Age | Commit message (Collapse) | Author | Lines |
|
Until now the directories were light green (similar to the Make files
in the data lineage plot) which was confusing. Now that they have a
different color, they can be better distinguished as directories
vs. files.
|
|
The file architecture plot was a little crowded, so I am now first
showing the files discussed in previous plots, and then showing the
whole thing. Also, the 4 commands necessary to reproduce a project are
added.
|
|
I just made this for the paper and thought it would be good to have
here is well.
|
|
The journal name and volume was added, also because its a little long,
the main terms were made all-caps.
|
|
This quote (that a paper is just an advertisement of the scholarship,
not the scholarship itself) beautifully summarizes the problem!
|
|
The image that is used for replicability was intentionally chosen
because it has an astronomer sitting behind the secondary mirror
(showing the hard work that goes into data collection). But the image
has too many details and the person can be completely missed by the
viewers. So I just darkened the outer parts and added an arrow to let
the audience understand the point of the image.
|
|
|
|
The previous definition of hardware/software reproducibility was
mainly my own. But I recently came across the National Academies
report that gives a precise definition, so I used those terms and
cited the National Academies report.
|
|
A verification step was recently added to the pipeline, so it was
necessary to add it here is well.
|
|
Since some of the audience is not yet fully familiar with Git, I
thought its necessary to remind them in the slide showing Git
branching, that every commit contains all aspects of the project.
|
|
Following the convention of the Make-demo plot (where green files are
input and blue files are built), the first graph's gray boxes were
also changed to blue.
Some other minor corrections were made.
|
|
The set of slides that show how the built files relate to each other
through Makefiles was edited to first show the final target, and track
it all the way to the raw configuration files. This is in-line with
how I describe Make (that Make starts from the end).
This great suggestion was made by Idafen Santana PĂ©rez.
|
|
To help the audience, the legend of the graph (the description at the
bottom of the page) now also says that the built files are shown in
the Makefiles that produces them.
|
|
Until now, the connection between the first and second slides (after
the title) wasn't clear, so the title and fir paragraph of the
NoiseChisel+M51 slide was slightly modified to be more in-line with
the previous slide.
|
|
As Raul pointed out, to avoid confusion for a color-blind audience, or
when its printed in black-and-white, its best to not just base the
source-build distinction in the graph based on color.
With this commit, the source files have sharp edges and the built
files have round edges.
|
|
After showing the graph (and how everything is connected), it was
useful to remind the viewers on the benefits of Make to create this
structure.
|
|
It could be confusing, but we are only using LaTeX in the template so
far, so its better to just use LaTeX.
|
|
Some minor modifications were made in the graph showing the
organization of the analysis with Make. Also, a `tex/preamble.tex' was
defined to simplify the main source.
|
|
Until now, there was only one slide, with many boxes and arrows
connecting them. It was too complicated to understand for the viewers.
With this commit, it is broken up into separate slides, with each
box/arrow added on each slide as we progress. This helps clearly show
the logic behind all the connections.
The LaTeX source of this graph is in the paper that describes the
project, I will later bring that source into the slides too (and
remove all the extra figures).
|
|
Some arrows and text were added over the shown Makefile to better help
the eye when the viewers aren't familiar with Make.
|
|
Until now, as an example of uploading to Zenodo (besides uploading to
arXiv), I was mentioning `zenodo.1164774'. But this link doesn't have
any software to show. So I just updated the Zenodo link to the more
recent upload of `zenodo.3408481'.
|
|
To help in continuity with the adoption funding possibility slide, its
better that the slide with the RDA grant notice is just before it.
|
|
The link short description above the link on the first slide is now
more clear.
|
|
Given that this project is now recieving funding from RDA EU 4.0, it
was necessary to add an EU flag/description for it. With the new EU
flag, it was necessary to re-arrange the logos to be more visually
appealing.
|
|
A standard GNU GPL copyright notice was added to `README.md' to be
complete.
|
|
The newly created `README.md' gives a short description of the
project, with a link to the final PDF, and the main template Gitlab
page, along with a short description of the LaTeX dependencies and how
to build the PDF.
|
|
The Git commit checksum has been added to the first slide to help
readers identify any change.
|
|
With this copyright, a random viewer will know their rights regarding
these files (that they are free software and that they can modify them
and redistribute them).
|
|
Until now they occupied a major part of the center of the first slide
and in two rows. With this commit, they are in one row and at the
bottom of the slide.
|
|
It helps the eyes find the Git checksum while also connecting it
visually with the previous slide showing the pictured happy
researcher.
|
|
A few other minor corrections were made to make them more clear.
|
|
I had mistakenly written `1/5' instead of `0.5' in the default S/N
value. So I changed it to `1/2' to be more comparable with the
optimized `1/4'.
|
|
Several figures were added to highlight the major points and add
better examples.
|
|
It is helpful to the argument to show the matplotlib dependencies
after we show the Makefiles used to build the programs. Also, since
work on the GNU C Library has already started, the colors of
NoiseChisel's GNU C Library components have been corrected.
|
|
To help with the discussion, that plot was also added after showing
the usage of LaTeX macros.
|
|
The information on the presentation place was updated for the RDA AMA
webinar and all the necessary logos that should have been added before
were now added. Also a diagram showing the relation between the source
and build files has been added after the introduction to Makefiles.
|
|
I had mistakenly forgot to include the four displayed M51 images into
the Git repository of these slides. They were being read from a
different directory on my system!
They are now in the `img/' directory and all references to them in the
LaTeX source of the slides has also been corrected.
This issue was raised by Charles Twardy.
Closes #4
|
|
Only the first slide (marking the location of the talk) has changed.
|
|
Everything else is the same, only the time and date in the first slide
are updated.
|
|
Due to a typo (incorrect macro naming), the credits of the "verified"
stamp were not being shown in the slide. This has been corrected.
|
|
The fact that we only displayed the Git checksum before showing it in
the paper could be a little confusing to people not much familiar with
Git. So a commit checksum (taken randomly from the history of these
slides) was added to it.
Also, some minor changes were done here and there.
|
|
A minor editing was made to make the NoiseChisel slide more
clear. Also the Repology information on Gnuastro and Astropy was set
to today.
|
|
I had used two `using's in one sentence. The second is now changed to
`with'.
|
|
The first slide was modified for this talk.
|
|
Just the first page was updated.
|
|
Since later we have Gnuastro on the right, it was set to the same
order here.
|
|
Until now I was only showing the good NoiseChisel results which don't
show the effect of paramter optimization. With this commit, we are
showing the default parameters, the optimized parameters and the deep
image.
|
|
Two images were taken from `repology.org' showing how Gnuastro and
Astropy's builds on different systems differ.
|
|
The slides were updated (most importantly removing references from the
IAU meeting) for a talk in Collibra.
|
|
To make things easier to understand.
|