Microsoft Invented Google Earth in the 90s, Then Blew It | Hacker News Reader

Microsoft Invented Google Earth in the 90s, Then Blew It

motherboard.vice.com

123 pointsjimsojim81 comments

Microsoft invented it? I recall working on a digital mapping system in the late 80's that was very similar. We had a very large drum scanner that we used to scan paper maps and imagery at various scales. First we'd rectify the result, then we'd do a color reduction to both cut down the file size and to match the capabilities of the machines of the day. Next we'd slice the image up into 128x128 pixel tiles and assemble the result into a database of various scales that allowed a user to zoom, pan from map to map, overlay symbology, etc. At the time we were working with 300 MB drives, so we were very restricted and had to create specific databases that matched a user's need.

That's very interesting. How was the digital mapping system named? Which company did you work for? Who were the customers?

I named it. I called it the "Digital Application Map Manager" or DAMM. That way my boss had to ask "When will those DAMM maps be ready?". We had a sub-contract with the Jet Propulsion Laboratory but the ultimate customer was the US Army.

Genius. :)

Having worked on one or two SIBR proposals, I appreciate a good (mis)use of the government's perverse acronym fetish. Bravo!

I had a professor who worked on an IBM research project back in the late 80's to build something like Google Earth and Maps. He also made some important contributions to GIS development and research. So yea... I don't think Microsoft invented it by a long stretch.

A guy I worked with at Xerox about 10 years ago worked on a government project similar to yours back in the mid-80s to mid-90s. He never said it directly but it is clear it was for the CIA. Apparently even back in the 90s they had shockingly high-res images of a huge amount of the planet.

One thing he said that always stuck with me was "whatever we have free access to today you can bet the [US] government had something just as good 20-30 years ago".

It is very difficult to process such an idea though, at least for me. When I look at what we have today I find it almost impossible to think such a thing was possible 20 years ago or that governments might have things 20 years ahead of us today. It seems like a few years ago things I use today were impossible. Technology moves so fast I can't really comprehend 2 or 3 years ahead anymore!

Very true- image mosaicking of aerial and satellite imagery has been around for decades. There were plenty of GIS and IMINT software systems way before TerraServer. What MS did was put all the pieces together with a web front end that nontechnical people could use. Hopefully that's what the article means by "invented", although that's still a misuse of the term IMO.

Tangential rant unrelated to the comment I'm responding to: the media misuses the term "invented" all the time as a synonym for "created a product". All it does is spawn a million pointless threads in tech forums about who invented what and who should get credit for what and why everybody should ignore the contributions of people who stood on the shoulders of anyone else, ignoring the fact that everyone does that always for nearly everything. Witness any thread where someone says something vaguely charitable about Bill Gates or Steve Jobs, and inevitably some first-year CS undergrad comes along and rants about how those guys did nothing because it was really their engineers that did everything, and inevitably some other first-year CS undergrad comes along and rants about how those engineers did nothing because they were just building off of what some lesser-known engineers did before, and eventually that regresses into yet another first-year CS undergrad ranting about how Dennis Ritchie should get all of the credit for everything that was ever done after 1969, while conveniently ignoring the giants whose shoulders DMR sat atop of.

And finally, it comes back around to some grungy middle-aged guy such as myself ranting on HN about people ranting about people ranting about things, so full circle I guess. :)

Important to note that there were a lot of mapping companies in the 90s/early 2000s, but the breakthroughs were really Google acquiring a number of mapping companies like Keyhole, where2. The big leap forward, of course, was the hyper responsive Ajax interface that google introduced around 2004.

Yeah. The seamless "scrolling a map in a browser thing" blew everyone's mind. It did a lot in cementing the image of Google as a visionary company.

Oh yeah. I remember when Google Maps first came out. Nobody thought that this degree of interactivity was even feasible in a browser. It was really revolutionary.

Yup. And originally done with a 1 pixel iframe rather than the XMLHttpRequest method if I remember correctly.

revolutionary ...

The emergence and craze of AJAX blew everyone away after that point. People realized there was a thing called XMLHTTPrequest. That technology had been my dream ever since I learned HTML. The tech came out in 1996 by Microsoft and nobody used it until 2003.

At this time, Firefox took the torch of keeping the web moving forward. So a whole buzz around web tech reignited.

Those were my days of inspiration.

You're probably off by a few years there and, of course, Microsoft used it for their browser-based Outlook client. One could write a similarly wrong-headed article about how Microsoft invented Gmail.

