4 ms·

There's quite the chorus of "Everyone is moving to airframes.io" or other similarly phrased posts in the ADS-B Exchange buyout thread[1]. Based on what I'm seei

There's quite the chorus of "Everyone is moving to airframes.io" or other similarly phrased posts in the ADS-B Exchange buyout thread[1]. Based on what I'm seeing, I feel like this is probably a premature place to settle on. 1. If you go to https://github.com/airframesio/ https://github.com/airframesio/ , it talks about how they want to eventually open source stuff. But mercy, like 90% of their stuff is private right now. 2. The Airframes roadmap has basically not been touched since 2019. https://trello.com/b/1vq5pHNq/acars-server-public-board https://trello.com/b/1vq5pHNq/acars-server-public-board (linked to as the "Airframes Public Trello Board" on airframes.io about page, both owned/authored by Kevin). 3. I also have found no public API documentation for reading or querying data. There appears to be a webapp one could pretty easily reverse engineer, but I haven't found any ToS or usage documentation that grant any rights to use the data. I would absolutely not go to airframes.io in it's current state. Side note, I think this is a super fascinating use-case for what p2p might want to target. Trying to draw together & perhaps correlate & index many highly active streams of data is a very compelling (and hard) problem. [1] https://www.jetnet.com/news/jetnet-acquires-ads-b-exchange.html https://www.jetnet.com/news/jetnet-acquires-ads-b-exchange.h... https://news.ycombinator.com/item?id=34520355 https://news.ycombinator.com/item?id=34520355

dx034 4y ago
Also, why not support Opensky which has an academic background and is already well established?
- rektide 4y ago
  I want very much to believe in both projects. Both seem well intended. Alas, for many similar reasons I'm unconvinced by Opensky as well. There are some projects up[1] but there doesn't appear to be any public development for the vast majority of the work. There are decent API docs, and ToS, which is a positive. However the rate limits are pretty low (1000 calls/day[2]) & the data does not appear to have any other distributions available. The internet has massive amounts of invisible labor that go into keeping communities & systems afloat. It's amazing how far it gets us. But it's always unclear, how do these projects scale as needed? What happens as costs grow? Lets look briefly at the example of another big data operator out there, npm, the javascript package manager & website/repository of packages. Npm was owned by Isaac Schlueter, who created in 2009. In 2020, IrisCouch volunteered to provide hosting services for free. They were acquired by Nodejitsu, who in 2013 ran a "Scalenpm" campaign to crowdsource some funding. In 2013 Isaac created npm, inc & began hosting the repo. Microsoft bought npm in 2020, & so far npm has continued to operate a free public registry. Figuring out how to keep this project going across a decade was hard work, with enormous sums spent & tons of work poured in. I know much less, but, like, Java's earlier pretty-big Maven ecosystem seems largely to have existed under the benevolent hand of SonaType (although other repositories have come and gone). There's definitely some goal/dream, that projects like IPFS or Nostro or someone can help marshal & make this data available. There's still a host of other sub-problems to figuring out how something like flight tracking or package managing would actually work atop such a system. Search & discovery, authenticity, trying to effectively index/collate data to make it consumeable are all various extra levels of problem, where again centralized infrastructure plus distributed storage might work, but ideally, we could really distribute the work out. Until then, it feels like we're kind of stuck trusting benevolent for now leaders, who have limited time and hardware resources to share. [1] https://github.com/openskynetwork https://github.com/openskynetwork [2] https://openskynetwork.github.io/opensky-api/rest.html#limitations-for-opensky-users https://openskynetwork.github.io/opensky-api/rest.html#limit...
xrayarx 4y ago
The difference is adsb exchange is for adsb data, this is for all kinds of data except adsb, so the game is who adds more data sources faster
kevinelliott 4y ago
Hey, I'm Kevin (founder of Airframes), and I'd like to give some background here. While the world has been obsessing about ADS-B, which is primarily focused on positional data, I've been building a platform that is heavily focused on ACARS (often referred to as Plain Old ACARS), VDL (the successor to ACARS), HFDL (HF-based, also containing ACARS payloads), and SATCOM (such as AoA - ACARS over Insmarsat, and AoI - ACARS over Iridium). The information in ACARS is quite interesting, and reveals significantly more about flights than ADS-B alone. The goal of the project has always been to provide a fuller picture of what is flying, including OOOI, routing, equipment status, fuel details, and more. Some very early feeders and enthusiasts have been feeding data and supporting the effort, and as many/most HN folks know, this is all very time consuming. Let me respond to your points too: 1. I'm a strong supporter of open source, and conceptually even open data. Your assessment that most of the stuff is still private is true. This is a strategic decision I have made. Having watched the aggregator space for many years now, my concern was that since we're still getting things built out (and I'm the only developer), a heavily budgeted competitor (such as FlightAware) could come along and enter the ACARS space long before we were ready and make the effort pointless. I do plan to continue to open source and open data, and as it makes sense to, at my own expense, I will do so. Especially as more people become involved and build interest in ACARS. 2. The roadmap on the trello has not been updated, because, there were very few people interested in the effort and did not really engage in using it. But there is most definitely a roadmap -- some of the things are related to API, desktop/mobile app, a radio-focused OS (initially aimed at making it easy to feed to Airframes and other aggregators, but with the goal of wider radio interests) similar to OctoPi for OctoPrint. 3. Documentation has been in the works. But, seeing as that I have a lot of priorities (and it's mostly just me at the moment), it's not complete. I was hoping to release it when more digestible/ready. https://docs.airframes.io https://docs.airframes.io 4. I would hope that folks do consider supporting the effort, and there are a lot of big plans for it in the ACARS space. Also, to put it bluntly, we were not yet ready for the events that unfolded. So there are some growing pains to deal with there.

