[Strategic] Fwd: Subject: Forks and such

[TimSC]

On 29/08/10 20:24, Jim Brown wrote:
> If the data is forked, the two (or more) forks will continue to diverge over time...  Dual contributions are possible of course of new data into both forks, but any kind of editing will increasingly be impossible to apply to both sets.
>
>    
[snip]
> So specifically, the impact of forked data sets is substantially different from " multiple editors, renderers, loggers, navigation apps, routing Applications" that are acting on the same data set.
>    
I agree that people who try to add data face a potential duplication of 
effort, particularly as databases diverge. Say I want to add a map icon 
for historic=wreck to all renderers. How much source code is in common 
between osmarender and mapnik? about zero. If I want to add a tool to 
the editors, how much source code is shared between potlatch and JOSM? 
About zero. So, we have already accepted that some effort might be 
duplicated. Otherwise we should ban the alternative editors and 
renderers. One might argue that two forked database branches will always 
have more in common between them than many OSM tool's source code! (As 
long as the same tagging conventions are used, of course. Many tools 
don't even share a common programming language.)

The point is these tools have different goals and different users. The 
duplication of effort in maintaining all these alternatives is justified 
because it suites the specific needs of different groups of users. 
Similarly, there are differences in peoples needs for licensing. Some UK 
mappers might want to be compatible with OS opendata. Some Australian 
mappers might want CC-BY-SA to allow imports. USGS wants PD, so they can 
reuse the information to create more PD data. I could list more 
examples. So the analogy of diversity in users requirements and 
solutions hold between licenses and OSM tools. Also, adding the same 
data to every forked database, or adding the same set of features to 
every editor would probably not even be attempted or be desirable.

> In essence, unless I am missing the point (and I may be) the ability for any individual mapper to edit upon, and contribute to, multiple forks decreases as the forks diverge.
>    
The effort needed to contribute to diversing forks increases, but not a 
mapper's ability.

TimSC