Old Maps vs. New Maps
vore.cc
vore.cc
This is about quality of search results.
Apple has had search in its products for a long time. It has search in its documentation, in Spotlight, in Mail.app.
All of this search technology is crap. Spotlight is fast, but it's fast crap.
I was in the bay area this week, and searched for San Francisco International. The data is not lacking for this search result. The roads are there in the Maps app database.
Anyone want to guess the result of my search?
Apple partisans want to believe that this is a data problem. If it were a data problem, then it'd only be a matter of time before Maps.app were a solved problem.
Unfortunately, it's not. It's a search problem. I know absolutely nothing about the technical details about how to implement a decent search, but I do know this: Apple has never, not once, delivered a product that gave decent search results.
So good luck, Apple. Hopefully this problem can be solved by throwing money at it. Something tells me that it can't, though.
2. I just went into Maps and searched for "San Francisco International". There were two hits, and the second hit was the airport. The other hit was for a business in SF with "International" as the first word in its name. These do not seem like completely awful search results to me.
3. Why are you assuming (and yes, it's a complete assumption which is free of evidence) that "Apple partisans want to believe" (you can read minds now too?) that it's only a data problem? I'm an Apple partisan, and I worked on the Maps team, specifically on quality of search results. I can tell you that what I believe is that it is both a problem of needing to refine the database, and also a problem of needing to continue to improve the search capability. Does this change your assumptions?
> 1. You were searching for an airport, left out the word
> "airport", which is the most important word, and it's
> Apple's fault?
Whether it is able achieve that feat because it knows that I have been to SFO within the last month, or knows that I have emails in my inbox that pertain to the airport, or because more users click through to SFO than other hits for that search phrase or because they use black magic is beside the point: Google's search results are simply better.
When most people refer to a place, they refer to it by it's common name. You fly to JFK, Heathrow, Charles de Gaule, Gatwick, Luton, O'Hare, etc. There is no need to add "airport" to something which so obviously is one in the context.
The familiarity of common names is why wikipedia uses them for article names: http://en.wikipedia.org/wiki/Wikipedia:Article_titles#Common...
SFO, of course, is. Which makes this particular search MUCH more ambiguous unless you include the word "airport".
Apple apologists are actually making Apple lazy and entitled. Any sub standard stuff needs to be pointed out. Not hand-waved away.
2. I'm no longer in the area, but at the time I made the search I got something like 5-10 results, none of which was clearly the airport.
3. That's my explanation for why nobody is talking about what I feel is the giant elephant sitting in the corner. Maybe I'm wrong. If Apple partisans actually think that the search is fine, that might be a better explanation.
Google maps is astonishingly good. The data is accurate, and it always maximizes the amount of information at any zoom level. The depth of quality isn't about features, it's about mastery of a complex dataset.
It's not just maps. Google Now vs Siri. Ask yourself which will evolve faster? Where will these products be in 2 years? 5 years? I predict an increasing gap.
Google has deep experience developing algorithms and managing data. Apple will have trouble matching this.
I love Apple. I used to work there. But they need to wake up a little.
And I don't see there being an increasing gap going forward. There simply isn't an unmanageably high turnover of street names/locations in most cities. And Yelp already handles turnover of POI data quite well.
I'm in Tokyo right now, and I use my Google maps constantly, but I can imagine how much easier they could make it. I can already read Japanese web pages via Google translate, soon I'm sure the integration will be even better so I dont need to copy and paste addresses between. And Google Now would integrate all that with my calendar and contacts in my address book. If you've ever tried to read Japanese or interpret Tokyo addresses before Google's products, you'd understand the breakthrough here.
The needs are way beyond a map dataset. And we'll see how easy it is for Apple to just go out and license a few of those existing datasets you mention to get a product a deep as Google's.
EDIT: Integration is what -I'M- talking about because that will make the long term difference in a successful mobile product. Which is the real underlying question. And I dont see how Yelp's data can help me navigate Shibuya station (missing from Apple maps, very detailed on Google maps. Good thing as I got lost within the station a few days ago).
The interesting question is whether Apple is even going to bother to license data. They could well rely on Yelp's geo expansion to do the work for them. God knows what will happen if Yelp goes out of business though.
Examples from last week:
- I stayed at the Radisson Blu Alcron hotel in Prague. Searching for 'radisson blu prag' (in German) with Google Maps immediately lead to a correct result. Apple Maps delivers no result. Using 'Prague' instead of 'Prag' didn't help either.
- I had an appointed at Skyguide's Zurich HQ (Skyguide is the Swiss air navigation service provider). Google Maps again showed the correct location while Apple Maps showed for some reason only results named 'Skyline'.
- I often used Google Maps for public transport routing. Apple Maps of course does not support this feature.
- On Apple Maps, the street I live on goes through my house. In Google Maps, it's where it's supposed to be, i.e., in front of my house.
- I went hiking in the mountains last weekend and the quality of Apple Maps' aerial and satellite images was simply awful in comparison to Google Maps.
- Etc.
Apple has four problems: (1) inaccurate street data coming from TomTom, (2) missing data for parks/green/open spaces, (3) missing internal data e.g. shopping centres, campuses and (4) missing POI data from Yelp.
The key difference between Google and Apple is that Google has initiated partnerships with key data providers in each country e.g. Whereis in Australia whilst Apple is relying mainly on two: TomTom and Yelp.
By the way couldn't you just have searched for SFO? That works even when looking at another part of the world. And it is much less typing.
