Session Title: Make a collaborative project for GOSHers
Date: 3/25/17
Attendees (who was there?):
Overview of topic (3-6 sentences):
Notes:
What kind of project do we want to do:
Guidelines:
Open
If we can scale cost down, we dramatically increase user base
Impact
Somehow broader functional appeal than just current open science movement (maybe simple to understand it’s purpose, implies low cost
Tiers of complexity
Room for collaboration outside the initial group
Create a successful example for how these kinds of collaborations can occur in the future in OSH.
maximizes global accessibility if at all possible
What can we make
Universal power supply xxx
Mass spectrometer xxxxxx
Raman Spectrometer xxx
UV-VIS-IR spec xx
fish tracking device xxx
human health…
environmental health…
CO2 sensor xx
pH probe xxxxx
low field MRI xxx
NMR xxx
gene sequencer xxxx
confocal microscopes xx
lens free microscope xxxxx
light tomography xx
space ship xxxx
smartphone with range of sensors integrated xxxxxx
autoclave - low power or solar power xxxxxx
Solve the supply problem in countries that have trouble importing
modular sensor connector xxxxxx
DIY solar cell manufacture xxxx
nitrate sensor xxxxxxx
train a neural network on open science hardware documentation xxx
smart toilet
time of flight xxxx
purify bioelectric signals to get rid of the background noise x
single pixel camera
Yes, I think there have been many good ideas! First, just want to lay out the goal of the collaboration (besides fun) -->
Can we collaborate effectively among many people on an OSH project? What are the best ways/tools to collaborate? What’s the best means to organize the collaboration?
Can we make something that shows that GOSH and a collaborative development can produce a influential, impactful piece of hardware?
I think next steps is to decide what we work on. There are some which have floated to the top. Perhaps everyone who wants to champion one specific piece of hardware can just reply with the reasons and reasoning, and we can try to come to consensus. Or if there are a few items that people are passionate about, we could have 2 groups.
I’ll start - Personally, I think the single pixel camera is possible, could be low cost and transformative for many people already in GOSH. The following is mostly a rephrasing of what Richard taught me yesterday plus some (maybe) clever ideas to lower the cost A single pixel camera allows you to use a single detector, which could be a (short enough wave) radio antenna, an infra-red (heat) image, or the full spectrum using a spectrometer (350 - 1000nm with 10nm resolution). Building a camera with many pixels to do this would be prohibitively expensive. I’m particularly excited by the hyperspectral imaging, which are commercially very expensive ($1000s - $10,000s of dollars).
It could potentially detect pollution from images or video, could be used to identify food quality, in agriculture to find areas of low nutrients or water, to provide hyperspectral imaging on an openROV, and in general purpose lab applications.
Also, should we start on github or gitlabs? I’m never used gitlabs but it’s fully open source. We could make GOSH as an organizations and all join it.
I’ve been using both github and gitlab for a while and I don’t notice a huge difference between them. Gitlab isn’t 100% as polished as github, but it is very close and supports a similar workflow. It is certainly quite usable and I’m a firm believer in using open source tools wherever we can.
In either case, I think having a wiki and issue tracker are essential for organizing this type of collaboration (both are included with either gitlab or github). My vote is to setup a GOSH community on gitlab and start filling in some details on a wiki there.
As for comparing these different projects, can we propose a few different metrics to judge them by? I would like to see:
Information on an existing commercial version (if it exists). How much does it cost? How many people use it/who uses it/what do they use it for?
What would we gain from an open version? e.g., Reduced cost? How much?
Identification of relevant IP
What new applications would an open version bring? i.e., what is the potential for a new project to be disruptive to the status quo?
How well would it scale? This relates to how easy it would it be to build once we have a working design
Anyone else have ideas for criteria to judge projects by?
I also love the OpenDrop (but I’m biased). My goal is to have OpenDrop and DropBot converge at some point in the not-so-distant future and would love to recruit anyone interested to the cause of building an awesome, open DMF/digital biology platform.