"Nobody used it" because until the early '00s bandwidth was so precious, and latency so huge, that making multiple HTTP calls was insane in most target markets (US/EU). So you would try to cram everything you could in one big request and push it out. JS was also very immature (both as a language and browser runtime, with differing DOM support etc), and computers could hang if your code was too expensive.

When DSL, cable and fiber started becoming commonplace, computers got faster and browsers started shipping compatible and optimized runtimes, then advanced techniques became really accessible. Eventually jQuery came out, making all browsers work the same, and the rest is history.

do you remember all those too-cool-for-school types who kept whining about the AJAX acronym and how it was overhyped and BLAH BLAH BLAH

Ironically, though, the original Google Maps didn't actually use XMLHTTPRequest, at least not for the trackable maps— that was all just moving <img> tags around and changing the src attributes on them.

XHR may have been used to fetch search results, although it later switched to a JSONP interface with generated script tags so that the various APIs (geocoding, directions, etc) could be used by third party sites on other domains.

Ironically, neither Google Maps nor Google Mail (GMail) used AJAX/XMLHTTPRequest in their first version. The first Google product that used AJAX came later, it was Google Instant Search (search box improvement). Google Maps and GMAIL used hidden iframes to load data in the background. Around 2004/5 only Internet Explorer 5+ and very recent Mozilla builds supported XMLHTTPRequest (without bugs). As far as I remember Maps and GMail switched to AJAX approach around 2006.

Indeed hidden iframes were the original way in which these effects were implemented.

BTW: MICROSOFT invented Ajax. Really! See Garret Smith on this topic.

http://garrettsmith.net/blog/archives/2006/01/microsoft_inve...

They didn't know what they had...

> The tech came out in 1996 by Microsoft and nobody used it until 2003.

Not really.

Ever since there iframes things like that could be implemented. People just didn't care...

I remember that back around 2003 I was making a highly interactive site that used iframes in a similar way - the parent page set the URL, the child one called window.parent.something() to pass data back.

I actually had no idea that I was on the cutting edge in this one-page app world at the time, I was just noodling around at university. But it was a lot of fun.

I used Java to accomplish the same thing as AJAX, back in 1996-97 or so. I embedded a hidden (1-pixel, transparent) applet that opened a TCP socket to a server.

The major browsers all allowed applets to expose methods to JavaScript, so I provided a simple API that the page used to send requests and receive responses, which the applet multiplexed on the same connection.

I also used it to receive notifications from the server; the system I built was a teleconferencing system, and so while a call was going on, the server continually sent status messages about parties entering or exiting the call, and flags about which party was speaking (based on audio monitoring). Requests were used for commands like disconnecting or muting.

It was pretty solid once we figured out how to make applet instantiation resilient (the browser/applet interface wasn't very stable). In fact, it was better than XMLHttpRequest, because it was just one, compact, long-running TCP stream; it was pretty much a websocket, before there were websockets.

Interestingly enough, the technique was used pretty heavily in a web-based BI software product that I wrote back in 1999. The UI was designed to be responsive and interactive, and I had to 'invent' a lot of the AJAX-type features and techniques that were later well-known. The problem was that most software back then (including this product) were 'enterprise software', limiting diffusion of ideas.

I used this and it was incredibly slow. Clicking the arrow buttons moved the map by a fixed amount. In my opinion the breakthrough of Google Maps was the Javascript interface that let you slide around and zoom in and out, plus the engineering behind it that served the map tiles quickly.

I used it too and, indeed, it was slow - but I was also using it on a Pentium 90 CPU and a 28.8kbit/sec internet connection.

I don't know if gMaps would have run any faster on that hardware back then, especially if it was all jscript-driven.

>especially if it was all jscript-driven

This is going to reveal an embarrassing lack of historical context, but could Google Maps have run in the Javascript of the '90s? Has the language changed significantly since then, or do we just... use it for everything now?

JScript was introduced with IE 3 (1996)

Dynamic HTML (modifying the page with javascript) was introduced with IE 4 (1997)

iframes were introduced with HTML 4 (1998) I think.

XMLHttpRequest was introduced with IE 5 (1999)

So 1998 era, maybe, but Javascript had a reputation as being slow and buggy. Flash, Java applets, and even active X were more acceptable technology at the time.

(Not trying to dismiss Netscape in my timeline -- I wrote it with respect to Microsoft implementing it.)

