Why oh why are we seeing this Map/Menu bar on all these new “Web 2.0″ mapping applications? Take this beta example from the USGS National Map Viewer:
So you’ve been sitting back, watching all this great new Web 2.0 stuff and this is what you bring to the table? I know, lets see how much junk we can throw into a JavaScript API. The whiteboard on this one must have been intense… Just take every idea that someone comes up with and toss it into a ribbon interface. Sweet! But this isn’t a complex, specialized, niche application we are talking about. This is our (well if you like me are a tax paying citizen) national map. Yes the American Map! It should be something we are proud of, something we would run of a flag pole and salute!
They do expose some other ways to access the data, but don’t be fooled by the names. The Google Maps, Bing Maps and the rest are all just links to the ArcGIS REST API. That is how The National Map should be exposed. ”Here are the services, use them and create your own maps”. Might be a better way to handle it because the future looks bloated.
Of course we can’t completely blame the USGS for this, ESRI’s JavaScript and Flex API sample viewers have a similar abomination. Clearly GIS Analysts shouldn’t be designing user interfaces. Are we really going to use this thing for every web map API?
I mean we all love to throw complex concepts under a widget icon of a box with gears on it. Me? I can never remember if I’m supposed to use the compass or the globe to zoom in or out.
The one saving grace is that one day Google will just enable all this in their web map viewer making everything else irrelevant.
34 Comments
Can anyone recommend a good OpenSource webGIS project that I can use for a client? All I need to do is display vector and raster data. It would be good if it was in Silverlight but not a requirement.
OpenLayer – http://www.openlayers.org/
You can even use it with the ArcGIS Server REST API – http://www.spatiallyadjusted.com/2009/03/25/using-openlayers-with-arcgis-server-rest-api/
Apparently we all have way more screenspace than we need (or maybe the developers do).
This may well be true. Maybe we should all test our apps on laptops with little screens instead of the big dual monitors on our developer rigs?
Yes! I specifically use a 15″ laptop as a check when I build web sites and try to make the pages fit comfortably on that size screen.
How about using the web standard 1024 X 768 (effectively 1004 X 598) when designing these interfaces.
I think I made similar comments to the head of the User-Centered Design project at the USGS this week. There are two basic problems:
The Palanterra viewer was designed by the NGA for their analysts. Did you click on the advanced tab and then Add Data? You get options to specify that your WMS data source is Unclassified/Secret/Top Secret… Intelligence Analysts live in a world overrun by complex arrays of data. Palanterra attempts to give a degree of control over that data within a browser-based interface. The main purpose of the National Map viewer is to expose the richness of the data sets backing The National Map – not to provide an “everyman’s map”.
The beta release is focused on the viewer – specifically to start getting this kind of feedback. Underlying the viewer, as you mentioned, is a whole suite of RESTful services that can be consumed directly. But this isn’t the focus of this release. The “Demo Examples” page you linked to (which I’m thinking wasn’t supposed to be released yet) is actually a subset of what we were had access to internally to the USGS. The idea was to try to release early and often rather than trying to hold everything back until it was “done”.
I highly encourage you to build your ideal “TNMlite viewer” – maybe mix in some of your own data. You’ll quickly notice which layers are cached tilesets and which aren’t.
I’m designing an icon of a satellite…a nice shortcut for my users to the unsupervised classification tools. Yep. That’ll be nice and quick for them!
This is one thing that is still a gray area imho. There are very few left. It’s easy enough for anyone to build a gis web app. It’s easy enough to have a beatifull base map. But can you design a beatifull and well designed UI? Think not… for the looks of what is out there, very rarely do you see a good UI in mapping apps. When I do find one I start taking notes like a madman!
Why is this? Are there guidelines? How many of these projects have a designer on board? Someone who knows specifically about designing a UI? And even then, those guys usually don’t know anything about mapping tools… it seems there’s no easy way out! Help!
Duarte
Try Business Analyst Online from ESRI
Hey James -
“But this isn’t a complex, specialized, niche application we are talking about.”
Compared to street maps and atlases, I don’t think there are that many people in the general public who use USGS maps. Those who do use them tend to be a bit more spatially literate and their requirements are a lot more diverse. I suspect they are accustomed to living with swiss army knife architectures. Specialized apps (developed by 3rd parties) will follow.
@Kirk
Parcels are niche as well, but you know this is going to happen.
http://blogs.weogeo.com/pbissett/2009/10/29/google-will-open-source-national-parcel-map/
A national map is just that, an atlas of the United States. USGS Topo is so interesting to people, that Google felt the need to copy it for their maps.
