Re: [Talk-ca] GitHub app for BC2020i Challenge
On Sep 6, 2018, at 4:30 PM, john whelan wrote: > The pilot project itself did manage to get a fair amount of accurate data > into OSM. That data is still there and can be used. It was instrumental in > supporting the HOT summit in Ottawa. It managed to raise awareness within > local government both in Canada and in Africa about how OSM could be useful > and it clarified a number of legal issues about importing data. Then, quite a number of events caused it to go off the rails. OK, that happens; no judgement on my part, I'm far, far happier to try to watch the smoke clear and "might I offer to help?" > > And I'm not boasting, but I did put some effort into the richest set of > > potential tags (harvested from our wikis) than I believe anybody else did, > > and I don't really have any specific interest in the project, except that > > it be a WELL RUN project. (So I tried very hard to "seed it well"). > > You mean me getting the recumbent trike out and site inspecting a few hundred > buildings and adding tags to them was for nought? How sad. John, I'm on your side, the side of OSM having great data while it builds communities across Canada, universities, high schools, libraries, etc. I'm happy the project has "stubbed in" (it's certainly "not nothing!") at least some data. Dust off your hands and keep going! (Is hopefully the attitude I'm encouraging you to adopt). > There is still a gentle movement to gather more data over time, whether we > are keeping the current building data up to date is a separate issue. Stats > is still trying to find ways to make building outlines available under their > Open Data license. A 100% separate track for which I wish them the very best of luck. That may or may not be successful, and CAN and SHOULD happen on a parallel track with OSM strategies which work. I'll stand shoulder-to-shoulder with THAT spirit, as it is what makes OSM a very powerful OSM. > The approach used on the pilot to import building outlines manually has been > picked up by Microsoft who have been making them available for the USA and I > understand Stats were involved with discussions with Microsoft about some > technical aspects. Truly, I'm glad to hear it. OSM and Microsoft (and now, it turns out, Stats Canada) do get symbiotic every once in a while, and we're all the better for it as/when it happens. > The Ottawa pilot was a perfect storm in many ways with many different players > involved coming together. Reproducing it is harder than it first seems. I get that, so do others. A "more valuable" aspect to pick up as a shiny coin from that is the learning experience that it is. In QA this is called a "post mortem" and is one of the most valuable data chunks for "the next phase" (and there always is a next phase). > What I would like to see is someone pick up doing analysis with R r.org to > see if we can build the feedback loops that might help motivate getting the > tags populated. Heh, you could sketch up a wiki to get them started...! :-) (It's not a bad idea, really). Steve ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] GitHub app for BC2020i Challenge
> Threads here, consensus and the wiki declared it dead, as it failed on a number of fronts, primarily that it didn't respect OSM in certain ways in which OSM must be respected. The pilot project itself did manage to get a fair amount of accurate data into OSM. That data is still there and can be used. It was instrumental in supporting the HOT summit in Ottawa. It managed to raise awareness within local government both in Canada and in Africa about how OSM could be useful and it clarified a number of legal issues about importing data. > And I'm not boasting, but I did put some effort into the richest set of potential tags (harvested from our wikis) than I believe anybody else did, and I don't really have any specific interest in the project, except that it be a WELL RUN project. (So I tried very hard to "seed it well"). You mean me getting the recumbent trike out and site inspecting a few hundred buildings and adding tags to them was for nought? How sad. There is still a gentle movement to gather more data over time, whether we are keeping the current building data up to date is a separate issue. Stats is still trying to find ways to make building outlines available under their Open Data license. The feedback I'm referring to was the number of buildings tagged and the total number of tags on the buildings. When I saw the analysis it was quite interesting to see the numbers going up. What I didn't see is those parts of the map that had low numbers of building tags. The approach used on the pilot to import building outlines manually has been picked up by Microsoft who have been making them available for the USA and I understand Stats were involved with discussions with Microsoft about some technical aspects. The Ottawa pilot was a perfect storm in many ways with many different players involved coming together. Reproducing it is harder than it first seems. What I would like to see is someone pick up doing analysis with R r.org to see if we can build the feedback loops that might help motivate getting the tags populated. Cheerio John On 6 September 2018 at 16:39, OSM Volunteer stevea < stevea...