Classic embrace, extend, extinguish, JScript(1996) was like JavaScript(1995), but a little late and not quite the same thing. Which made JavaScript seem a lot more buggy than it was because MS implemented something just a little different.

As I understand (per Wikipedia [0]) Microsoft invented XMLHttpRequest in 1999 and nothing we now consider Ajax could really exist before that.

[0] https://en.m.wikipedia.org/wiki/XMLHttpRequest

We now have the horsepower in the browser to do it. We're a long way from 80586s at 90Mhz with 4MB of RAM and 60MB of disk.

Not just that, but the current engines are much more efficient as well. For example, when the first JIT engine was released for Firefox, it was an order of magnitude faster than the previous engine, even on the same hardware (for pure computation).

Even if similar code could have been written, it couldn't have been usable on the average home internet connection at the time...

I also used it just when it was introduced, but with a 2 Mbit/s uplink at work. And a 10 Mbit/s student residential uplink at home connected to a 34 Mbit/s link that the Linköping University had via Swedish University Network (SUNET). (Back then 34 Mbit/s was something fantastic!) I probably used a P133 with Linux at home and a 110 Mhz Sparc 5 at work. And Netscape was the browser in both cases.

My vague memory from then is that the server was super slow, and that my connection wasn't the limiting factor. Still, it was fascinating to play with for a few hours.

This was before P2P pirating was a big thing. Pirating was mostly done via SMB shares in the local network (no-one was afraid of getting caught since the authorities had no clue at all, and also because it felt like we were spearheading something new and interesting), so the external link didn't suffer. Online video at that time was like 50 kbps RealVideo streams, at most.

>Back then 34 Mbit/s was something fantastic!

I can't even get 34 Mbps now. This was 20 years ago?

Well, it was shared by like 20k (occasional) users and 1-2k students with campus Ethernet network connections. (Still, it never really got saturated. There wasn't that many interesting AND bandwidth-consuming services out there in the rest of the world.)

I think Sweden was quite good at getting people connected at decent speeds. In retrospect, I think this was a very wise way to spend tax money.

My personal Internet-connection history goes like this:

1994: 28.8 kbps modem at home in a small village. It cost like $2/hour.

1996: 10 Mbit/s - residential student university network, for like $10/month

2000: 10 Mbit/s - moved away from the student area to an apartment with a commercial connection (Bredbandsbolaget - they got loads of investements, some of it in the form of hardware from Cisco). $30/month

2002: Upgraded to 100 Mbit/s. $30/month

2009: My homeowner association accepts a 10 year deal where all apartments get 100 Mbit/s Ethernet-based access for $10/month (in exchange for "monopoly" - everyone needs to pay $10).

(Since 1996, it's been all direct Ethernet, no silly old stuff like DSL or cable.)

UK: I'm on adsl over copper (actually solid aluminium) wire from a phone exchange 6km away. No plans for fibre upgrade. 2.4Mbit/s down and something like 0.8Mbit/s up costing £12 per month to the ISP and around £65 per quarter to the national phone company for landline service. All this and I'm about 2km from the centre of a major city.

"There wasn't that many interesting AND bandwidth-consuming services out there in the rest of the world."

MIT had a video capture system. If you told it your IP address, and did the xhost + correctly (we were so trusting back then), it would pop up an X window on your machine and show video. We watched part of the 1994 World Cup that way, at a few frames per second and 1/3rd of a continent away. And no sound.

Regarding network in Sweden, part of me drools over the $100/month 1Gbit/s offerings from Com Hem and Bredbandbolaget. The bigger part of me wonders if the extra bandwidth would be worthwhile.

Welcome to Sweden! I signed up for 100/10 mbps when I moved in to my (then newly built) apartment in Stockholm, in 2006. Recently (last year I think) my provider was installing some new equipment, and when noticing the plan I was on asked if I wanted to upgrade to 200/100 mbps at the same cost. Said and done.

The idea of city wide LANs started taking off in the late 90s and early 00s, and many (most?) new buildings from the mid 00s onward have fiber optics wired throughout for the day they do get connected. This was the case for my building, but it is by no means unique. Many older apartment buildings and single family residences also get retrofitted to take advantage; an operation which these days is usually pretty cheap all things considered.

Even smaller cities often invest in such infrastructure. For example, in the small city of Strömstad (population ca 12000) you can connect to the local fiber network from nearly every home. The explicit goal is to get everyone connected.

