Right now, I don’t have a good idea of who is actively using PLR. It seems like there are at least 3 people developing it: @rickwierenga@CamilloMoschner and me. Possibly there are more people developing on top of it?
For the most interesting use-cases, developers are likely busy & also need to keep their developments private to protect IP in their drug programs.
If we don’t record the meetings, then they aren’t as useful as the forum, because new people won’t be able to discover what we’re working on asynchronously.
Some work is going on in the background. Marielle from Sculpting Evolution is working on pump arrays and @vkorotkov is working on integrating their Crystal Powerdose powder dispenser (front end already merged).
Either recordings or well-written notes are extremely important. I wonder if whisper + gpt could transcribe and summarize automatically
I think this is a fantastic idea, and I would definitely join.
IP security for industry contributors
I agree with @ben that appropriate IP management is of top priority to industrial collaborators, and since most academic labs cannot afford the robots we are aiming to develop PLR for, the main PLR target user is likely industry-based.
But this is a common situation for many companies (github.blog: Why more companies are investing in open source program offices). So it is the responsibility of individual contributors to ensure their companies’ interests are maintained
After all, anyone can make a PLR copy and develop it in house. The question is then why develop PLR:main openly, and there are many answers here including developing speed, maintenance, crowd-sourcing problem solving (which has time and time again been proven to be the most efficient strategy; see wazoku), and many more.
Ideas regarding meeting
Yes, I think a regular meetup of something like a “PLR Board” that oversees and structures the development of PLR would be very beneficial. I found forum threads are very good for ideas and troubleshooting but bigger projects require more organisation that I don’t believe a thread can provide.
Therefore the purpose of the PLR Board would be to guide and oversee the development of PLR in a structured manner.
Like with most societies/clubs/projects the actual number of people joining will fluctuate but usually a small core team that contributes consistently emerges. I believe in the powerful incentives of PLR to attract people over time
I think every other week is a bit too frequent. Since most people will not be able to work on PLR dev full time there might not be enough progress to ‘oversee and guide’ in 2 weeks, which then raises questions of the utility of the individual meeting.
Once a month at a fixed timepoint, like first Tuesday, is likely easiest to maintain over an extended period. If we have to fix the PLR board meeting every time around all people who might want to attend it will be very difficult.
To get the okay of industry collaborators to join the meeting during company time it would also be vital to stick to a maximum time limit of e.g. 1.5 hours for planning and approval purposes.
Yes notes are very important. But recordings are a handicap in my opinion: imagine having to go through 24 recordings of 1.5 hours after just 2 years of doing this… the time to go through these would be too long.
Well written, concise notes should be able to be read in 10-20 min.
Yes, and no: I have used whisper before to transcribe recordings and it works very well for seminars/lectures (i.e. one speaker). I have not yet tested it for recordings with multiple speakers and therefore don’t know whether it would recognise our voices automatically.
But as long as it records and summarises all the information of the meeting, I can see a fast minute-taking workflow:
generate voice recording of board meeting
transcribe with whisper
summarise with GPT4
maybe ask attendees to check accuracy and add images they might have shown during a meeting to the summary (where appropriate)
upload to development meeting minutes page/folder on the forum(?) for everyone to freely access and maintain and open record.
What do you think?
Potential topics to discuss in board meetings
What does ‘guide and oversee the development of PLR in a structured manner’ mean?
We could discuss and divide workload amongst volunteers for (i.a.)…
new implementations (modules, functionalities, entire new robots to add,…) and their priority for the open-source community
the development of a “PLR Guidebook” (a PDF with a lot more documentation and examples than fits into the docs, and allows people an easy reading option; also less intimidating to newcomers); a .tex document hosted on the GitHub repo would work well I think
a written-down code styleguide that defines the PLR-best-practices (I was a bit thrown off by the two-space tab indent but am happy to adjust to anything consistent)
necessary code change to enable upgrade to Python 3.11 (that funky business with asyncio’s coroutine decorator…)