical
values produced by FDS but perhaps not the overall performance (FDS
predictions might vary at an instant in time / location but the
overall prediction will remain unchanged).  In a case like that
repeating the V&V suite might in large part depend upon the acceptance
criteria.

On Feb 4, 4:18 pm, bryan.kl...@nist.gov wrote:

"Maintenance releases are more frequent and are comprised of bug
fixes
and small refinements that restore the intended functionality of the
previous release of the program.  A bug is defined as being an error
that somehow over-writes the computed value of the expected quantity.
A number of bugs that are fixed are fatal errors that occur
during input processing or during runtime when a user tries to do
something that is not anticipated by the developers.  For this type of
a bug fix, it isn't likely to impact the results of test cases
performed for the V&V.

Other bug fixes are errors within an algorithm, and could result in
changes in the specific numerical values produced by FDS, but not the
overall performance.  In this type of a bug fix, rerunning the V&V
suite is dependent on the acceptance criteria.”

Following a maintenance release, it is encouraged by the developer to
upgrade FDS to ensure that the output quantity is correct.”

Again, thanks for you time.

Gail

On Feb 4, 8:37 am, Gail <gwestle...@gmail.com> wrote:

We consider a "bug" to be an error in the software or computer
program, not the model itself. In other words, the model is the set of
mathematical equations that we are solving. In a maintenance update,
we are not changing the model, but rather fixing a glitch in the
program that implements the model. When we do a minor release, we are
changing, even if slightly, the mathematical model.

Bryan's text above is fairly well-stated. I would use that.

On Feb 5, 11:52 am, Gail <gwestle...@gmail.com> wrote: