Mapping Google Maps

jgwebber writes "Google Maps is starting to cause a bit of a stir as Google makes the browser do still more backflips than most expected. In the tradition of dissecting Google Suggest and GMail, I've done a little dissecting of this newest service."

what about plotting waypoints on the map?

[garcia]

What I would like to see them add is something like what GPSVisualizer [gpsvisualizer.com] does. It will allow you to upload a GPX or LOC file of waypoints (from your GPS or various other programs) and plot them on a map. Because GPSVisualizer requires the SVG plugin (or native support) it would be nice to have an advanced application like Google has that doesn't require such support yet is as smooth/speedy as Google Maps is.

It would be awesome if Google could completely take over the commercial mapping software application market (ie Streets and Trips/Mappoint and Street Atlas) by enabling routing/directions between the points on the map. Hell, allow us to then download the planned route back to the GPSs via a GPX and that would really rock. I mean web-based applications such as maps.google.com and maps.yahoo.com have already taken over from older programs like Automap which just gave text directions and simple maps. Why can't they add even more features? I don't know anyone that asks for directions anymore. Everyone just uses the web-based software.

For now I'm just happy being impressed by the pretty scrolling. I'm excited to see what comes of this after the finish up the Beta.

Still doesn't work with Safari

[Anonymous Coward]

I find this interesting because Google's response, if you load maps.google.com in safari, isn't "we don't care about your platform, bugger off", it's a short, apologetic note saying that they don't work in Safari yet but you can try one of these other browsers. This seems to indicate the problem isn't with Google's javascript, it's with Safari, Google's javascript is more than Safari can handle.

Hell if I were a browser company I'd pay Google a small consulting fee just to find bugs in my browser. You know, give them some cash and say "have your javascript fellows write the most fucked up thing you can, i am paying you to break my browser".

Combination

[augustz]

Google:

- Nice company
- Cool services
- Sweet interfaces

That is a rocking combination.

The fact that they seem to be making stuff available under Firefox as well is also great.

Safari support

[Staplerh]

Hmm, I wonder if this is why I can't use Google Maps with Safari:

I also think it bears noting that Google is pulling out all the stops to build rich web apps, no matter how weirdly they have to hack the browser to make them go. And I strongly believe that this is a trend that is here to stay -- XHTML Strict/CSS/etc be damned. At the end of the day, what really matters to users is compelling apps that let them get their work done quickly.

I'm not necessarily complaining, as I can use Firefox, but it is too bad that even Google can't get a webpage to render properly on any modern browser, such as Safari.

Oh well, I don't know that much about any of these 'browser hacks' that Google is doing, but hopefully their promised Safari support will come soon.

Usage...

[Avyakata]

I think another, non-technical, issue is popularity. Online maps are a good idea, but most people who find them useful are already accustomed to Mapquest. Is there anything groundbreaking enough with Google Maps to persuade them to switch?

Re:what about plotting waypoints on the map?

[Moby Cock]

Remember that Google purchased Keyhole a while back. They are a satellite imaging firm. I wonder (or hope) that somehow this technology could be merged with the excellent Google maps. Imagine a "See Photo" button once you have found the location of whatever it was you were looking for. Now that would be cool.

Re:what about plotting waypoints on the map?

[ackthpt]

It would be awesome if Google could completely take over the commercial mapping software application market (ie Streets and Trips/Mappoint and Street Atlas) by enabling routing/directions between the points on the map. Hell, allow us to then download the planned route back to the GPSs via a GPX and that would really rock.

While I've fiddled with it and found the interface to have it's up-sides and down-sides (not really very big considering it's a web app.) I think you're looking for functionality at a whole other level, i.e. competing with Delorme or such, which could prove a challenge. I've got the Delorme package on my PC at home and the volume of information its processing to produce things would kill Google's servers. Besides, Delorme is Topo and Google's isn't, and Topo information is actually very useful when travelling. When I moved west I drove a truck, pulling a trailer with my car on it, up a road I never should have taken. Perhaps the squiggly road on a decent map would have given the appropriate clue, not to take that road because of the high pass. Google could at least include some elevation data in their images as an option.