View Larger Map
Great post. About time someone pointed this out. Seems like more and more sites are popping up these days overloaded with widgets and options. Great for those suffering from ADD, but these sites lack any practical focus.
Now… how about some examples of good web map interfaces?
Everyone says the one I designed for the st louis county crime map sucks
Glad I get to do spatial location/allocation models for the next few months instead.
(Though in one of those stupidly simple moves, bolding all the drop down selection boxes made it a lot easier to use.)
James:
I’m somewhat empathetic given that UI has never been a GIS strong point and there is little in the way of a set of accepted best practices. That’s why an initiative such as Cartography 2.0 (http://cartography2.org/) can fill a real need. My guess is that as templates become easier to work with, you’ll see a flowering of different UI experiences more closely tied to an online map’s specific purpose.
My other guess is that much of the momentum will come from the design/carto end of the spectrum and NOT the GIS community: we’re too obsessed with functionality over form and tool menus nested within other menus to understand the average user experience.
Personally, I spend the most time pondering the design decisions made by the digital crew at the NY Times. These are the pros pros who are getting some sophisticated data visualization techniques in front of a large general audience. Seeing the choices they are making makes my stuff immediately better.
Brian
yeah, the guys at the Times are good. When I get folks asking about whether we can do something like that with our online mapping system at the City, I regularly have to bring up the fact that stuff like “crime maps” are often have a very specific purpose and aren’t designing to show copious amounts of layers from Transportation, Parks, Aerial images, Cadastre, etc
That being said, there are a few good well-designed sites out there. I liked what the folks at Hanover County have created.
Now I just wish I could find their new online GIS website. It looks soooo much better than what I could find here: http://www.hanovercountygis.org/hanover/
Their new site is still in development. I will respond here once it is in production.
Stamen and Michal Migurski did a nice thing with Oakland Crimespotting (http://oakland.crimespotting.org) proving the point with the design/carto part.
I believe that UI will be more and more important. We can see this at a municipality level here in Sweden where I work – if the web map application isn’t as intuitive and fast as lets say googlemaps our information departments will use googlemaps instead. Old clunky ArcIMS webapps just wont do anymore.
And clicking the little black triangle in the upper left corner of the bar closes it up and the map bar is gone. All you see is a tiny square with a little black triangle – not obtrusive at all.
What am I missing? Is the selection of icons bad? Is the options the icons provide us bad?
Like Brett, I would ask that some examples of “good web map interfaces” be shared.
How does one define an interface for something that might be used by a very wide ranging audience?
Who is the audience of this “National Map”? Why shouldn’t aspects of this be a complex, specialized niche app? What exactly should the “National Map” be?
@HeavyG:
Closing the map bar does nothing. It still sits there calling out that it is a junk drawer of bad design.
Some good examples of maps?
Leaving it up the the user, not matter now many PhDs they have to figure out what analysis is under what tab or icon is bad design. The map bar interface is just a collector for these poor decisions. There is a reason why project such as MapFish (http://www.mapfish.org/) are getting such good traction, present the information in ways the user can understand. Not toss all the balls in a black bag and hope the user can figure it out.
Try Business Analyst Online from ESRI, it is easy to use and simple.
James -
Each of the sites you cite are focused on a small set of use cases. I suspect the USGS designers have not been given a very well defined set of use-cases.
While the map is not google-ish, awareness of the dangers posed by google cartography is growing … http://www.bustedtees.com/mapmarkerdeath
Which is why I said they should just provide the services and let people develop the maps. Google can just throw the layers in Google Earth and Google Maps and we’ll have more access to the data than this thing.
I tried adding the services to ArcMap and to GE.
No go on ArcMap, but I wish them luck as I could really use this type of service.
The GoogleEarth, now that’s a different subject. The feature I love the most about GE- arbitrary POV- is best implemented the way GE and WorldWind have done it. Labels are fixed relative to the view, not to North. This feature crushes other globe viewers, IMO. This feature fails when tiled caches have been built so that North-is-Up. I don’t know who is going to make that work, but until I can view my target from 315&45 and have the labels face me, I’ll pass.
The National Map may yet provide us with useful data. I just hope someone figures out how to do it right.
Federal govt and “do it right”! That is like trying to mix oil and water and produce a useable solution. Not going to happen anytime soon.
