LUMIERA.clone/doc/devel/meeting_summary/2023-09-13.txt
Ichthyostega b556d2b770 clean-up: comb through the historical pages to fix markup errors
Some sections of the Lumiera website document meeting minutes,
discussion protocols and design proposals from the early days
of the project; these pages were initially authored in the
»Moin Moin Wiki« operated by Cehteh on pipapo.org at that time;
this wiki backed the first publications of the »Cinelerra-3«
initiative, which turned into the Lumiera project eventually.

Some years later, those pages were transliterated into Asciidoc
semi-automatically, resulting in a lot of broken markup and links.
This is a long standing maintenance problem problem plaguing the
Lumiera website, since those breakages cause a lot of warnings
and flood the logs of any linkchecker run.
2025-09-21 05:40:15 +02:00

124 lines
5.5 KiB
Text

2023-09-13 Lumiera Developers Meeting
=====================================
:Author: Ichthyo, Benny
:Date: 2023-09-14
Sep 13, 2023 on #lumiera 20:00 - 22:30 UTC +
__Participants__
* ichthyo
* benny
* cehteh
_Summary written by Ichthyo and reviewed by Benny_
''''
Focus Topics
------------
_Ichthyo_ started to define a collection of *FocusTopics*.
These are topics towards which we focus our attention for further development *after* the
current Engine integration is completed. Some tickets currently of interest were enhanced
and extended, and new tickets created, which were all marked with the `FocusTopic` Tag.
-> See https://issues.lumiera.org/report/17[Report-17] in the Lumiera Issue Tracker
_Ichthyo_ identified some difficult issues concerning session modifications and change logging,
especially in connection with rules based configuration.
We discussed in some detail event sourcing, session modification, session configuration
and representation of a configuration possibly as a context in the session. Of particular
interest is the portability of a session from one machine to another and, consequently,
the problems that can arise when a change in configuration occurs: these should be
recorded as an event in the Event-Log. However, this requires the representation of the
entire configuration in a consolidated form as part of The Model. A change in the
configuration could then be detected and recorded as an event.
Another task of concern is how the actual Diff Messages will be generated. We could first
manipulate the data model and then _detect_ a diff. Yet the favourable solution would be
to interpret the commands and directly derive the events associated with the commands,
with an attached diff, which can then be applied against the data model as a _projection._
A discussion followed to sort out which topics should be addressed first.
It seems the topics regarding Event Log and Session persistence
https://issues.lumiera.org/ticket/1234[#1234]
https://issues.lumiera.org/ticket/540[#540] and
https://issues.lumiera.org/ticket/569[#569]
should be addressed first, since these may lead to decisions relevant to many other parts
of the system. Possibly also
https://issues.lumiera.org/ticket/332[#332] regarding a full definition of placement capabilities.
The topic of Placement interpretation will be addressed in the *next* vertical slice,
where we intend to add a clip to the timeline. To do this, we'll require at least a
_simple_ resolution capability of Placements, because we'll attempt do the groundwork for
The Builder.
_Benny_ proposes to add »interface« and »mock« tickets for Event-Log and Placements,
blocking the full implementation follow-up tickets. Then these preliminary parts can be
addressed first and integrated as part of the »Add Clip« Vertical Slice. For Context:
currently the »Playback« Vertical Slice is being worked on, aiming towards a first
integration of the Render Engine.
Playback Vertical Slice
-----------------------
_Ichthyo_ reports on the progress of the »Playback Vertical Slice« effort.
- Memory Manager is implemented and tested
- Activity Language is also completed and tested
- a passive worker pool has been implemented, based on the C++17 concurrency framework
Next steps:
- retrofit our existing thread wrapper (which was based on POSIX) to use C\+\+17 thread support;
this should be a drop-in replacement, since our own old design almost completely matches the
design and feature set standardised with C++17
- complete implementation of the Scheduler Layer-2, integrating Activity-Language and worker pool
- integration test of The Scheduler
Documentation Index and Glossary
--------------------------------
_Benny_ reworked and extended the documentation overview in the Lumiera website.
This can be published now.
Moreover, _Benny_ investigated Glossary and Index generation. It turns out that
ASCIIDOC already provides an index generation feature, which could be configured and used
to build an index. But there is no support to generate a glossary which we'll build soon.
The next goal would be to build a prototype for testing and to see how this can be
integrated into the website infrastructure.
Library solutions for Video Display
-----------------------------------
_Benny_ conducted some research regarding library solutions for *video display*, notably
ffmpeg, GStreamer and libVLC. It was rather easy to code up an ``hello world'' example based on
GStreamer, while it was difficult to find suitable documentation regarding the ffmpeg based
solution. Moreover, GStreamer seems to have a much more active and helpful community.
Next steps would be to look in more depth into libVLC, and especially to compare the
feature sets of these solutions. _Cehteh_ points out that MPV (formerly MPlayer) would also
be an interesting point of reference. _Benny_ mentions that GTK4 offers an integrated
video display capability. We'll investigate these frameworks further.
Extended discussion ensued regarding video frameworks and the difficulties regarding the plethora
of interfaces, especially when it comes to hardware acceleration support. Ideally, the Lumiera
project should follow the solutions provided by a sufficiently active community, capable of following
the constant evolution of interfaces and frameworks.
It was agreed upon that the »Playback Vertical Slice« just needs a very simple solution to output
raw video (RGB888 or YUV), and offers a good opportunity for experimentation to find a suitable
support framework.
''''
Next meeting
------------
The next meeting will be Wednesday **Oct 11, 20:00 UTC**.
You are welcome to join.
''''