OpenStreetMap has had a spike in tiles usage and, as I understand, the decision among those in charge was to limit tiles usage. Ilya Zverev has written a good summary on the decision in his blog: (use Google Translate) http://shtosm.ru/all/dlya-kogo-tayly/ His main point is that OSM is not liable to provide tiles to anyone but the mappers.
OSM owes you nothing, to put it shortly. And it’s not a commercial provider to do it for a reasonable rate. That’s perfectly true. Except that this case is another step to a death spiral and makes the project end come closer.
I saw exactly this happen in an amateur community network that I used. Its approach was exactly the same and it led to an eventual death of the project. OSM resembles it in many details, decisions and explanations. The community network also was very righteous in its statements.
(I have to clarify the state of the internet providers, to make it clear. You may omit this paragraph.) In 1999, the internet providers in the city consisted almost entirely of (audio) dial-up, that is 56Kbit top speed. The phone network was state-owned, it tried hard to jump on the DSL wagon, but only upgraded its stations by 2004. Unlimited Ethernet or DSL appeared only in 2006-7.
So, in 1997, some 5 guys in a big condo near me, tired of modem speed, connected their computers with coaxial cables and Ethernet cards to make a local network. They were very enthusiastic and made all sorts of things: bought 56K modem and an unlimited plan, installed 24/7 servers, made an internal website with lots of information, and so on. In 2000, they organized consumer cooperative and finally bought in bulk a 1Mbit Ethernet line. By 2001 there were about a 100 users, and I was in 110s when I joined it in October or November.
The service was just fantastic compared to what was available: $4..5 per month, $0.06 per megabyte of data. Overall, I paid around $20..25 per month. With nobody wanting to pay a lot, the 1 Megabit connection to the outer world was always available. Inside the network, everything was free, we had 10..100 Megabits connections and shared lots of files.
There was another similar network in neighbor area, which eventually occupied the same area as ours, and in 2003 that network became a commercial provider.
That was the tipping point. From that moment that commercial network tried harder and harder to implement all the newest things, and our non commercial network invested in cables and hardware but had no means to improve the services and organization. In 2007 I finally left it because the commercial guys offered unlimited data plans, payment via bank card or payment terminals in shops, guaranteed support, etc.
Meanwhile our noncommercial network, had same old tools from 2002. For instance, to buy data traffic, we had to print a bill, go pay it in a bank, then input the numbers in an old form in the internal website, which looked the same as in 1997. Then the admin would just add data bytes to your account in the gateway. To open the gateway (it was closed to prevent your money drain), there was another web form, and no tools to simplify the workflow.
By 2007 when I left, the network had more than a 1000 users, but in mid-2000s many already started switching to commercial providers.
Looking back, in 2001, the noncom network offered the best service in the city. By 2007, step by step, the commercialized networks upgraded their services, meanwhile every time the consumer cooperative made righteous explanations and did next to no improvements.
That’s no secret: in every commercial company someone is motivated to shut up and work better and harder. He either earns big money on this, or at least guarantees the company survival. In noncommercial organization, everybody wins a little bit, but nobody is incentivised this much to make big enough an effort.
What are the reasons for users to switch to OSM now?
A tile layer is too big a commodity to be offered for free, but the point is that the less gateways to OSM exist, the less reasons are for it to exist at all.
In 2006 our noncommercial network rejected the idea of unlimited plans as economically unsound. In reality, unlimited plans are very economically sound. The management insisted that free data traffic would be abused. Same way, OSM insists it’s unsound to offer tiles for free.
I’m no expert to give advices. Limiting the tile usage might be necessary right now. My point is that this is another step in a way from the users and the world, which leads to project not needed by anyone. Ask yourself, why WILL one use OSM? (will, not should)
Comment from Hjart on 19 August 2016 at 11:20
I think you are missing the point of OSM, which is not to be a tile provider or an end user product at all, so I think you can’t really compare your noncommercial network to OSM.
Our core “business” is in creating data, not in providing services to anyone.
Note that anyone can copy our data and create tiles based on them (just like i.e. Mapbox does). If you think it’s so “economically sound” you are perfectly free to offer your very own unlimited plans providing tiles based on our data.