Google Search Results

[prakslash]

For some reason, if one enters an address in Google Search to find a location on a map, the resulting search results still point to MapQuest and Yahoo!Maps. (See example) [google.com]

They need to update that.

new Google browser (alpha) is intriguing

[peter303]

I'm hoping they decide to ship it. There are several very inventive features. And solves some of the issues mentioned in this thread.

I like it!

[__aammuz5019]

Hello all,

I am going on travel this afternoon and I had already printed out my maps and directions from "another service (MapQuest).

Today, I did the same using Google Maps, and I found the interface much easier to use, the maps pretty good and the driving directions less complex and easier to read than the ones I usually get from other services.

I say "pretty good" on the maps, because of two things: First, several of the maps had slight discontinuities when I printed them, versus the way they looked on screen. At the left edge, there was a vertical seam where the map pieces were shifted with respect to each other. Not a big deal, as this had no great effect, but has anyone else seen this? Second, the shadows from the "start" and "end" push pins obscured some of the map information and blobbed the end of the route a bit. Does anyone know how to turn off the shadows?

Just my experience, YMMV.

smp

Nice as a video game engine

[192939495969798999]

I'd like to see a MMORPG ported for this, like a web-enabled version of ultima 1 that shows where everyone's looking, and we can all interact. How awesome would that be? Totally.

Google not promoting their own map app?

[danielcole]

Do a phone number or address search from the main page, and Google suggests you can look at Yahoo! Maps, and MapQuest for directions, but not their own service.

Why aren't competitors beating Google to market?

[Anonymous Cowdog]

Google is hitting a lot of the obvious sweetspots for improving the user experience. Some of them are obvious only in retrospect. But we know their competitors have smart people, and they do UI research, and they have resources. Why does Google come out with innovation after innovation?

I have three answers. I wonder which ones are valid:

1. Laziness
2. Encumberance with legacy political and business issues (is feature x threatening to partner Fooinc, how can we hang ads on this, etc.)
3. Focus on fancy-pants analysis of numbers (data mining to try to optimise, rather than revolutionize), leading them to be blind to simple measures like using Javascript and caching lots of content in the client.

What other reasons are there?

Re:Quick review

[DoorFrame]

Yuck, just tried map24 for the first time and was not happy with the EXTREME load times. I don't want to sit around while silly java applets load up. Also, it didn't actually pinpoint my home address, just gave me the street. Google maps is better.

Re:Quick review

[Proney]

While on the continent view in google maps Canada isn't marked at all, but if you zoom in on it the streets are just as finely marked as their American counterparts. The directions functionality also works just fine with Canadian addresses, although it did choke a little coming up with directions from my native Toronto to my current Seattle... ;)

What WOULD you call Google's approach?

[dpbsmith]

It's not quite AI, yet Google comes closer to realizing the fantasy of Isaac Asimov's Multivac than anything else I've experienced before. It's very weird: the impression that Google gives is that it does NOT understand your question, yet it DOES manage to find the answers you want.

It's not quite user-interface, in the sense of elegant widgets or consistency or any of that stuff. Google's traditional search features could almost run on Lynx on a green screen. Maybe they can. Google Maps is visually spiffy by comparison to Mapquest, but it's nothing we haven't seen in standalone programs years ago.

It isn't really "search." Or at least, if it is, with every new thing they roll out, Google does an amazing job of expanding my notion of what "search" means. What does it mean to "search" on "250 pounds in kilograms?"

Something that Google seems to share with Apple is some sort of courtesy or kindness or service orientation to the end-user. It just works. And unlike Microsoft or Apple, Google's services seem to come with fewer strings attached.

One of the things that delights me about Google is a certain kind of freshness I haven't seen elsewhere as often as I'd like. They have the characteristic you used to see in innovative software that when you describe the latest Google feature, it doesn't sound all that new, yet when you use it you get that feeling that something unexpected has been revealed.