@softworkers.com> wrote: > On Sep 6, 2018, at 1:14 PM, john whelan wrote > (replying to me, stevea): > > > Hm, we tried to revive the wiki, a tried-and-true OSM methodology for > doing EXACTLY that. Is there something wrong with that idea? > > > > No this project was initiated by Stats Canada, but without clear > requirements or feedback about what had been achieved. The Stats Can side > wasn't dependant on normal OSM mappers but my understanding was it was > hoping to draw in new mappers. > > John, BC2020i (note the i) was initiated by Stats Canada. Threads here, > consensus and the wiki declared it dead, as it failed on a number of > fronts, primarily that it didn't respect OSM in certain ways in which OSM > must be respected. It MIGHT have become resurrected as BC2020 (no i) and > the wiki were attempts to do that, especially as Stats Canada was out of > the picture by then, as they weren't going to contribute to either the wiki > or the BC2020 itself, though, like anybody who "takes" (uses, and not in a > bad way) OSM data, StatsCanada would be welcome to the results of BC2020 > (and BC2020i is dead, I'll say it one more time). We stripped away what > was wrong with "i" and slightly renamed the project to conform to the way > that OSM has, can, does and will complete projects (including, but > requiring that we use wikis). BC2020 seems to have become moribund and > ineffective, though I continue to hold out high hopes that it can be > successful. > > OSM actually DOES have clear requirements or feedback, part of those are > naturally "built in" to crowdsourced projects, part of those need a bit of > goosing along by prompting volunteers to communicate well (wiki is ONE way, > not the only way) via simple things like reporting progress and/or > continuing to sharpen up focus because tagging started out as an early > draft, but now is a "more complete" or "final" draft. Sure, any good > project wants to start out with clear goals (and should) but a modest bit > of mid-course correction certainly won't prevent successful completion. > > > Fine but a couple of maperthons that were organised had data quality > issues and no clear guidance about what tags were most valuable. > > That's because no QA was planned up front, just like I suggested to do > last year and into January of this year. And I'm not boasting, but I did > put some effort into the richest set of potential tags (harvested from our > wikis) than I believe anybody else did, and I don't really have any > specific interest in the project, except that it be a WELL RUN project. > (So I tried very hard to "seed it well"). In crowdsourcing, yes, this can > be challenging, but communication is the lynchpin that allows it. Wikis, > at least in OSM can be and often are a critical component of the successful > ongoing (status reporting, et
Re: [Talk-ca] GitHub app for BC2020i Challenge
On Sep 6, 2018, at 1:14 PM, john whelan wrote (replying to me, stevea): > > Hm, we tried to revive the wiki, a tried-and-true OSM methodology for doing > > EXACTLY that. Is there something wrong with that idea? > > No this project was initiated by Stats Canada, but without clear requirements > or feedback about what had been achieved. The Stats Can side wasn't > dependant on normal OSM mappers but my understanding was it was hoping to > draw in new mappers. John, BC2020i (note the i) was initiated by Stats Canada. Threads here, consensus and the wiki declared it dead, as it failed on a number of fronts, primarily that it didn't respect OSM in certain ways in which OSM must be respected. It MIGHT have become resurrected as BC2020 (no i) and the wiki were attempts to do that, especially as Stats Canada was out of the picture by then, as they weren't going to contribute to either the wiki or the BC2020 itself, though, like anybody who "takes" (uses, and not in a bad way) OSM data, StatsCanada would be welcome to the results of BC2020 (and BC2020i is dead, I'll say it one more time). We stripped away what was