dx034 4y ago

Also, why not support Opensky which has an academic background and is already well established?

rektide 4y ago
I want very much to believe in both projects. Both seem well intended. Alas, for many similar reasons I'm unconvinced by Opensky as well. There are some projects up[1] but there doesn't appear to be any public development for the vast majority of the work. There are decent API docs, and ToS, which is a positive. However the rate limits are pretty low (1000 calls/day[2]) & the data does not appear to have any other distributions available. The internet has massive amounts of invisible labor that go into keeping communities & systems afloat. It's amazing how far it gets us. But it's always unclear, how do these projects scale as needed? What happens as costs grow? Lets look briefly at the example of another big data operator out there, npm, the javascript package manager & website/repository of packages. Npm was owned by Isaac Schlueter, who created in 2009. In 2020, IrisCouch volunteered to provide hosting services for free. They were acquired by Nodejitsu, who in 2013 ran a "Scalenpm" campaign to crowdsource some funding. In 2013 Isaac created npm, inc & began hosting the repo. Microsoft bought npm in 2020, & so far npm has continued to operate a free public registry. Figuring out how to keep this project going across a decade was hard work, with enormous sums spent & tons of work poured in. I know much less, but, like, Java's earlier pretty-big Maven ecosystem seems largely to have existed under the benevolent hand of SonaType (although other repositories have come and gone). There's definitely some goal/dream, that projects like IPFS or Nostro or someone can help marshal & make this data available. There's still a host of other sub-problems to figuring out how something like flight tracking or package managing would actually work atop such a system. Search & discovery, authenticity, trying to effectively index/collate data to make it consumeable are all various extra levels of problem, where again centralized infrastructure plus distributed storage might work, but ideally, we could really distribute the work out. Until then, it feels like we're kind of stuck trusting benevolent for now leaders, who have limited time and hardware resources to share. [1] https://github.com/openskynetwork https://github.com/openskynetwork [2] https://openskynetwork.github.io/opensky-api/rest.html#limitations-for-opensky-users https://openskynetwork.github.io/opensky-api/rest.html#limit...

xrayarx 4y ago

The difference is adsb exchange is for adsb data, this is for all kinds of data except adsb, so the game is who adds more data sources faster

kevinelliott 4y ago

Hey, I'm Kevin (founder of Airframes), and I'd like to give some background here. While the world has been obsessing about ADS-B, which is primarily focused on positional data, I've been building a platform that is heavily focused on ACARS (often referred to as Plain Old ACARS), VDL (the successor to ACARS), HFDL (HF-based, also containing ACARS payloads), and SATCOM (such as AoA - ACARS over Insmarsat, and AoI - ACARS over Iridium). The information in ACARS is quite interesting, and reveals significantly more about flights than ADS-B alone. The goal of the project has always been to provide a fuller picture of what is flying, including OOOI, routing, equipment status, fuel details, and more. Some very early feeders and enthusiasts have been feeding data and supporting the effort, and as many/most HN folks know, this is all very time consuming. Let me respond to your points too: 1. I'm a strong supporter of open source, and conceptually even open data. Your assessment that most of the stuff is still private is true. This is a strategic decision I have made. Having watched the aggregator space for many years now, my concern was that since we're still getting things built out (and I'm the only developer), a heavily budgeted competitor (such as FlightAware) could come along and enter the ACARS space long before we were ready and make the effort pointless. I do plan to continue to open source and open data, and as it makes sense to, at my own expense, I will do so. Especially as more people become involved and build interest in ACARS. 2. The roadmap on the trello has not been updated, because, there were very few people interested in the effort and did not really engage in using it. But there is most definitely a roadmap -- some of the things are related to API, desktop/mobile app, a radio-focused OS (initially aimed at making it easy to feed to Airframes and other aggregators, but with the goal of wider radio interests) similar to OctoPi for OctoPrint. 3. Documentation has been in the works. But, seeing as that I have a lot of priorities (and it's mostly just me at the moment), it's not complete. I was hoping to release it when more digestible/ready. https://docs.airframes.io https://docs.airframes.io 4. I would hope that folks do consider supporting the effort, and there are a lot of big plans for it in the ACARS space. Also, to put it bluntly, we were not yet ready for the events that unfolded. So there are some growing pains to deal with there.