Sweden has great IT infrastructure, a testament to forward thinking in the 90s. (Probably fueled by lobbying from technology companies like Ericsson.)

Just before the 2001/2002 crash, there was a huge wave of IP network investments in Sweden. A lot of it came from America - US pension funds investing in Cisco and Bredbandsbolaget, in turn investing in Swedish networking infrastructure.

In the 30-apartment house where I've lived since year 2000, there's a monster Cisco router delivered in year 2000. I think the street value then was around $1-2M.

Thank you america :).

You're welcome. Any chance you could send it back?

This headline is pernicious clickbait. Why not say "Microsoft Invented a Google Earth precursor, but never took it seriously"? Because that would not force you to click through to understand how they "totally blew it."

Spoiler alert: They built TeraServer map service to demonstrate that their database product could handle a terabyte or more of data. They got tons of user interest, but didn't use the information they got from searches or the user emails to iterate on their project. They could have developed a dominant mapping service (the author's thesis, sort of), but they let the project languish.

"According to a USA Today article from June 22, 1998, the initial plan with Terraserver was to list every single transaction in the history of the New York Stock Exchange online and make it searchable. But that was only a half terabyte of data. Microsoft needed something larger."

I'd be curious to see what the size of that collection is now, post advent of high frequency trading.

Side note, the article in no way explains why Google Earth would be considered a success story. The lead engineer from this project went on to Bing Maps, it's not as if MS decided geographic data wasn't worth doing.

The US equity market does about 4b in transactions a day; let's say the average trade size is $1000 dollars (probably an aggressive estimate; I'd expect it actually to be higher), and you record the ticker (4bytes) + price(8) + quantity(8) for all the transactions, you'd get --

(4+8+8)(4e9/1000) 252 (trading days) / 1e9 = 20GB/year

lets say you have these for 100 years, you'd still only have 2TB. And this if for the entire US equity market -- NYSE is only one of 11 protected exchanges (and only trades tape A), so the above estimates are pretty optimistic.

Now, if you wanted to record every message and not just every transaction, then you get into serious data (on the order of ~2-5TB/day).

Heh, I remember TerraServer. It was pretty amazing to me at the time (late '90s), the first thing I did was find my grandparents' property and sure enough saw a few shadowy figures which might as well have been me with them and the dog given the date on the tile. When Google Maps (and maybe MapQuest too?) started doing maps in color I was a suitably impressed kid.

I almost feel bad for Microsoft. I think they were far ahead of their time in many ways, and ended up getting in trouble for it.

They had some pretty cool tablets, but unfortunately the hardware / internet infrastructure wasn't really there to support it. They were the first (major at least) to build an OS underpinned by a browser. They were early to the game of running a real OS on a phone.

Unfortunately they weren't really able to take advantage of those. They seem slow to get into the tablet game the 2nd time around, and even now they're still playing catch-up (although I've heard the Surface things are pretty good). Their phones are consistently underachieving for whatever reason. They're still struggling to find the right model for Windows ever since XP/7 (I think they've finally gotten on board with the Apple model of the OS is freely updatable, but I'm not 100% sure of that).

They've done some great work, but man, they can't seem to get out of their own way sometimes.

The TerraServer project split into two branches, read both Wikipedia pages: https://en.wikipedia.org/wiki/TerraServer

Also NASA World Wind got popular, just before Google Earth, and was initially more advanced: https://en.wikipedia.org/wiki/NASA_World_Wind

I completely forgot about the response to aerial photography when it first came out. I remember having some concerns about the privacy implications and feeling 'exposed' on the internet that anyone could see my home. Funny how much my mindset has changed.

Except if you look at the screenshots, there's no information.

That's the key thing with google maps or google earth, it's the information around what you're looking at that is important, and google understood that.

I remember using this on my first 486, maybe a 386 on a CD that came bundled with Microsoft Encarta. This was just a hair before the internet started taking off on modems and such, I beleive we might have had an internet connection but it was still at the point where sharing floppy's and CD's were the best way to share software.

Heck, Google didn't even invent Google Earth, they bought a company named Keyhole and renamed it.

Fun Fact:

Microsoft had Instant IMAGE Search over 10 years ago when they launched the live.com product (if anyone remembers). Much much before google had instant web search. Start typing and images would instantly fill the screen. This was at a time when AJAX was getting popular. Like a lot of cool Microsoft tech, they didn't invest in it for the long run and just killed it along with EXPO, the craigslist competitor.