wrong with "i" and slightly renamed the project to conform to the way that OSM has, can, does and will complete projects (including, but requiring that we use wikis). BC2020 seems to have become moribund and ineffective, though I continue to hold out high hopes that it can be successful. OSM actually DOES have clear requirements or feedback, part of those are naturally "built in" to crowdsourced projects, part of those need a bit of goosing along by prompting volunteers to communicate well (wiki is ONE way, not the only way) via simple things like reporting progress and/or continuing to sharpen up focus because tagging started out as an early draft, but now is a "more complete" or "final" draft. Sure, any good project wants to start out with clear goals (and should) but a modest bit of mid-course correction certainly won't prevent successful completion. > Fine but a couple of maperthons that were organised had data quality issues > and no clear guidance about what tags were most valuable. That's because no QA was planned up front, just like I suggested to do last year and into January of this year. And I'm not boasting, but I did put some effort into the richest set of potential tags (harvested from our wikis) than I believe anybody else did, and I don't really have any specific interest in the project, except that it be a WELL RUN project. (So I tried very hard to "seed it well"). In crowdsourcing, yes, this can be challenging, but communication is the lynchpin that allows it. Wikis, at least in OSM can be and often are a critical component of the successful ongoing (status reporting, etc.) and completion of projects. > I could be wrong but I'm not aware of any significant movement on the project. OSM (worldwide, not "just" in Canada or any particular country) seems to be in a communication crisis, where everybody thinks that some sort of "secret-special-sauce walkie-talkie" (like GitHub or Slack) will solve everything. No. While those have their place (let me emphasize, I truly mean that) they will continue to Balkanize (fracture) and legally bind (have you READ the contracts GitHub and Slack ask you to agree to?!) OSM volunteers far past the state of hobble-and-wobble, it will kill us. Talking about GitHub is like pilot-radio-chatter: specialized, harmful to BUILDING new community (which is CRITICAL in BC2020) and will keep you grounded as certain as a hurricane. OSM Canada knows how to crawl, and even walk. To run, and even fly, especially with BC2020, use what we have. The fancy stuff might (MIGHT!) be used later. SteveA ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] GitHub app for BC2020i Challenge
> Hm, we tried to revive the wiki, a tried-and-true OSM methodology for doing EXACTLY that. Is there something wrong with that idea? No this project was initiated by Stats Canada, but without clear requirements or feedback about what had been achieved. The Stats Can side wasn't dependant on normal OSM mappers but my understanding was it was hoping to draw in new mappers. Fine but a couple of maperthons that were organised had data quality issues and no clear guidance about what tags were most valuable. I could be wrong but I'm not aware of any significant movement on the project. Cheerio John On 6 September 2018 at 15:58, OSM Volunteer stevea < stevea...@softworkers.com> wrote: > > Personally I think if the BC2020i is to be revived mappers really need > some feedback on what has been done and what tags are of interest. > > Hm, we tried to revive the wiki, a tried-and-true OSM methodology for > doing EXACTLY that. Is there something wrong with that idea? > > I've been trying to keep "the embers orange and warm" on this project (via > its wiki) since January. > > https://wiki.osm.org/wiki/WikiProject_Canada/Building_Canada_2020 > > If there is something wrong with that wiki (John Whelan, you have > described its history both here in talk-ca and to me via private email any > number of times), then re-write it. I think it is at least a start to > describe what you (Canada) are trying to do, so if it or parts of it are > useful, use it. > > On an upside, there are a lot of data there, (though some of it might be > junk or outdated), and so as a corollary, on a minor downside, it could be > said to be loquacious/wordy/overly detailed. On a MAJOR downside, "BC2020" > (not suffixed with an "i" as that is rightly declared to be dead) "needs > reviving." OK, revive it. If not via wiki, "because mappers really need > some feedback" (it DOES mention Active Monitoring tools) and "what tags are > of interest" (it DOES mention Tag Standardization" and "The data that could > be mapped"), then HOW? The answer: (or at least an excellent one): use > our already-built, well-established, good-for-our-community tools which > WORK. > > Go! > > SteveA > OSM Volunteer since 2009 ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] GitHub app for BC2020i Challenge
