[OSM-talk] Things People Say

[Michal Migurski]

On Dec 28, 2011, at 2:21 AM, Frederik Ramm wrote:

> Google has a sales force - people who are paid to make you use their product. If Google brainwash their staff well enough, then it may even be a personal goal for their staff to make you use their product; but the second the employment contract ends, that's it.

Strictly speaking that's true, but Google's product is also good enough to overcome their lackluster sales team. Remember that the API was effectively beaten out of them by Paul Rademacher, Adrian Holovaty and others after their public release. Google Maps is *really good* at being a map portal, and we can't dismiss their organizational effectiveness by assuming that they're brainwashed.


> I hope that we are all very clear on this issue - OSM does not want a sales force. Advocates who come to like OSM and "sell" it to their friends are of course fine!

I agree that you are very clear on this issue, but I don't think the rest of us are. Successful projects grow to a point where specialization is needed, and I think OSM should have a better sales or evangelism function. Some companies call this "sales engineering", which means technically-oriented people who also see their job as smoothing the path for new users. Think about the change in Mapnik when Dane Springmeyer joined Artem Pavlenko and over a period of two years the "hard to install" joke became a thing of the past.


>> I believe it's time to
>> remove a few fire hydrants and turning circles and bump up some font
>> sizes in the service of a less-busy, more-comprehensible selection of
>> elements.
> 
> Doesn't MapQuest do that reasonably well, or were you thinking something different? We already have their tiles on our www.openstreetmap.org page, and we could easily switch the default style to use their tiles if everyone believes they are nicer. Saves us a lot of traffic too.

MapQuest's cartography is completely awesome, but it's not exactly what I was thinking of.

I like that tile-making is currently in the hands of users, but I'm also arguing that if we have a map on the front page, it should be a real, official set of tiles as a reliable service that people can really use. When I was making my terrain layer recently, a friend kept prodding me with the final goal: "all this code is nice, but the most widely-useful products is tiles."


>>> ETA: And as I go to OpenStreetMaps and see it struggle to serve me
>>> a page showing my house (and showing exactly the same wrong place
>>> for my house that Google Maps does, because it's using the exact
> 
> [...]
> 
>> This person is caught between the "it's a map" / "it's a project"
>> divide.
> 
> I think that this person is also caught between the "I'm in the USA" / "I'm not in the USA" divide.

I should be more specific: this person goes to maps.google.com, they see a big map with a search box, they enter an address and pan around to look at their house or hometown. They go to openstreetmap.org, and they see a big map with a search box so they assume OSM is filling the same need. It's clear from your mails that you think OSM fills a lower-level, more data-oriented need so we should *change our public presentation to fit what we're actually trying to do.*


> I'm a little tired of people like that and I hope that by drastically reducing the amount of map on our front page we will get rid of them.


Yes. Well, I would phrase this as "shape their expectations" but yes. =)


> Our product needs a bit of pretty packaging and customer service added before it can compete with the consumer friendliness of something like Google Maps; such pretty packaging and customer service can be provided by enthusiastic individuals, or nonprofits, or commercial entities - maybe even by other open projects. But I don't see this on our plate.

I believe that you are underestimating the importance of first impressions by using words like "pretty". I'm thinking of something like accessibility, in all its forms - how can OSM help the new visitor? If someone searches for information on what they can do with OSM data, does a useful wiki page come up in the results? Do we have a log of search terms people have used when entering the wiki or other pages? Can it be made public so we know what people out there think they're looking for when they find us?

Part of this process is acknowledging that certain tools and services have graduated from being somebody's volunteer pet project to core elements of the critical infrastructure, and should be handled as such. Osmosis, Osm2pgsql, the Planet downloads, and Potlatch all spring to mind.


>> Setting up your own tile
>> infrastructure is something that's easier now thanks to all the work
>> that's gone into Mapnik, Planet replication and other tools. Can we
>> make it easier in the eyes of a business, by showing how the costs
>> too are predictable and doable?
> 
> I think we could, but then I don't think that we need to worry. I do this as a business, and I am by far not the only person to do it.

So do we, but that's not really what I mean. There's a paragraph in the Nestoria blog post where they talk about hiring Andy to consult on running a tile server and ultimately decide that it's too hard or expensive or whatever. So, I'm not talking about businesses in the GIS / map space, I'm talking about businesses that are staring down the gun barrel of Google's sales force for the first time and trying to understand what alternatives they have to the GMaps API. Dating sites, real estate search engines like Nestoria, directory services, etc.

DevelopmentSeed has done amazing work with Tilemill, but I believe that it's not a useful answer to street-scale maps with countrywide coverage. There's still a hole here; I've been working to fill it with things like High Road and the new Planet page but through threads like the one on Hacker News and conversations with people who are interested in paying attention to OSM, I believe that we are representing ourselves in a confused light.

The biggest problem I can imagine solving is something you're hinting at above: if we are primarily shepherds of data and the means of its production, why do we look like a map portal?

-mike.

----------------------------------------------------------------
michal migurski- mike at stamen.com
                 415.558.1610