As an occasional user, what I really liked in your previous version was that Phind always gave the list of sources.
So my workflow was :
- ask a question
- quickly read the given answer
- go to the main sources to decide if they are trustworthy or not
I noticed that the new version does not always give the sources, even when I ask. Is this intentional or do you plan to revert this in the near future ?
It also solves the problem of not reducing the traffic to the content creators you use.
Anyway, congratulations for making this tool and offering a free plan to test it.
Interesting, do you mind sharing some of the queries you tried? We avoid doing automatic searches for certain followups if the existing sources already contain the correct information, but we should be giving improved sources compared to the original version of Phind.
[0] is a thread reproducing what I described. In that case I would like Phind's last answer to provide me with a link to the documentation or source code of 'NormalizedLandmarkList'.
I just discovered Josh Newans' projects and tutorials about mobile robotics on ROS2.
I find them particularly clear and straightforward, and I think some HN people here might find them interesting as well.
It can be great for beginners to explore the project, but as a robotics engineer myself, I also find some specific topics really useful and well explained. The ROS2 documentation is not always as clear as it could be.
My personal website was originally written this way. The compile and build script was about 20 lines of shell scripting (relying on pandoc), and could have probably been reduced further. It worked very well and I stuck with it for about 2 years before moving on to Hugo. It was a good experience to hand-roll all the components of a static site, and something every web dev should be required to do early in their career.
> it occurs to me that your system only requires an occasional movement rather than a rapid movement for tracking purposes
Right, actually the system wakes up every 10 seconds to check if the angle needs to be adjusted. It often does not and waits for the next 10 seconds.
There is an opportunity to implement true "hibernation" while the system is waiting to save some more power, but it's not done yet.
> Secondly, the motion system appears to be single-ended and based upon a rope and a stepper motor. It would perhaps be useful to consider conversion to a more rigid system.
Actually there are two geared motors, allowing to control both angles.
Yes, I use simple ropes, the system is stable enough thanks to the counterweight (all ropes are always under tension, making the whole system stable).
> In short, if you have knowledge of your own geometry, have known orientation relative that fixed point, and can before activation determine the TOF distance to your target, and have a GPS fix thus solar inclination at that moment is calculable, then it should be possible to track the target without the need for visual feedback.
My project takes the opposite approach: using a low-cost camera board to avoid having to measure all the geometric aspects precisely.
> In terms of temperature sensing, you can obtain remote temperatures using infra-red linear systems which should be effective.
Good idea, I haven't implemented temperature sensing yet, but I'll look into infrared sensors.
> In terms of safety, laser TOF ensuring no change in distance would ensure the subject is still within the programmed range and an intermediate object hasn't been introduced for combustion.
It may be too late, the intermediate object/person is already heating up by the time we detect it.
> In terms of gathering interest, not sure where you are based in France but I could suggest structuring the system as a public sculpture and holding events including eating food cooked by the system which could involve the additional sponsorship of local wineries, cider, etc. to garner social support.
Yes I'm in France, it would be nice to present the system at such events. For the moment I'd much rather get to work on the technical side.
> Smiles from Sydney. I applaud your work toward the application of technology to social and environmental concern, we need more of this.
Taking a computer out of the loop is the best way to guarantee speed and reliability.
Any safety response is going to rely on speed of detection and speed of response. Almost no sensor system will be faster than laser TOF and IMS. They are both very fast and very low power, you can afford to sense at high frequencies and this costs almost no power. Some models may include programmable interrupt lines to further reduce aggregate power utilization by avoiding the need for polling.
In terms of speed of response, anything within 1-2 seconds should be safe. Perhaps having an emergency actuation function in which the array is inverted to prevent convergence (and attract attention as a side effect!) may be safest. You could also cheaply and easily add a siren or audio announcement.
Well short of an enclosure, the standard for power tools is to project some form of laser mark. Only when the operator verifies the mark is in the correct position should they activate the system. This would be a viable approach for a CNC-shaped array of curves and would not require a complete enclosure, which has various downsides (windage, size, weight, potential for damage, suitability for rough terrain, etc.)
I've not heard of it since, no. It's got some steep hills to climb as a concept before it's a better option for building materials than just pouring more concrete.
Not sure what those hills are, but it’s always surprised me that this type of hobby project hasn’t been ever been scaled up industrially.
Setting aside applications for moon base 1 and other sci-fi, there’s a lot of desert on earth. If the construction method is automatic or even semi automatic, and costs almost no energy, then who cares if it’s slow? A legion of robots that can’t even make other robots but can make glass bricks from sand seems like it could be paving the Empty Quarter one decade, finishing the glittering glass towers in the next.
If you're going to start slinging youtubes at people, isn't it common courtesy around here that it should be cool machine-shops or tokamak reactor videos or something kinda educational instead of this awful attempt at comedy/memes? Get out of here with that.
So my workflow was :
- ask a question
- quickly read the given answer
- go to the main sources to decide if they are trustworthy or not
I noticed that the new version does not always give the sources, even when I ask. Is this intentional or do you plan to revert this in the near future ?
It also solves the problem of not reducing the traffic to the content creators you use.
Anyway, congratulations for making this tool and offering a free plan to test it.