I wish people would stop blogging on Google plus. Period.
It's the worst reading experience ever on the iPhone - nearly impossible as the text blurs, and you have to drag the article around to see it all and then wait for the blur to go away.
Does the iPhone not open Google Plus posts in the Google Plus app? Or do you not have it installed? The reason your experience is awful is because Google doesn't care to make the G+ mobile web experience perfect since their app is the way to go.
On Android, when I click one of these links it opens the app and the reading experience is very pleasant. Maybe I've just been spoiled, but I find the experience to be pretty good.
Edit: Also, I think it's kind of silly that you wish people would stop publicly posting on Google Plus due to your phone experience being subpar. I think the problem lies much more with you and your phone than it does with Google Plus as a writing platform.
Should you need a third party app installed for a good reading experience? I don't use Google+ and don't want the app on iOS, even if it was launched when I click a Google+ link (which, fwiw, you are correct: it currently does not).
(On the other hand, Tumblr looks great on mobile, and it doesn't need an app. Facebook, too, is at least readable on their mobile site.)
Does anyone actually use Tumblr for original content, or for that matter, text? 99% or more of the Tumblrs I've seen are just people recirculating other Tumblr photos.
The current Facebook mobile site is almost 1:1 with their app (on Android, anyway). I wouldn't surprised to know the app is one giant WebView with a few system hooks sprinkled in.
I never understood why websites created apps of their websites. The browser is ubiquitous. Use it. I don't want to install seven different apps for my seven favourite sites. I don't see why this is acceptable on a mobile platform any more than it would be on a desktop platform.
Let me get this straight: You think it's more reasonable for Apple to have to build in special support so requests for webpages on plus.google.com go to the G+ app instead of Safari than for Google to just make their site work on mobile platforms like everybody else on the planet?
Well, not exactly. But that is the way that Android works, and it's pretty fantastic. When an app takes an action, it launches an Intent that notifies the OS of the details of that action. The appropriate application is launched (or the user is presented with options, in the case where there are two or more applications that know how to handle that Intent. In this case, it's both the Browser and Google+).
I was mostly asking if iOS had similar functionality, but it appears not.
Yes, but to then go on and say Google shouldn't worry about whether their site is even legible because, hey, it's the user's fault for choosing a phone that isn't based on intents? That's just silly.
I didn't mean to say that they shouldn't worry about it. I more meant that it's probably lower priority, and for an understandable reason. I'm sure they want to fix the web page and I'm sure they will - but the iPhone Safari market share is probably very, very small when looking at the Google+ user stats.
Also, I thought it relevant to mention here that you can indeed read the post very clearly on the iPhone - but a seeming lack of this feature (Intents) in iOS makes this awkward.
I've never developed for iOS, but on my iPhone when I go to okcupid.com it somehow 'redirects' to the okcupid app on my phone. So it seems there is some capability for this built in that website developers can take advantage of.
There is: Apps can register to handle certain URL schemes, so all you need to do is put in a redirect to (for example) "okcupid://profile/109" and, since Safari doesn't handle URLs with the okcupid scheme, the OKCupid app will be asked to open the URL instead. But again, this would require the plus.google.com team to actually implement it.
It looks like they basically ask whether you want to use the site or the app the first time you visit on a device, and if you choose the app, it sends you to the App Store and sets a localstorage setting that triggers a redirect on subsequent visits. (Edit: OK, it's a little bit more sophisticated than that. But I think that's not too far off.)
Mobile Safari does support apps registering custom URL handlers, so google could rewrite the links as gplus:// and they could open in the G+ app. Unfortunately I don't think this behaves nicely if you don't have the content handling app installed since it just errors.
That's correct. If you don't have the app in question installed, you get a nasty "Page could not be found" error. There are hacky and unstable JS tricks you can pull with setTimeout to send the user to the app store if they don't have the app installed, but I haven't heard of any reliable way to deal with this (pretty glaring) failure case.
Even just the ability to suppress the error message would be more or less fine.
How about loading the gplus:// link in a hidden iframe. Then continue to load the page in the browser. If they have the app installed, it will redirect, otherwise the page will just load in the browser.
So just load up the page (on iOS user-agents) with a big red button up top which allows the user to self-redirect, or default a preference on the G+ profile (if they're logged in).
