2:01 p.m.
Hi everyone,
TSC already contains a lot of original art that is not found elsewhere.
While our current license does permit reuse of most assets, I would
like to mark their origin in the TSC project.
SVG files contain metadata in form of the Dublin Core tags. They should
be used to indicate they're from the TSC project, e.g. like so:
<dc:source>The Secret Chronicles of Dr. M.</dc:source>
<dc:relation>http://www.secretchronicles.de</dc:relation>
Inkscape allows you to easily set this metadata via the document
settings dialogue.
The OGG music container format also supports metadata. For these, I
think it would be good if the music files were marked to belong to the
"album" "The Secret Chronicles of Dr. M.", and if possible, include
the
link to http://www.secretchronicles.de.
I also think that the license information should be included into the
metadata of assets, but as we currently have the .settings and .txt
files for this, please don't start editing the licensing information
into the metadata now. We'll first have to determine if this is
possible for all formats we use.
Any artist of course is free to add the name of the work and his name
as author to the metadata of any asset. This information will not be
removed.
Does everyone think that normalising the metadata is a good idea?
Valete,
Marvin
--
Blog: http://www.guelkerdev.de
PGP/GPG ID: F1D8799FBCC8BC4F
4:45 p.m.
I think including the origin of assets is a great idea. I found it
useful to include the URL of websites from which I found sound effects.
In some cases, we may not be able to find a URL, but including what we
do know about the origin is a good idea. Also, keep in mind that URL's
may eventually go bad.
I'm a bit hesitant with the idea of getting the origin into every
Inkscape file, as it sounds a bit tedious. This would let the artists
set it themselves really easily, though many of them will probably
forget anyways. Also, I'm not sure if there's an equivalent for this
info within music files -- I think it might be more consistent just to
use text files for both images and sound/music files.
It would make sense to make this change in the current repository before
we start the new codebase.
-datahead
7:02 p.m.
It's time for another one of Luiji's Random Sporadic Drop-ins!
Information about metadata with OGG files is surprisingly
inconsistent. However, normalizing metadata would be a great idea! I'd
be willing to write up a Ruby script to verify metadata in the assets
as part of a consistency check if you would like. Something like:
$ ./check-asset-metadata
data/here/there/however.format: dc:relation should be
"http://www.secretchronicles.de" but was not set
data/there/like/this.another: ALBUM should be "The Secret Chronicles
of Dr. M" but was "Sample"
2 error detected.
$ exit
(It'll even have colorized output. :P)
I imagine we'd want to have the metadata on PNG files, too, yes? You
use XMP metadata in an iTXt chunk for that. If we do that, though, we
have to check and see if optipng will mess it up.
-Luiji
On Sat, Jul 23, 2016 at 12:45 PM, Chris Jacobsen <cjj_009(a)yahoo.com> wrote:
I think including the origin of assets is a great idea. I found it
useful
to include the URL of websites from which I found sound effects.
In some cases, we may not be able to find a URL, but including what we do
know about the origin is a good idea. Also, keep in mind that URL's may
eventually go bad.
I'm a bit hesitant with the idea of getting the origin into every Inkscape
file, as it sounds a bit tedious. This would let the artists set it
themselves really easily, though many of them will probably forget anyways.
Also, I'm not sure if there's an equivalent for this info within music files
-- I think it might be more consistent just to use text files for both
images and sound/music files.
It would make sense to make this change in the current repository before we
start the new codebase.
-datahead
8:20 p.m.
Hi Luiji,
On Mon, Aug 01, 2016 at 03:02:56PM -0400, Luiji Maryo wrote:
It's time for another one of Luiji's Random Sporadic
Drop-ins!
It is nice to see you around here again! Here is the latest information
for you: we have decided we are going to rewrite TSC from scratch. Thus,
it is now a good opportunity to take influence on the new code
architecture and even re-join development again on a clean codebase if
you like. Take a look at this ticket:
https://github.com/Secretchronicles/TSC/issues/534
Meanwhile, we are going to release 2.1.0 probably soon (see my recent
message about this). If you feel like doing some testing, try to compile
and run `devel'.
Feel free to give your opinion here on the ML if the tracker doesn't fit
you. I'd love to see more discussion on the ML anyway. Just look at how
the ticket encapsulates multiple discussions in a linear line, which
would have had a nice threaded tree in an ML thread!
Also, didn't you want to hang around in #secretchronicles
occasionally? :-P
Information about metadata with OGG files is surprisingly
inconsistent.
Hm. ffmpeg displays metadata on OGG files, so some kind of standard has
to exist there.
However, normalizing metadata would be a great idea! I'd
be willing to write up a Ruby script to verify metadata in the assets
as part of a consistency check if you would like.
Sure. However, we have to gather all the metadata of the assets before
this is going to be really useful, and that probably is quite a bit of
work on its own. As of now, nearly no asset has any metadata set.
I imagine we'd want to have the metadata on PNG files, too, yes?
You
use XMP metadata in an iTXt chunk for that. If we do that, though, we
have to check and see if optipng will mess it up.
I don't think so. The PNG files are generated from the assets in SVG
format (except for very few cases we urgently need to replace) so one
can just as well take a look at the original SVGs. For the rewrite, we
have even been thinking about using sprite sheets, which would have no
useful place for this information. What would be nice however is to
generate a list of authors of asset files automatically from the
metadata in the assets so that it can be included into distributions of
the game to fulfill the author naming duty most licenses require.
-Luiji
Vale,
Quintus
--
Blog: http://www.guelkerdev.de
PGP/GPG ID: F1D8799FBCC8BC4F
12:02 a.m.
However, normalizing metadata would be a great idea! I'd be
willing
to write up a Ruby script to verify metadata in the assets as part of a
consistency check if you would like.
If someone borrows a file from the project, they will probably just dump
it into a new project, without thinking to look at our license text
files. Putting meta data into the content files is actually a really
good idea, as it strongly encourages people to attribute the source and
abide by licensing. I still like the idea of having text files for
quick reading, though.
For the rewrite, we have even been thinking about using sprite
sheets, which would have no useful place for this information.
Are you saying that if multiple authors contribute subimages within a
sprite sheet it is hard to attribute ownership?
-datahead
8:38 a.m.
Hi,
On Mon, Aug 01, 2016 at 07:02:23PM -0500, Chris Jacobsen wrote:
If someone borrows a file from the project, they will probably just
dump
it into a new project, without thinking to look at our license text
files.
He may do that, but if he doesn't follow the license terms, it is a
license violation on his part.
The license files are just our way to fulfill the naming duties from
various licenses.
Putting meta data into the content files is actually a really
good idea, as it strongly encourages people to attribute the source and
abide by licensing. I still like the idea of having text files for
quick reading, though.
Yes.
> For the rewrite, we have even been thinking about using sprite
sheets, which would have no useful place for this information.
Are you saying that if multiple authors contribute subimages within a
sprite sheet it is hard to attribute ownership?
How do you want to include that into the metadata of a sprite sheet
which contains dozens of different images? Thus, the attribution
requirement is easier fulfilled by adding a separate text file listing
all authors.
-datahead
Vale,
Quintus
--
PGP/GPG ID: F1D8799FBCC8BC4F
Date: Tue, 02 Aug 2016 10:37:57 +0200
Message-ID: <8760rjftwa.fsf(a)hades.cable.internal.west-ik.de>
1671
days inactive
1681
days old
tsc-devel@lists.secretchronicles.org
6 comments
3 participants
participants (3)
-
Chris Jacobsen -
Luiji Maryo -
Marvin Gülker