4 ms·

Have they started sending around apps to beta testers? You can sign up[0] and I'm pretty sure I did, but I haven't heard anything. [0] http://crayfis.io/ http:

by privong 10y ago

Have they started sending around apps to beta testers? You can sign up[0] and I'm pretty sure I did, but I haven't heard anything. [0] http://crayfis.io/ http://crayfis.io/

marvel_boy 10y ago
Is the app open sourced? any link?
- privong 10y ago
  > Is the app open sourced? any link? I'm guessing not, since they currently seem to be only letting authorized beta testers use it. For what it's worth, there haven't been any blog posts since Feb 2015[0], so I'm not sure how active the project is? [0] http://blog.crayfis.io/ http://blog.crayfis.io/
  - cshimmin 10y ago
    We're here! Unfortunately I'm terrible at social media and especially blogging, so from the outside we must seem pretty dead. See my response to GP, but basically we've pumped the brakes a bit after a small initial beta roll-out in order to ensure we are accurately modeling the detector response to muons. This is essential to ensure we're not throwing the baby out with the bathwater when deciding which data to keep (i.e. upload to the server). Since we scan HD video at ~30fps to look for the cosmic ray signatures, we have to very aggressively throw away uninteresting data in order to keep the data rates reasonable. If anyone here is really interested in giving the beta test a go (android only for now), feel free to email me directly (look elsewhere in the thread for my address). It would be good to hear some feedback from the pro hackers here on hn :)
- cshimmin 10y ago
  Unfortunately not open sourced (yet). The main reason is because we're physicists, not web security experts (although I do have some background in this). I was afraid that if we open source the code and make the protocol way too obvious, the bar will be just too low for some troll to spam our data acquisition servers, making it hard for us to do science. So for now we're kind of relying on good-old "security through obscurity". In the future we hope to harden the online system further, and eventually open source the code for both the android/iOS apps, as well as the online website which functions as a high-throughput distributed DAQ with online analytics.
  - privong 10y ago
    Have you thought about trying to partner with someone to do a Google Summer of Code (or something similar) for this? It seems specific and targeted enough that it might have a chance of getting support.
    - cshimmin 10y ago
      That's a really cool idea, I don't think we've even considered this before. If you have information about how to get this kind of thing going, please get in touch! (My email is my hn username at cern.ch)
      - privong 10y ago
        I've not been involved with SoC myself, but check out the website: https://developers.google.com/open-source/gsoc/ https://developers.google.com/open-source/gsoc/ Unfortunately the deadline for this summer has already passed.
cshimmin 10y ago
We've invited a small number of beta testers for the android platform, mostly to test that our website (which functions as our DAQ) is working and to get a baseline dataset to characterize the kind of challenges we'll have in processing the data. AWS gave us a rather generous grant of credit for their services, and we spent a good amount of time developing a system for DAQ/live-metrics that should scale to at least a few hundred thousand users. The devices have to upload their data every ~3min, so it's a fairly significant amount of traffic. Beyond that we'd have to put either a lot more time into it or find "real" scalability experts, which means we'd need to get some $$$ first ;) The main thing that has caused our beta rollout to stall is the fact that we have not yet been able to successfully calibrate our response to muons, which are an important component of the UHECR shower. Whereas we were able to calibrate the response to x-ray/gamma-ray photons fairly accurately using common radioactive sources, there is no easy way to generate muons under controlled circumstances (you basically need a particle accelerator). We are applying for beam time at Fermilab, so hopefully we will be able to do these measurements soon. In parallel, we are also trying the "bottom up" approach, by developing detailed simulations to model how energetic particles interact with CMOS pixels. For this we're using GEANT, which does a great job of simulating how particle deposit energy into materials, but the hard part is modeling how that turns into actual charges that get read out by the CMOS electronics. We're getting close, but with nobody working full-time on this project and no funding, most development has happened in spurts and bursts whenever someone happens to have time away from their "day job". In order to look for cosmic rays we have to scan HD video frames at ~30fps, which is a ton of data. So that means we have to very aggressively throw away "uninteresting" data before selecting candidates to upload to our servers (in physics jargon this is called triggering). Since the muons are not currently understood, we are afraid that we cannot design an effective trigger. Which means that if we continued the beta roll out, we might blow through all of our AWS cash and then realize retrospectively that all the data we have taken is useless. Hence we made the decision to pump the brakes on the beta testing until we can resolve the instrumental questions. P.S.: If anyone really wants to give the app a try (unfortunately android only at the moment), feel free to shoot me an email directly. My email is my hn username at cern.ch.

