Mercator Projection Hating Continues
Jeff Thurston asks a simple question:
Why take perfectly projected GIS data and stick it into Bing Maps or Google Maps? Isn’t it time that the 49th parallel was not a straight line?
I hate to break it to Jeff but 90% of the world has no idea that Bing or Google maps have a projection. People expect their web maps to look a certain way (you know where Greenland is bigger than Brazil) and because of that we’ll continue “sticking” our data in Bing or Google Maps. Anyway when projections matter, use something other than Mercator. I talked a little bit about my problems trying to work with the poles before here and here.
Much like the flat earth society, these Mercator haters want to make lives harder for average users to navigate maps. If projecting your data to Mercator is causing it to be incorrect, then you are obviously doing it wrong.
The GeoMonkey has always enjoyed the Mercator projection because he doesn't like going to the poles
90% of the world has no idea what a map projection is.
90% of the world thinks non-Euclidian geometry is a trick question.
90% doesn’t know what a map projection is? You’re off my a statistical measure or two…I’ll wager that about 99.999% of the world has no idea what a map projection is. Nor should they care, quite frankly. That’s shite for us wankers.
BTW – word on the street is ESRI is switching gears and going to Gerry’s world in the next release. That’s 440 years too late, if you ask a Binger. Probably a different perspective from the cognoscenti.
I agree and was going to suggest 99% myself, but opted instead for using repetition as a literary device.
The only projection that counts is the cube projection. You can print it out and have a nice cube. You can build a chair from that, or use it for soccer, put some things in it, make a Rubick out of it…
Hmmm…: An Earth-Rubicks-Cube would be a very cool idea.
This post raises a very pragmatic issue, but inspires me 2 more philosophical thoughts:
Now, philosophy apart, pragmatism rules: mainstream tools are good to reach the masses, so if this is your objective, long life to Mercator and to [put your preferred Internet major's name here]!
Projections are just the start of it…
See the world from Australia/NZ.
http://flourish.org/upsidedownmap/hobodyer-large.jpg
And how the US media see the world.
http://www.ted.com/talks/alisa_miller_shares_the_news_about_the_news.html
“Anyway when projections matter, use something other than Mercator.” — James, didn’t we all learn in Geography school that projections always matter? Selecting a projection is a key decision that goes into cartographic display, and I for one am sad to see the web mapping environment standardizing to web mercator, with limited tools for alternatives. Yes, as you point out, expectations matter (and let’s be honest, mercator caught on as the “standard” world projection long before the web) – but not always. frequently, chosing the wrong projection amounts to disinformation — we should be increasingly careful with thematic overlays on mercator bases. Is there any real reason why we shouldn’t have a multitude of options when it comes to projection? As a community of developers and geographers, we should demand it.
@Julie K:
Here here.
If Mercator “amounts to disinformation” as Julie presents, then why should we continue to use it when we all know it misrepresents geographic data? Sure Mercator is what MS and Google use but isn’t it the purpose of forums like this to pressure these organization to develop better products?
A key component to disinformation is intent. To that end, Mercator does not amount to disinformation. But Julie did not say that. She said: “chosing [sic] the wrong projection amounts to disinformation”. Depending on the chooser’s intent, Julie’s statement may or may not be true.
Good point,
But even if we can agree that ‘one projection fits for all’ is not a best approach, we have to consider that supporting 2 projections for web map services means setting up 2 caches (because don’t even dream about Google Maps-like performance if you re-project on-the-fly), which is a huge burden in terms of infrastructure.
Who will invest massive amounts of money to cover it without strong operational justification? (and I’m affraid ‘to be fair with everybody in the way we represent the world’ is not a strong operational justification).
Fair? – that’s something we all (should have) left on the schoolyard.
Is it not more reasonable to see more data presented on a familiar ground that is widely known and accepted, or should we make multiple allowances to accomodate a minority view?
It’s a free web, people. If you’d like to use something else, go ahead. I like that the ‘big uns’ are using Gerry’s view…the angles look pretty good in the area where i live. If it were up to me, I’d have to say I like Bucky’s Dymaxionic.
Web maps are mostly used in a highly-zoomed-in state, where issue of lack of a true scale disappears. Particularly since in web maps it is possible to provide (and the maps do provide) a dynamic scale bar that is always correct no matter what your zoom level or viewing latitude.
Things only fall apart when looking at the whole world at once (or trying to look at the poles), but that’s not the primary use case for these maps, I would argue (perhaps unconvincingly).
Convinced me.
Paul & Archie: Quite, but there are many online map applications that use Google/Bing/whatever to plot global data because there aren’t any realistic alternatives. Occasionally Mercator is okay (although it is ugly for this case), but in most cases it is an extremely poor choice of projection. For example, the much publicised map of world marine preserves would appear to suggest that virtually all of the world’s marine preserves are held by Denmark and one group of small Russian islands!
As it happens, we have a series of articles due for publication in a few weeks covering these subjects. They will end with a demonstration web site and how-to that shows you how to do a better job. (yes the demo website is not designed for zooming in to individual cities where these issues are generally not of concern)
Paul Ramsey’s comment is to the point – despite being poor for large areas, the Mercator works fine for the small areas that most people are looking at on the Web.
Still…there are applications where this is not the case, weather maps for example. It would be nice if the regional or continental maps were in a more appropriate projection, but that would require work and investment, and since nobody is clamoring for it or paying for it, the offerings sink to the lowest acceptable level, i.e., put it all in lat/long unprojected or Mercator.
I know from experience that a map of New York City’s five boroughs in unprojected lat/long (i.e. on a Cartesian, or is Platte Carre projection?) hardly raises an eyebrow here, even though everything looks skewed to me!
If you were an early explorer with only a sextant and a crappy clock (not until 1700′s BTW) to guide you, you would appreciate the finer points of the Mercator projection.
How about an octant?
Nowt for time and an astrolabe anyone?
take a page from Polynesians and their etak view of the world – everything was local.
Maps and the internet. Serious business.
[...] project, now an OSGeo project. MapServer renders individual tiles, at specified zoom levels, using the Web Mercator projection. These tiles are then cached using an aptly named Python application called TileCache. The map is [...]