> Personally I think if the BC2020i is to be revived mappers really need some > feedback on what has been done and what tags are of interest. Hm, we tried to revive the wiki, a tried-and-true OSM methodology for doing EXACTLY that. Is there something wrong with that idea? I've been trying to keep "the embers orange and warm" on this project (via its wiki) since January. https://wiki.osm.org/wiki/WikiProject_Canada/Building_Canada_2020 If there is something wrong with that wiki (John Whelan, you have described its history both here in talk-ca and to me via private email any number of times), then re-write it. I think it is at least a start to describe what you (Canada) are trying to do, so if it or parts of it are useful, use it. On an upside, there are a lot of data there, (though some of it might be junk or outdated), and so as a corollary, on a minor downside, it could be said to be loquacious/wordy/overly detailed. On a MAJOR downside, "BC2020" (not suffixed with an "i" as that is rightly declared to be dead) "needs reviving." OK, revive it. If not via wiki, "because mappers really need some feedback" (it DOES mention Active Monitoring tools) and "what tags are of interest" (it DOES mention Tag Standardization" and "The data that could be mapped"), then HOW? The answer: (or at least an excellent one): use our already-built, well-established, good-for-our-community tools which WORK. Go! SteveA OSM Volunteer since 2009 ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] GitHub app for BC2020i Challenge
It looks pretty but misses the detail in the magic step then a miracle occurs. Currently wind is the cheapest renewable energy source. Off shore works well but the central area of North America works fine. Unfortunately New York, Toronto, Montreal were all established in relatively wind free areas many years ago so they need a power grid to get the energy to them. Wind also has the advantage of being available when all those office lights in the buildings were on in the video. There are a couple of problems with solar, first the cost of connecting to the grid is quite high. More than 50% of a home roof top installation costs are for permits and connections. That's why OREC.ca likes bigger roofs and ground solar arrays. The second problem is roof vents, they limit where solar panels can be placed on the roof. Currently there are very few roof vents in OSM to see where solar panels could be placed. Very few buildings have enough roof detail to see if the slope is in the right direction for solar panels. OREC accept that wind turbines are nice but the most efficient ones today are big and that means expensive $3-4 million dollars each which they think they can't affordon a local coop basis. A wind farm makes sense from a logistics point of view but a decent wind farm has many turbines, we aren't talking local here. The bigger they are the more efficient they are and the more likely to produce a steady stream of electricity. Most are around 499 feet tall, the FAA require extra steps in their approval process if its higher than 499 feet. New ones in the pipeline are 850 feet tall. On the mapping side I haven't seen any feedback on how much detail has been attached to buildings as part of the BC2020i project. If anywhere I'd expect Ottawa to have the richest tag set per building. Personally I think if the BC2020i is to be revived mappers really need some feedback on what has been done and what tags are of interest. Cheerio John On 6 September 2018 at 11:06, Jonathan Brown wrote: > This is cool. Could we not develop a BC2020i challenge similar to the ECCE > annual challenge? (See McMaster University Team’s winning ECCE 2018 entry: > https://www.youtube.com/watch?v=yRb_g1llT00&feature=youtu.be&list= > PLdgq5G0ox73VEQFJd4No6peb4NP-GFbdU > > > > Jonathan > > > > ___ > Talk-ca mailing list > Talk-ca@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-ca > > ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca
Re: [Talk-ca] GitHub app for BC2020i Challenge
This is cool. Could we not develop a BC2020i challenge similar to the ECCE annual challenge? (See McMaster University Team’s winning ECCE 2018 entry: https://www.youtube.com/watch?v=yRb_g1llT00&feature=youtu.be&list=PLdgq5G0ox73VEQFJd4No6peb4NP-GFbdU Jonathan ___ Talk-ca mailing list Talk-ca@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-ca