by privong 10y ago

Have they started sending around apps to beta testers? You can sign up[0] and I'm pretty sure I did, but I haven't heard anything. [0] http://crayfis.io/ http://crayfis.io/

marvel_boy 10y ago

Is the app open sourced? any link?

privong 10y ago
> Is the app open sourced? any link? I'm guessing not, since they currently seem to be only letting authorized beta testers use it. For what it's worth, there haven't been any blog posts since Feb 2015[0], so I'm not sure how active the project is? [0] http://blog.crayfis.io/ http://blog.crayfis.io/
- cshimmin 10y ago
  We're here! Unfortunately I'm terrible at social media and especially blogging, so from the outside we must seem pretty dead. See my response to GP, but basically we've pumped the brakes a bit after a small initial beta roll-out in order to ensure we are accurately modeling the detector response to muons. This is essential to ensure we're not throwing the baby out with the bathwater when deciding which data to keep (i.e. upload to the server). Since we scan HD video at ~30fps to look for the cosmic ray signatures, we have to very aggressively throw away uninteresting data in order to keep the data rates reasonable. If anyone here is really interested in giving the beta test a go (android only for now), feel free to email me directly (look elsewhere in the thread for my address). It would be good to hear some feedback from the pro hackers here on hn :)
cshimmin 10y ago
Unfortunately not open sourced (yet). The main reason is because we're physicists, not web security experts (although I do have some background in this). I was afraid that if we open source the code and make the protocol way too obvious, the bar will be just too low for some troll to spam our data acquisition servers, making it hard for us to do science. So for now we're kind of relying on good-old "security through obscurity". In the future we hope to harden the online system further, and eventually open source the code for both the android/iOS apps, as well as the online website which functions as a high-throughput distributed DAQ with online analytics.
- privong 10y ago
  Have you thought about trying to partner with someone to do a Google Summer of Code (or something similar) for this? It seems specific and targeted enough that it might have a chance of getting support.
  - cshimmin 10y ago
    That's a really cool idea, I don't think we've even considered this before. If you have information about how to get this kind of thing going, please get in touch! (My email is my hn username at cern.ch)
    - privong 10y ago
      I've not been involved with SoC myself, but check out the website: https://developers.google.com/open-source/gsoc/ https://developers.google.com/open-source/gsoc/ Unfortunately the deadline for this summer has already passed.

cshimmin 10y ago

We've invited a small number of beta testers for the android platform, mostly to test that our website (which functions as our DAQ) is working and to get a baseline dataset to characterize the kind of challenges we'll have in processing the data. AWS gave us a rather generous grant of credit for their services, and we spent a good amount of time developing a system for DAQ/live-metrics that should scale to at least a few hundred thousand users. The devices have to upload their data every ~3min, so it's a fairly significant amount of traffic. Beyond that we'd have to put either a lot more time into it or find "real" scalability experts, which means we'd need to get some $$$ first ;) The main thing that has caused our beta rollout to stall is the fact that we have not yet been able to successfully calibrate our response to muons, which are an important component of the UHECR shower. Whereas we were able to calibrate the response to x-ray/gamma-ray photons fairly accurately using common radioactive sources, there is no easy way to generate muons under controlled circumstances (you basically need a particle accelerator). We are applying for beam time at Fermilab, so hopefully we will be able to do these measurements soon. In parallel, we are also trying the "bottom up" approach, by developing detailed simulations to model how energetic particles interact with CMOS pixels. For this we're using GEANT, which does a great job of simulating how particle deposit energy into materials, but the hard part is modeling how that turns into actual charges that get read out by the CMOS electronics. We're getting close, but with nobody working full-time on this project and no funding, most development has happened in spurts and bursts whenever someone happens to have time away from their "day job". In order to look for cosmic rays we have to scan HD video frames at ~30fps, which is a ton of data. So that means we have to very aggressively throw away "uninteresting" data before selecting candidates to upload to our servers (in physics jargon this is called triggering). Since the muons are not currently understood, we are afraid that we cannot design an effective trigger. Which means that if we continued the beta roll out, we might blow through all of our AWS cash and then realize retrospectively that all the data we have taken is useless. Hence we made the decision to pump the brakes on the beta testing until we can resolve the instrumental questions. P.S.: If anyone really wants to give the app a try (unfortunately android only at the moment), feel free to shoot me an email directly. My email is my hn username at cern.ch.