Loomio
Mon 14 Apr 2014

Openstreetmap.in do we really want to make this happen ?

DU
[deactivated account] Public Seen by 16

please refer to the discussion so far with respect to the openstreetmap.in domain here https://lists.openstreetmap.org/pipermail/talk-in/2013-June/001829.html , i wanted to move this item either way from my list of items to do hence this discussion.

DU

[deactivated account] started a proposal Mon 14 Apr 2014

OpenStreetMap.in Community made maps, for Indians. Closed Thu 1 May 2014

This was something which was proposed by Arun Ganesh way back in 2013 , this was not made public and was just cirulated among few people .

https://docs.google.com/document/d/1nXz_oWkb29GBx-wGuMlD5yWUs_p2VRb8invFTVt8fu4/edit?usp=sharing

Agree - 8
Abstain - 0
Disagree - 1
Block - 1
10 people have voted (71%)
HSR

Hardeep Singh Rai
Agree
Mon 14 Apr 2014

It is rather late. It should have happen a lot time earlier.

AG

Arun Ganesh
Agree
Mon 14 Apr 2014

About time, especially since OSM is gaining traction in India.

IC

Ishan Chattopadhyaya
Block
Mon 14 Apr 2014

I agree on principle, but don’t think Loomio is the correct forum for such discussions / voting.

DU

[deactivated account]
Abstain
Tue 15 Apr 2014

I am ok either way since its a collective decision , would like to see all those voices heard and individual commitments and technical issues we can face while on such a project.

S

sumandro
Disagree
Tue 15 Apr 2014

Can we change "for Indians" to "for India"? That is: "Community Made Maps, for India".

Also, agree with Ishan that perhaps we should do this over the OSM India mailing list. Lazy consensus should be appropriate for taking decisions.

SB

Sanjay Bhangar
Block
Tue 15 Apr 2014

until there is some technical consensus on how to deal with the issue of "forking" main OSM data and where new updates will go and how they will be synced with main OSM, I think it maybe dangerous to go ahead with an instance.

DT

Devdatta Tengshe
Agree
Wed 16 Apr 2014

I generally agree that this is a good idea, and we should do this.
I foresee some technical difficulties, but they can be solved;

SB

Sanjay Bhangar
Agree
Mon 21 Apr 2014

I believe the forking issue I had raised is satisfactorily resolved, so I remove my block on this.

DU

[deactivated account]
Agree
Tue 22 Apr 2014

Looking at the progress and people involvement so far on the mailing list and here this looks like a feasible project.

SA

Sajjad Anwar
Agree
Wed 30 Apr 2014

I agree that we need a separate tile server. But want to resonate what Ishan and Sumandro mentioned earlier. We should just use the mailing list for this and not a separate channel. Leveraging the mailing list will get more attention and support.

DU

[deactivated account] Mon 14 Apr 2014

Before you vote and start talking about it read the proposal document carefully , also the document is open for editing if you think the proposal is not ok you can either make changes to it or , you can make your own and bring it here for discussion .

DU

[deactivated account] Mon 14 Apr 2014

@ishanchattopadhyay we are not here to discuss about the platform , its ok to agree and disagree but state which parts you do not agree and disagree . Better if you can write a proposal and present your case here.

DU

[deactivated account] Mon 14 Apr 2014

You’ll be prompted to make a short statement about the reason for your decision. This makes it easy to see a summary of what everyone thinks and why. You can change your mind and edit your decision freely until the proposal closes.

DU

[deactivated account] Tue 15 Apr 2014

one comment i heard a lot was why ? and how are you we going to sync our data from upstream ? There was no straight answer to this one . Also one feeling i got was like this is too much of repetitive work .

IC

Ishan Chattopadhyaya Tue 15 Apr 2014

we are not here to discuss about the platform

I wasn't here to do that either. But when I was here, I realized that you run the show here and require every signup to be moderated by you. This is against regular culture of open community projects like OpenStreetMap. By requiring community members to join closed groups like Loomio, you are encouraging discussions to happen away from the community, which should be avoided at all costs. I suggest you use the OpenStreetMap wiki to document your proposal and use the mailing list for discussions.