Re:Why aren't competitors beating Google to market

[Ced_Ex]

4. Google allows side projects which likely hold more interest than actual business projects, hence are produced with creativity and passion.

Where this is going...

[RCulpepper]

I see some features that will tie in well with this. It already makes Local Search a lot more handy. I could see Google using aggregated GSM phone locator signals to forecast traffic patterns and then, after asking you when you intend to start and end your trip (so it can route you around traffic), estimating when you'll want to eat lunch, etc, so that bricks-and-mortar restaurants, gas stations on the selected route can pay for advertising - it's one segment of the economy Google has not yet touched.

Re:Google + DOM = Mozilla Juggernaut

[bcmm]

Have you seen this [google.com]

Re:Quick review

[Anonymous Coward]

Sorry, map24 is nonsense. A poor interface and poor quality data - even for canada. It looks like a front-end to an ESRI ArcIMS server - and from my dev experience with it, I know that that technology is a beastly dinosaur. Google's solution is more elegant and effective - AND, from what I can see (as a Canadian) hosts more accurate maps.

Will this bother some Privacy Fanatics ?

[brandonp]

Of course, will this bother some people who are fanatical about Privacy issues?

John Smith in New York City, NY [google.com]

Depending on how the results are categorized and obtained, this seems like it could be a hot issue.

Brandon Petersen

This might be intentional.

[Kickasso]

Map publishers often insert small errors in their products to thwart copyright violations.

Made from images?

[Anonymous Coward]

I live in a semi-rural area where we have pretty long driveways (.25 mi). My driveway, and some (but not all) of my neighbors show up on this map. Interesting.

Doesn't comply with W3C-suggestions

[DancesWithBlowTorch]

I also think it bears noting that Google is pulling out all the stops to build rich web apps, no matter how weirdly they have to hack the browser to make them go. And I strongly believe that this is a trend that is here to stay -- XHTML Strict/CSS/etc be damned. At the end of the day, what really matters to users is compelling apps that let them get their work done quickly.

At the end of the day, Google _has_ cared about XHTML Strict/CSS/etc. From the head of googles web app:

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd ">
<html xmlns="http://www.w3.org/1999/xhtml">

And I'm quite happy about that. There's a reason why the w3c gives out definitions for html et al.

Google uses XUL

[fulldecent]

Here's the big secret:

Google uses XUL to develop all their rich websites. For example: Gmail, Maps, Groups and others on the way. This natively XUL interface is then converted to HTML/CSS/JavaScript that we can see and run. This conversion is done by a program Google wrote a while ago and the conversion is very simple. Of course, it's not perfect and needs to be loked over by hand. This is how Gmail is compatible now with all the other browsers.

In the future, when they decide it is time, they will publish their XUL interfaces side-by-side with their current interfaces. I'm not trying to give any hints, but this is related to a large push that Google is going to make to support XUL technology and will happen by the end of this year or early 2006.

Re:Google uses XUL

[christopherfinke]

An interesting example of Google XUL was mentioned earlier in the thread: Google XUL search page [google.com].

In addition to that,

[jdgreen7]

Imagine if they started using this for real estate searches, too. Just link all the local back-end MLS listings to whichever region you're searching in and end up with an MLS service that's better than what most realtors pay for. By partnering with some of the larger realty companies (Century 21, Remax, etc.), they could probably take over a good chunk of the industry in a matter of a few years and make it much easier for individuals to shop for houses without the assistance of a realtor...

Another example of an excellent online map

[Bio]

I want to point you to another example of an online map:

map.search.ch [search.ch]

This is Switzerland and not the US and it uses aerial photos with an overlay of vector street data. The resolution is amazingly good.

Furthermore:

• It is implemented with (D)HTML and Javascript only (like Google)
• You can move the map holding the mouse button pressed (like Google)
• You zoom in and out using mouse buttons 2 and 3

For me it's one of the most fascinating applications of web/html/javascript technique.

It's not easy.