I think this has the potential to become the GCC compiler for biology/chemistry.
My intervention is based on @gbathree clarifications:
For me, we should keep in mind that 2 forms of collaboration can be distinguished: Local/national collaboration & international/online collaboration. For the last, platform and web are the keys to succeed. But for the first, thoughtful for my african context, I have proposed the idea of Sciences Shop, through which we can collaborate with grassroots and know their real needs. Then the science shop manager can share and collaborate with you everywhere (international/online collaboration). I am aware of the different existing contexts, but by this patterns GOSH can have a real impact in the global south.
What can we do? I can be interressant if GOSH engage in Green projects (solar, waste management…). For example, if we consider plastic waste management, which is a real problem in my country (and others). Can GOSH create or reproduce apparatus of plastic recuperation like presented by precious plastic? This recycled plastic can be used to create others objects. Such project can be very impactful, since at the same time we are dealing with environnement, we can also create a lot of jobs, making GOSH engage in circular/green economy which is anticapitalist.
Yeah Thomas, totally agree. I think part of the purpose of open development is that it can be modified and applied in different contexts. It also means that for international collaborations, regardless of their location, everyone can keep in mind a wider range of applications while developing. This may not mean that the finished product is 100% appropriate for those applications, but it may mean it’s much closer to being appropriate than it would have been otherwise. That sounds wishy washy but I swear it’s not, the MultispeQ is a good example of keeping many applications in mind for a single instrument.
Wow that precious plastic link is cool, and it’s actually already open source, plans and all. Has anyone built one in your area?
Also, in terms of the single pixel camera…Here’s an interesting link from a paper that @zegarac.robert shared, perhaps an even simpler strategy -
So for the single pixel camera / hyperspectral imager, Ryan made a gitlabs page, as well as a gitlabs account for the GOSH Community collaborations:
I started writing out some Background, Design Ideas, and Applications for the single pixel camera. Please add your contributions here!
If you’re not familiar with how Git works, just create log in at gitlab.com, and request access to join the the project. Once joined, you can just edit those documents directly in the browser, or you can clone the repository and do it on your computer. GitLabs and Github are useful because they do amazing version control, and allow many people to collaborate without stepping on each others toes. We can also store all our other files (firmware, kicad, design, etc.) here as well.
If you want to make a different project to collaborate on (like some of the other ideas that emerged from our session), join the gosh-community and create more projects.
Might be an idea to enter the single pixel camera (and any other ideas) into the Hackaday Prize which gives you some deadlines and the potential for some funding. The first round is especially for submitting concepts rather than ongoing projects and closes at the end of April.
Yeah, let’s do it! Actually May 1 is a great date to set. Maybe folks interested could have a call and divy up some of the work and go from there - what do you all think?
Here’s a good video walking through the different current technologies for creating hyperspectral cameras.
Also, I’m really interested in applying for the hackaday prize for the single pixel camera - is anyone seriously interested in that? @kaspar@ryanfobel@rbowman maybe @gaudi or @dusjagr … anyone else?
I am super interested in this project, but am totally swamped right now (definitely won’t have any time before the first round of the Hackaday prize). I could probably find some time to make meaningful contributions prior to the round 5 “Anything goes” (September 4th - Ends October 16th).
I am afraid I am too busy to dedicate any significant amount of time at this point as well. I would enjoy watching it develop and helping out with little bits here and there if I can.
ok I’ll keep pushing it forward as I have time - I’d love any comments or support from anyone if/when you all have time. I have a conversation with hamamatsu corporate today about the project, I’ll post here what their response is, maybe if I can get some outside funding or support that’d help.
- Personally, I think the single pixel camera is possible, could be low cost and transformative for many people already in GOSH. The following is mostly a rephrasing of what Richard taught me yesterday plus some (maybe) clever ideas to lower the cost