SB

Sanjay Bhangar Tue 15 Apr 2014

Does anyone have an answer to the problem of data-sync? What happens if we set up osm.in? Can we easily continually update data from main OSM and re-sync our data back? Or do we, at that point, create a hard fork of the data and go from there?

I think its a bit dangerous to setup an instance until that question is answered.

SB

Sanjay Bhangar Tue 15 Apr 2014

(also, while I find loomio to be an interesting decision-making platform, I agree that this is probably better done on the mailing list).

DU

[deactivated account] Tue 15 Apr 2014

@ishanchattopadhyay agreed to your mod point so whomsoever in on this group is made a mod , talking about mailing lists the discussions and talks keeps on going without owning things and getting things done , more over i wanted to check out this platform for these kinds of decisions . if you have noticed my vote and comment i am ok either way .

DU

[deactivated account] Tue 15 Apr 2014

@ishanchattopadhyay let me clarify this proposal is not mine , its now open on mailing list , feel free to propose whatever you think is your take on this topic anywhere. again i am sticking to this platform to brainstorm.

IC

Ishan Chattopadhyaya Tue 15 Apr 2014

@satyakamgoswami Please don't get me wrong. I appreciate your enthusiasm and support this initiative. That's why I mentioned, in principle I agree with the proposition. However, this is not an appropriate forum for any discussions on the issue. Previously, discussions on the topic have happened at talk-in@ and I don't see any reason why discussions there can result in "talks going on without owning things" and a different outcome can happen if such discussions happen here.

More discussions should happen on @sanjaybhangar's concerns on updates. There are many other concerns that need to be addressed, but none of these appear as blockers to me.

IC

Ishan Chattopadhyaya Tue 15 Apr 2014

let me clarify this proposal is not mine
I don't really care. The proposal is good.

its now open on mailing list , feel free to propose
whatever you think is your take on this topic
anywhere. again i am sticking to this platform to
brainstorm.

This "my way or the highway" attitude can be really detrimental to the cause for which all of us have so far remained deeply united and passionate about.

DU

[deactivated account] Tue 15 Apr 2014

@ishanchattopadhyay its not about anyone's ego's here yes i wanted to shake the hornets nest and come quickly to a decision , i own the domain name anyone interested in owning it are free to raise there hand and do whatever they want with it .
Coming back to the issues which @sanjaybhangar mentioned how do you think its an non issue ? Can you list the concerns which are like blockers .

DU

[deactivated account] Tue 15 Apr 2014

@sumandro the doc is open for edit with link so feel free to make any changes. i think lazy consensus is a difficult one on this one since there are many stakeholders, resources and unknowns we are dealing. i do not want this proposal to fall through the gap and five years from now we are again talking same.

SB

Sanjay Bhangar Tue 15 Apr 2014

@satyakamgoswami i'm fine using this space to discuss as long as someone summarizes / copies the discussions over to the mailing list and a final decision is taken only after it has been put on the ML. I do feel that many may not sign-up for a new platform and shouldn't be excluded from the process.

About the updates question, I really know nothing about the OSM tech-stack apart from spending one day setting it up on my laptop for kicks, so people who are more knowledgeable will need to answer to that. It's really more just a common-sense type question and getting the sense that this may not be a trivial to do. I definitely think this question is a blocker though - its totally possible there is some easy solution that I am not aware of, but until this is figured out, we need to think hard about why we want our own instance if it is going to create a hard fork, and if there are ways to accomplish what we need without creating a hard fork of the OSM data.

From looking at Arun's proposal, it seems like some ideas are around better localized styling, labelling, etc. A lot of that stuff seems like it can happen without forking the data / using main OSM to store canonical data and just creating differently styled front-ends. This should happen anyways if Arun / anyone else is willing to contribute styles and designs, I don't see any good reason not to have more styles and styles better suited to local conditions.