While we’re all talkin’ smack on the USGS (c’mon folks, that’s waaaaay too easy) did anyone notice the credit under “Viewer”
= Powered by Palanterra x3 =
Not sure if this community is aware, but Palanterra is the ESRI code-bolus our National Geospatial-Intelligence Agency (NGA) promotes as the solution for ‘information-sharing, situational awareness, interoperable collaboration, and geospatial intelligence.’
Anyone else troubled that the USGS National Map being “powered by” a tool that’s controlled by the US military mapping agency?
How long before “critical infrastructure” like schools, fire stations, city halls, police stations, and hospitals disappear from The National Map?
I’m just sayin’….
Yes the map bar is mysterious but there are other issues like “double click zoom and center” does not work. All you get is zoom.
How more fundamental can this get?
The point of focus for a user is where the cursor is – its leading the attention around the map and one should not be faced with traveling pixel miles to go and drag the map to center.
This is similar to why menu bars and all out-of-the-box commercial and technical map web map interfaces pose classic problems.
Who said you have to put all this hierarchical navigation space on the left? Its just another throwback based on static geo-enhanced thinking that has not evolved since Windows file folder navigation took over the world.
Why should I be forced to spend my time hunting and testing with map widgets and not being given tools to enhance my experience.
This is evidence that there is a long way to go and I agree that geo-centrics will not likely lead this next wave of development.
Here are two great USGS maps built on the Google API for the National Water Information System and the Annual Water Data Reports.
http://wdr.water.usgs.gov/nwisgmap/
http://wdr.water.usgs.gov/adrgmap/index.html
They use an ArcGIS server tile cache. Granted, they have a specific purpose, which allows for simple workflows to be built and dont leave you searching for a tool or purpose. I was impressed when i first saw these maps, and for someone who accesses this type of data often, i can appreciate the simplicity.
I agree about the map bar. Just because you can do something doesn’t always mean you should. But a universal way to display large amounts of information across multiple skill levels doesn’t exist. You have to know your audience.
Here’s my attempt.
http://www.portablemaps.com/example
Theres my application, we have been working in GIS applications for Guatemala. If you could check it out and send us some feedback would be great.
http://www.guaterutas.com/
Behind your rant, you raise a good point James – what is the mandate of the USGS? 1) to provide data like the seamless server (http://seamless.usgs.gov/) 2) to provide a front-end like a) Earth Explorer (http://earthexplorer.usgs.gov/) b) or EROS viewer (http://glovis.usgs.gov/) c) or Oil&Gas Assessments (http://certmapper.cr.usgs.gov) etc. 3) and how about GOS (http://www.geodata.gov), National Atlas (http://nationalatlas.gov/), etc.?
Doesn’t it look like USGS tried to rationalise the plethora of sites using NGA’s Palanterra, and provide user community a chance for feedback on your National Map (meaning US taxpayers)? This discussion will NOT go away anytime soon, judging by the same in other countries! Thanks for asking those uncomfortable questions…
Job Security!
You raise good points under your rant, but I’m not sure whether you’ve personally tackled the problem of building general purpose powerful yet easy-to-use map viewers for general users (not geo/analyst type people), because you act like it’s trivial. It’s really not.
Google Maps is great for a lot of tasks, I use it all the time, but I wouldn’t call it especially powerful. You can’t even easily measure straight line distances out of the box. I realize that’s not necessary because its target audience is people that are getting addresses, looking for a place to eat, or figuring out how long it will take them to get to a shopping mall or their cousin’s house. Which of course is the point…
Now if you do have additional requirements (like many people do), say to do mensuration, overlays, animated overlays, bookmarks in the viewer, drawing, exporting drawings for powerpoint presentations with annotations, letting people add georss feeds.. and so on. I can’t show animated cloud cover or add caveats and classifications. I can’t do simple buffering, geoprocessing services.. etc.
Google Maps can’t do that. It basically displays 2 basemaps and lets you get directions. And it’s GREAT at that. But it also has tons of money backing it, a huge team, and it’s targetting a very specialized usage case..
So guess what? The rest of us end up with a lot more stuff and you have to put it somewhere. Now I’ll agree that most web map apps out there don’t have great UIs. Our enterprise one sucks, that’s why I’ve started rewriting to make it more “task-centric” rather than “tool-centric”.
It doesn’t mean we’re not aware of the problem; most of us live in the real world where we have a thousand other things to do and while I’d love to have a dedicated UX person and designer doing it fulltime – it’s not gonna happen.