This, for me, captures exactly why Apple deserve the backlash they've seen from this whole Maps thing (and why Tim Cook's apology was much needed). Their special events are always filled with the same bullshit: 'Amazing'. 'Magical'. 'Incredible'. 'Unbelievable'. Ok, Apple have very talented people working on their products, and they're extremely proud of what they do (they should be), so I get that they can't help but roll out the same marketing spin for everything they release.
This of course backfires when we get these products in our hands and they're not everything we expected them to be. The biggest change that needs to happen at Apple now is that they need to breed humility into their culture, and show it more in public.
The last hint of humility I saw from them in recent years was the word 'beta' slapped onto Siri at its original announcement, and yet Forstall still put on a flawless presentation that showed Siri as far more capable and responsive than it's been in the wild.
The most recent of course, was Tim Cook's Maps apology. Steve Jobs would never have written that letter, and I'm glad that there is now someone running Apple that knows when an apology is due, especially one that recommends alternative products in the meantime, rather than shovelling shit on them in an Antennagate-style "we suck, but look, everyone else sucks more, na na na na na".
It takes a certain kind of person to get a company like Apple to the top. It takes an entirely different kind of person to keep them there. I can only hope that this is first step in Apple forging a more honest, humble, real relationship with it's consumers, because in time that will be the only thing that can secure their future.
Big eyeroll here. Returning a probabilistic match is a nefarious plot for revenue. OK...
That may be an anecdote, but that single data point is enough to make Apple maps utterly useless for Auckland.
Also, I don't agree with the assertion that Apple maps won't return far-fetched results. I searched for "PB Tech", a technology shop chain, and it took me to Mexico.
But I believe they will be there soon as they just launched in Australia and bought the NZ domain name. Yelp's strategy has been to partner with the local YellowPages equivalent.
So what you will see is a jump from bad data now to almost perfect data overnight.
I spot-checked on the Google API a few entries that the spreadsheet showed in the wrong country (Waldemar, Coldstream, Osceola, Snow Road Station, Belgrave, Winger, Heidelberg, Priceville, Vernon, Blackwater). I only got one result that was wrong (Vernon) with the default settings. With a country bias to Canada, which I hope would have been there for an API user from that country, that one is resolved to a sensible looking locality as well.
There clearly are some queries in that set for which Google returns the wrong answer, but I'd be the frequency of those problems is an order of magnitude lower than this post suggests. It'd be interesting to know what the reason for this discrepancy is, but without seeing the exact queries on the wire one can only speculate.
The only way to use Google's forward geocoder data is to make a REST request with your own API key using NSURLConnection yourself.
In iOS 4.x you can use MKReverseGeocoder which would return results from Google, but this class was rewritten in iOS 5 to be a thin layer on top of CLReverseGeocoder (using Apple's servers).
My iphone 4 had google maps on it as I drove to Miami beach, I actually missed my turn twice and I am not a bad driver, it's just that you need to merge very quickly in one particular spot twice (basically stay left or stay right). Google maps wasn't keeping up and even after using maps for a year or more or three, it still happened. It just was unfamiliar and the app didn't provide any additional help other than trying to flip through the directions very fast while driving which isn't safe as it is.
I finally got there, got my iphone 5 and was BLOWN AWAY at the maps app. At first I was annoyed coming out of the parking garage because I thought "you cant skip forward ahead in the directions??". It must be what everyone is annoyed at. Suddenly, I pulled out of the garage and magically it has 3d view, which rotates as you drive, and talks to you. It was an INCREDIBLE comparison because not 45 mins before I was missing turns, and on my way out of miami i could see every turn PERFECTLY, and not only that I didnt have to hold my phone not keeping my eyes on the road.
I understand the data/search is what most people are referring to but the point above is massive. It blows google maps away x10000. I played with the GPS the entire way home, especially the part where you go around the cloverleaf loops on the highway merge junctions, and watching the maps follow around with it.
As far as search, I have tried searching 'sushi' while somewhat zoomed in. whereas google would search a given radius if it had low results, apple maps stays with the low results unless you manually zoom out.
I have also seen it be somewhat picky if you don't specify a road correctly (in the jungle where my parents live it didn't recognize "CR" as county road, so I typed that manually). These problems are hardly more than just an annoyance, and the benefits far outweigh the gains imo
None of this blows away Google Maps. It blows away Apple's implementation of Google Maps, which was subpar. So, now Apple's implementation is on par with everyone else, but with subpar data.
Any time my friend hands me his droid phone to text while hes driving its just a mess.
I'll stick with iphone. Theres a reason it takes features longer to get there I think, its because they actually need to really make sure they 'fit in' in really smooth cohesive way. if they just accepted every single suggestion or feedback from customers it would be just as cluttered as droid
(It's not like the old Maps app was flawless. It's just that it usually worked and the new one usually doesn't. I'm in greater Washington DC, I'm sure your mileage will vary.)
The same can be said for Waze -- I often would convert a descriptive location (such as "San Francisco International" per another post) to a street address then copy and paste it into Waze. I use the conditional tense because I consider the new iOS Maps app to be superior to Waze for navigation, once you've found an address.
It seems to me Apple could easily cut a deal with a third-party-search-engine-that-isn't-Google to improve this simple flaw.
It's bound to fail.
What's interesting about this is that it beyond a doubt confirms that Apple is now playing catch up to Google (and Android).
I want to see more of that and less of the anecdotes. I’m under no illusion that this is any sort of conclusive test, but it’s a start. (This, for example, doesn’t test for POIs or more complicated and fuzzy search queries, which for me – anecdote alarm! – are the huge issue and both often unusable – while all the rest works like a charm for me, except for some small gripes I have with how maps are displayed.)
"tl;dr" is to save other people the effort.
This would be correct, however 'tl;dr: <summary>' is pretty common for providing a summary.