[Openstreetmap] The bigger picture

[Tom Carden]

On 2/14/06, Richard Fairhurst <richard at systemed.net> wrote:
> Quoting Tom Carden <tom at tom-carden.co.uk>:
>
> > So you're just replacing 'work' in the current license* with
> > 'geodata'?  How does that help?  (Not being sarcastic or rhetorical
> > there - I'm not clear on the difference).
>
> Ok.
>
> "Geodata" is: there's a road from lat/long 51.13425,0.2342345 to
> 51.34652,0.13245.* It's called Albert Street. It's one-way, there's a pavement
> for pedestrians to use, and it's part of National Cycle Network route 71. (And
> any other amount of key/value stuff you want.) This is the essence of OSM.
>
> A "map" is: here's a black line on a piece of paper (or a JPEG, or an
> Illustrator file), depicting Albert Street. Because it's one way, we've put
> some little arrows along the side of it. (2pt stroke, solid arrowhead, 100% C
> 50% M 0% Y 0% K). We've written "Albert Street" along it in 3pt Myriad Pro
> Semibold. We're not producing a walking or a cycling map, so we've ignored the
> stuff about the pavement and the NCN route.
>
> The map is clearly a derived work from the geodata.
>

Sure. I pretty much accept those definitions, though the piece of
paper bit is irrelevant, and I wonder about a half-way thing - a
similar description of a vector file format which could be printed to
produce the 'map' above but which could also be transformed back into
the 'geodata' above.

> So why does it help to replace "work" with "geodata"? Because, to quote the
> first sentence on the wiki, "OpenStreetMap is a project aimed squarely at
> creating and providing free geographic data".

Quoting the wiki is kind of irrelevant.  There are lots of other
things on the wiki that are wrong :)

> We're not trying to create a
> cartographic reference library full of different styles of maps, just as long
> as the geodata coverage expands through the goodness of sharealike (as
> you say,
> "set the data free and all that").

Not sure what you mean by 'different styles of maps' - the aim of any
of my involvement in OSM has been to make the creation of maps from
GPS data as easy as possible.  I say maps there and not geodata
deliberately.  The fact is, the most flexible way to make editable
maps is to store the underlying geodata and edit that, then the
presentation is independent of the data.

>
> Three of the reasons why this clarification is important.
>
> One is the suggestion I made about producing lovely-looking maps. By
> clarifying
> ShareAlike in this way, we get easy funding for OSM without any loss to the
> project aims. (And yes, I'm volunteering to do the work.)
>

You're suggesting that by sleight-of-hand with the wording you can
define your way into a situation where a "map" (under your above
definition) derived from "geodata" (under your above definition) is
not required to be distributed under a sharealike license?

> The second is something that was mentioned earlier (can't remember by who,
> sorry).

Me :)

> It'd be great to do a Google Maps-compatible API using OSM data - a
> really superb advert for free geodata. At present, if we were to do that, the
> entire contents of any site using that API could be deemed a derivative work
> under CC-BY-SA; so you've got to make sure that every single piece of data on
> your site can be relicenced under ShareAlike, even assuming you want
> to. And it
> wouldn't be so great to have a "free" API that you could actually use in fewer
> situations than the proprietary one.

Well, we already have proprietary APIs that are free as in beer, but
restrict the creation of derived works.  Why not a free as in freedom
API that requires any derived works to be under a CC sharealike
license too?  Again, this doesn't prohibit people from selling their
data, just requires them to distribute it under the same terms as the
OSM data they derived it from.  That sounds totally fair to me - OSM
gets the benefit too since the derived works could go back into the
main database.

(I'm advocating a 'viral' GPL-y way of doing things here but I see how
if the license could be written as you suggest, we could go a MIT-y or
BSD-y way - people who "get it" would submit back to OSM regardless,
and it's a waste of time to force people who don't "get it" to play by
our rules.  See this recent Ruby on Rails interview for an example of
this sentiment in action
http://uk.builder.com/architecture/oop/0,39026558,39297360-3,00.htm)

>
> And the third is that, by removing limitations on the ways in which
> OSM-derived
> maps can appear, we make it more likely that these maps will be produced.

I think they'll get produced anyway, and if you loosen the
restrictions then you just encourage people to go off and do it alone
without feeding back into the project.  Rather the attempts to profit
from OSM be tightly coupled to the project than parasitically skimming
off the top of it.

> The
> other part of CC-BY-SA is Attribution, and I would love to see printed maps
> springing up here, there and everywhere with little OSM logos, saying "go to
> openstreetmap.org to find out how to help".
>

Me too!

> What I'm suggesting isn't anything new. In the full CC-BY-SA Legal Code, they
> already have an exemption like this: the concept of a Collective Work. This
> says, broadly, that if you incorporate it into an
> encyclopaedia/anthology/whatnot, then SA only applies to this one
> entry. CC's a
> literary licence, not a geodata one: we need the geodata equivalent.
>

I wish you'd said that earlier - it is an interesting distinction when
you look at it like that.  I just don't think it's one worth making.

Tom.