I think the big sticky point is: "Conform to local laws pertaining to the display of India’s external boundaries". I think this is a big reason why many institutions and people are looking for an India-specific OSM, so that we can use OSM for government / education projects without getting into trouble with the law. Is there a way to do this without hard forking OSM? Or just show the "legal" boundaries for the india OSM but have all other data remain synced / being pulled from the main OSM? Here is where I have no idea how the stack works / how this maybe accomplished, but I think needs to be addressed.

It maybe useful to split up the use-cases into things that can be accomplished without forking data, and things that would require forking of data, and then see if someone has a tech solution to the problem.

If we are going to hard-fork OSM to create an india-specific version, I really think we need to think long and hard before doing so as it seems quite dangerous.

DU

[deactivated account] Tue 15 Apr 2014

i wanted this space to be like SIG. Absolutely no decision will be taken in isolation , the whole discussion i will document and take back to mailing list or OSM Wiki.

me too have no clue about how are we going to handle the updates ,there are more experienced people on this group who can talk about it .

AG

Arun Ganesh Tue 15 Apr 2014

@sanjaybhangar
Forking the data:
Not required, one would still make edits to the main db and we can just update the tiles using daily dumps or minutely diffs. http://wiki.openstreetmap.org/wiki/Planet.osm/diffs#Minute.2C_Hour.2C_and_Day_Files_Organisation

Borders:
Its fairly trivial to create another postgres table with the borders we want and have mapnik render that instead of the default ones.

AG

Arun Ganesh Tue 15 Apr 2014

Btw, the idea of forking the data is not part of the proposal in any way. This is more about custom tiles and customized location based services like http://www.openstreetmap.de/karte.html

DU

[deactivated account] Tue 15 Apr 2014

Thanks to @ishanchattopadhyay the document has been converted to a wiki page https://wiki.openstreetmap.org/wiki/IN:Proposal feel free to contribute.

DU

[deactivated account] Tue 15 Apr 2014

can everyone give there OSM handles in order to quote on the wiki page.

DT

Devdatta Tengshe Wed 16 Apr 2014

Firstly, my OpenStreetmaps User page is: http://www.openstreetmap.org/user/Devdatta

I see that there are multiple issues that we are discussing here. Firstly do we want an India specific page with differently styled tiles? The answer is a resounding Yes.

The second about whether discussion should happen here or on the mailing list: I have no hard position on this.

The third sticking point about the technical difficulties and whether we should fork the database or not. I say that we should cross the bridge when we come to it. Once we all agree that this should be done, a group of technically proficient people should look into how exactly it can be done. (I agree with Arun's comment, that this can be done without forking the database). it shouldn't be a blocker for the decision to go forward.

Let us first agree that this should be done, and set up a committee of interested persons who want to take responsibility. All the detailed decisions can be taken after due discussion.

DU

[deactivated account] Wed 16 Apr 2014

So technically we will not be forking , but will be using the existing data with styled tiles . What are the Legal angles to it i recall Survey of India going after Google for just conducting a Mapathon , technically on the same basis they can come after the OSM community too .
http://www.medianama.com/2013/04/223-issues-that-survey-of-indias-complaints-against-googles-mapathon-raises-about-indias-maps-policy/

SB

Sanjay Bhangar Mon 21 Apr 2014

So, to setup a Tileserver that just provides custom styles would definitely be nice and I see no issues (just someone needs to do the work).

A big motivation for this is definitely to be able to create "legal" borders for India to be able to use OSM in govt. / educational projects, etc. This does involve "forking" the data and there are issues we need to deal with there. However, after a discussion with Sajjad as well, I am led to believe by people who know more than me, that this is a problem that can be figured out (though may not be trivial).

Anyways, so in principle I think this is a good project, and its great to see the enthusiasm around it - it will require a bunch of work to get up and running and sustained maintainence though, so this should be discussed, and that discussion is probably better had on the ML :-)

DU

[deactivated account] Thu 1 May 2014

@sajjadanwar agree to what everyone says , this topic has been lying dormant for last 1.5 years , i am happy things have moved a bit further.This proposal as planned on loomio will be closing on this site in next 7 hours , thanks to everyone who participated .

I@

i@thejeshgn.com Wed 14 May 2014

There could be legal issues. Just make sure we are legally (allowed by Indian laws) allowed to run map server in India.