I am sure there were other things like this kicking around at the time as people have been discussing. I worked on improving the performance of the terraserver project a little bit. I was part of the team that designed and wrote sql server 7.0 (a db rewrite from scratch of the purchased sybase, including parser, optimizer, QE, storage engine, but keeping the stored procedure programming language that was from sybase). I believe this was 1998.

We all knew it was a pretty impressive combination of maps, jpeg image, software and web browsers. Microsoft definitely had an idea that they didn't want to make the web too great, because that could lessen the importance of windows. As others say, Microsoft didn't know what they had with ajax.

I remember this was one of the first, early, high query examples we had, other than tpcc and tpcd. There was a stored procedure call in the application that accidentally lead to basically a lock that was held, blocking everything else. That was a really hard problem to figure out.

I used terraserver to manually download the entire tile set for my town (right click, save as). Then I laboriously used Photoshop 5 to align and remove the watermarks. It took an entire week but by the end I had a clean ~150 MB jpg that I could scroll around in as needed.

We were writing web based GIS in 1996, using ESRI's MapObjects IMS. Tiling, projections, yes these were solved problems by then using available commercial software.

We were storing data in Oracle or DB/2 with ESRI's SDE sitting on top, because at that time no mainstream RDBMS had spatial types.

I don't recall if SDE supported SQL Server back then. That aside, you could have built TerraServer with COTS tech in 1997.

TerraServer was impressive due to being publically available. Nothing we built at the time was available to the public as far as I remember. But it existed.

And that's just web based stuff, which really is the easy part. Large scale GIS had been around for decades prior.

They're not the only ones apparently:

http://www.prnewswire.com/news-releases/origination-of-googl...

They invented Google Earth, yes, but looking at satellite images is mostly a novelty unless you're searching for concentration camps in North Korea or nuclear plants in Israel.

Google Maps is a much more generally useful product. And Microsoft was well ahead of Google, they sold numerous offline mapping products.

How I wish MSFT executives were more competitive and less dismissive of non-OS related technologies back in the day. I am really glad that things are quite different today and they are -atleast- competing with other tech companies if not leading them.

I wish they were more open to let the whole world use the standards they set. MS was an amazing platform company they set some of the greatest standards we have come to known. It is sad to see those standards hit the dust in the face of more inviting technologies.

AFAIK, Microsoft used public (US Government) imagery. Google bought Keyhole and that enabled them to own the sensors which gave them more power in terms of capturing content.

If you're interested in developing virtual globe applications, take a look at http://cesiumjs.org

Whenever I read stories like this about Microsoft I have to chuckle about all the similarities to Aperture-Science from Portal 2...

It's been a while since I played the series, but what similarities exist there?

Developing all kind of cool tech, but never releasing it?

I remember that good old around year 2000 Microsoft web design playing with them fonts and such. Ah, good memories!

> Microsoft, the corporation, didn't seem to care very little about...

Very much.

Microsoft Invented Google Chrome in the 90s, Then Blew It.. Eleven times.

Microsoft's focus in the 90s was to kill Netscape, they never had a grand vision for the internet.

Wow, I had totally forgotten about Terra server. I remember this now.

SQL Server Rocks!

Similarly, The first pc user interface was created in Xerox Labs, But Apple was the first realizing its importance.

I use Google Earth almost daily, and they've never made a penny off of me. It doesn't even have in-app ads. I guess these days "Totally Blew It" means "Wisely decided not to spend the next two decades hemorrhaging money on the thing."

So you think Google has been needlessly hemorrhaging money for two decades? Seems like an odd statement.

Clearly Google sees the value in it. Even if it is simply brand identity, or "vendor lockin" (for lack of a better description). Clearly there is something that is very meaningful to Google to do it.

I bet there are a ton of companies that would love to have their maps product as the "maps app" of choice by the masses. Likewise, i'm sure Microsoft would love if i had to login to bing every time i used Navigation, compared to me never using bing for anything ever. So yea, i don't think "Totally Blew It" is that far off the mark.

To list just one value: Google maps is a very strong reason why Android device manufacturers feel like they need to go with a Google-service model and follow Google's requirements, rather than do it Amazon Fire style. This alone is worth billions in search revenue. (And Google was smart enough to let it take a decade for this value to materialize.)

No, Google Earth has.

Charging for the Google Maps API calls is probably where they make money off of it. Lots of apps use it.

And prolific users pay millions/yr for it.

Reply on news.ycombinator.com