Comment from Zverik on 19 August 2016 at 13:19
$4..5 per month is not “noncommercial”, it’s just cheap. There were two providers, one failed to build a sustainable business model. That’s all.
There are many services providing data extracts, tiles, iframes, routing, geocoding. Just for tiles, there are 20+ competing commercial providers. OSMF doesn’t want to become 21st and enter the competition, because that will lower resources to do our job: providing the data.
Comment from siberiano on 19 August 2016 at 13:32
@Zverik: You’re wrong on figures. In 2001 this was 120 rubles, which for a fast connection was almost free. No way you could find even dial-up cheaper than 500..1000 per month.
Comment from Zverik on 19 August 2016 at 13:41
Doesn’t matter. Had profit = commercial. Registered as a company = commercial.
Comment from siberiano on 19 August 2016 at 13:46
@Hjart: I personally have no business that will benefit on free tile servers, so I won’t do it. Never the less other companies offer a lot for free. Free OSM tiles are one of the reasons people install it in their websites and others learn about OSM. Of course, I have no data to draw conclusions and give advices.
Probably, you both are 100% right, tiles for big websites is too much of a free service.
My problem is with the reasoning behind this: “we don’t owe you” in 5 years made a project that nobody needed.
It’s pretty easy to make this with OSM. You may tell others to render tiles themselves, but this will work only until a commercial provider offers vector maps at reasonable rate and quality. For big websites this makes sense. Small websites always choose Google and other providers over OSM.
Data researchers (I’m one of them) use OSM for data. But the data is not free. Neither as in freedom (ODBL), nor as in beer: assembling and cleaning the data is really labour intensive. Community maintenance – commercial companes do this too, better and better. Nothing keeps users attached to OSM.
Comment from siberiano on 19 August 2016 at 13:47
@Zverik: it was non-commercial consumer partnership. It can’t make profit by our civil code.
Comment from Zverik on 19 August 2016 at 13:56
We don’t need users attached to OSM services: the more profit they bring 3-party service providers, the better. The data itself is free, and is guaranteed to be free forever. Services for processing the data are paid-for, and that’s a job for 3-party services. We offer data and we offer a choice. OSM owes you the data, but doesn’t owe anything on top of that. OSM is not a business, it can’t lose any competition because of that.
Please do not mix data and tiles/services.
Comment from siberiano on 19 August 2016 at 14:13
Well, tiles are not my business, it’s up to you.
But they are a gateway for newcomers to OSM. Do you agree on this or not?
But you do compete with commercial providers. To have money in the foundation, and contributors, it has to show value. The existing value of the data, as I pointed, is not forever.
Comment from Hjart on 19 August 2016 at 16:07
@siberiano. Please understand that OSM is not a business or commercial company. It’s a complete non-profit organisation, 100% run by hobbyists and volunteers. Our servers are paid for with donated money, so we just plain can’t afford to provide free tiles for everyone and it’s not our mission at all anyway.
OSM may to some degree compete with commercial providers, but competing is not the intention of our project at all and it doesn’t really need to. OSM exists because a lot of people want data that can’t be had at an for most people affordable price in any other way. For me OSM is my primary hobby. I’ve been contributing for 6 years and have seen several other attempts at predicting the death of OSM and despite that I still see OSM picking up steam beyond my wildest dreams. Ok, everything that was born or started at some point must also have an end, but I just can’t see OSM come to that for quite some years.
Comment from Zkir on 7 November 2016 at 08:23
If OSM dies for the described reason (lost completion for commercial cartography providers) it will be very positive thing.
For example, you condo has not lost internet. Quite opposite, you now have hi-speed internet with unlimited traffic for reasonable prices. Non-commercial consumer cooperative is no longer needed.
The same with OSM. When we started mapping Russia in OSM about 9 years ago, available commercial maps were awful (e.g. used in PNAs of that time). We needed maps and we had to start creating them ourselves.
Now commercial maps (especially end-users services) are much, much better.
If commercial maps (both tiles, navigation services and raw data) are available for low prices, why to bother with OSM?
Comment from Zkir on 7 November 2016 at 08:24
Comment from siberiano on 7 November 2016 at 13:56
@Zkir: makes sense.