[Kickasso]

The map is tiled, so you need to somehow produce a big image from all the pieces and then convert it to JPEG. Coding it in Javascript would be a real mess, and doing it server-side kinda defeats the point.

You can easily convert PS or PDF to JPEG with ImageMagick (yes I'm a geek).

Re:Google + DOM = Mozilla Juggernaut

[hey!]

Yeah, javascript might be a toy, although there's a bit more there than meets the eye, but DOM is definitely not a toy.

Languages are important, but even more important is the runtime environment they have access to. If the environment has the basic stuff you need, then even a crappy language would be pretty powerful. Think of a templating language like velocity -- it's not designed to be powerful by itself, but to be very convenient to integrate with a context that supplies it with everything it needs to do powerful things.

Years ago, in the era of of the 16MHz microprocessor, I had the problem of writing an Exel spreadsheet that required lookups from huge tables. Using VLOOKUP took hours. So I implemented a double hash algorithm in the Excel macro language. Mind you, this wasn't VB for apps, this was the nasty old lotus-y macro language. It turned out to be easy, because the spreadsheet environment provided most of the lumber I needed, I just had to snap it together.

Re:JS / HTML graphics: iWon Prize Machine

[nautical9]

Although this doesn't talk to a server while it's running, it's still the most impressive DHTML app I've ever seen: DHTML Lemmings [snesorama.net].

Really opened my eyes to the possibilities of what JavaScript/DOM can do. Glad to see Google, iWon, and other sites finally starting to make use of it.

Re:what about plotting waypoints on the map?

[parkrrrr]

And fortunately for us, Google makes it easy by giving us JavaScript code to decrypt the "points" string, and it does indeed contain latitudes and longitudes. Here's the equivalent Perl code, with almost all the same variable names as their JS code, with leading indentation stripped (thanks Slashdot!) Encoded string in $XX, decoded lat/lon pairs to stdout in CSV format. Feed that output to GPSBabel as "arc" format, and you should be able to simplify it and upload it to your GPS receiver.

my $Ch=length($XX);
my $pb = 0;
my @aa;
my $Ka = 0;
my $Pa = 0;

while($pb<$Ch) {
my $ub;
my $oc=0;
my $Fa=0;
do {
$ub=ord(substr($XX,$pb++,1))-63;
$Fa |= ($ub&31)<<$oc;
$oc+=5;
} while($ub>=32);

$Ka=$Ka+(($Fa&1)?((1-$Fa)/2):($Fa>>1));
push @aa, $Ka;
$oc=0;
$Fa=0;
do {
$ub=ord(substr($XX, $pb++, 1))-63;
$Fa |= ($ub&31)<<$oc;
$oc+=5;
} while($ub>=32);
$Pa=$Pa+(($Fa&1)?((1-$Fa)/2):($Fa>>1));
push @aa, $Pa;
}

for (my $i = 0; $i< $#aa; $i++) {
print $aa[$i]/1e5 . ', ' . $aa[++$i]/1e5 . "\n";
}

Re:what about plotting waypoints on the map?

[Deviate_X]

http://map.search.ch/ [search.ch] is more similar to Keyhole (on the dynamics) and and even more interesting use of dhtml than googles first attempt.

Re:what about plotting waypoints on the map?

[parkrrrr]

And of course it has bugs (mine, not theirs.) The two places where it says (1-$Fa) should say (-1-$Fa).

Re:Old info

[Leatherman]

Strangly the map of my hometown is a mix of outdated and new info. There is a street on the map that was put in about 3 years ago, but a road around an entire man made island is missing. The island has been there at least 20 years.

http://map.search.ch/ rocks!

[jpatokal]

I thought Google Maps was spiffy, but this [search.ch] is just insanely cool! Took me less than 1 minute to spot the house [search.ch] where I lived near Geneva over 10 years ago, and the satellite map even shows the community swimming pool and the nearby farm sheds full of stinky fertilizer. I can almost hear the cowbells ringing...

Note that you can also check the "Strasseskarte" box to switch between the satellite view and the just-the-facts-ma'am road map view.

Cheers,
-j.