[CREATE] Proposing OpenRaster Enhancements

[InkLab App]

Hi All,

I had an awesome time meeting and discussing ORA with all of you! I can't
wait to (hopefully) meet some of you in-person next year at LGM in Rennes!

I wanted to tentatively summarize some of the discussion from the meeting
today:

-a single <layer> or <stack> element may now be able to support both
"composite-op" and "blend-op" (names not yet finalized) attributes. This
would allow, for example, a layer to be "clipped" (src-atop composite mode)
and also blended with multiply mode, without needing to duplicate layers,
create extra groups, etc.

-SVG layers will be supported, possibly by the <vector> tag. These layers
will source a .svg file inside of the ORA data folder, and use standard SVG
rendering libraries / specs to render them. IF a program does not support
SVG, the layer should not be shown. If the program can support SVG render,
but not edit, optionally the rasterized layer can be duplicated / edited to
another layer, or overwrite the SVG layer on the user's request with a
warning, etc. (I am simplifying here)

-Animation and Multiple pages will likely require a different file format
optimized for these use cases, and we will likely not touch on this
decision for another year.

-Additional named (implemented in the client program) blend or filter modes
will need to have an open source implementation and use case discussed and
documented before it is considered to be added to the existing list of
blend modes. The proposals should do through the ORA website gitlab and
will be looked at on a case-by-case basis.

-Generic blend modes or filter algorithms (somehow storing the actual
implementation of a blend mode or filter inside the ORA file) seemed to be
a very exciting topic for many of the participants. However, there were
doubts about the efficiency and practicality of supporting something which
may be relatively cutting-edge work. The idea was not dismissed, but a
concrete proof of concept implementation was determined to be needed before
further discussion into ORA integration is considered.

-We didn't have time to go over: Affine Masks, Opacity Masks,
Alternative image containers for CMYK, UUIDs, Palette, Fonts, Undo History,
and additional Composite modes this time, but we can definitely bring them
up for discussion again on the gitlab ORA site!

Thanks again everyone, and I am sorry if the meeting was either too long or
too short for some of you! I appreciate all of your time greatly in
discussing this topic!

-Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/create/attachments/20200530/5ddc05f6/attachment.htm>