Reasonable or not, that is the expectation Apple has encouraged iPhone users to hold - every website should have its own app in a way reminiscent of AOL keywords.
Expectation for who? It has never occurred to me to install an app to view a single website. Maybe this is something you do, and therefore think that other people also do.
But you're mistaken. I browse the web using a web browser. Even on my iPhone.
Since I don't own an iPhone, your assumption is incorrect. My spouse and many other non-technical people I know do own iPhones. And it is very common for them to use apps rather than their browser to access web content.
I call shenanigans. Practically every "content" site out there has its own app, and loudly proclaims its existence whenever you visit its site with a compatible device. It's simply not possible that "it never occurred" to you. You can obviously choose not to install a Google+ app if you like. But I don't believe for a minute that you are surprised at the need for it.
That's how Android Intents work, Apps can register to handle domains.
In my opinion, it's a far more elegant solution than encouraging developers to register pseudo-protocols to use in URLs that also result in very awkward interactions if the app's mobile site tries to redirect to the app and it is not installed. (Then again, this gives users a "choice" (!) of whether to open links or Intents in the mobile app or native app.)
Does the iPhone not open Google Plus posts in the Google Plus app?
Nope
On Android, when I click one of these links it opens the app and the reading experience is very pleasant. Maybe I've just been spoiled, but I find the experience to be pretty good.
This doesn't happen for me. I have a Galaxy Nexus with ICS, and I have Google+ installed and linked to my account, and this doesn't work - it opens the post in the browser. Is there some magic fairy dust that needs to be sprinkled into my microusb port for posts to be redirected to the Google+ app?
What's much worse, though, is that Google+ posts actually reliably crash the ICS chrome browser - as in 100% of the time. Whenever anyone posts anything on Google+, I have to open it in Firefox to read it.
Clearly Google needs to work on this.
Edit: Also, I think it's kind of silly that you wish people would stop publicly posting on Google Plus due to your phone experience being subpar. I think the problem lies much more with you and your phone than it does with Google Plus as a writing platform.
This isn't really that different than putting the post behind a paywall or some other inconvenient barrier to your audience. It might be a nice writing platform, but I think it's a pretty lousy reading platform.
> I think the problem lies much more with you and your phone than it does with Google Plus as a writing platform.
The problem definitely lies with Google Plus as a web app. Whatever they're doing, it's just not a very good experience on a mobile device. Maybe on Android they redirect Google+ pages to their native app, but they don't even offer a way to do that for other platforms even if you have the native app installed (even a simple link on the page that opens it in the native app would be fine). It seems obvious that they're actively trying to promote their own platform over all other reading experiences, which is a radical departure from the ideals of the web that Google has historically cared about.
I don't think a conspiracy is obvious at all. I think this is just a common Google problem; when you hire A+ level talent they don't want to spend their time adjusting css for every platform. That's not an interesting problem to them; self-driving cars is interesting.
This is why I think every company should hire B and even a few C+ level talent. Someone has to do the boring details.
It's not a simple matter of CSS. I have no idea how, but Google+ on MobileSafari periodically dumps the entire tile cache and forces it to re-render (which is why the text goes all blurry). I've never seen that on any other website, ever.
I think this is just a common Google problem; when you hire A+
level talent they don't want to spend their time adjusting css
for every platform.
Maybe I'm wrong, but I thought that the criteria for defining A+ talent is people that (in short): Defines correct specs for a given project; does the jobs to the specs; Works fast; Makes the code as maintainable as the situation may require, according to change and growth prospects; Does not waste time in unnecessary or fancy stuff; Communicates well with the team;
In this case, Google plus as a blog, obviously all the platforms are targets. So the CSS should be adjusted for all the platforms, and that must be part of the specs. If you have A+ talent executing, either it will be done, or they haven't implemented that yet according to their priorities, or there is something more unknown to us.
I'd be surprised if you could easily go from an Mobile Safari (or another iPhone browser) to the Google+ app when you click on a Google+ post. The way that works on Android is via Intents and, AFAIK, iOS doesn't have anything like that.
The blurring is a bug in Safari that is sometimes (I haven't investigated the exact cause) triggered when you scroll content below an element with position: fixed
In the case of Google+ that's the navigation bar.
Strangely, I could not trigger the issue when I was reading this post on my iPad. Now either Google has figured out the cause and found a workaround or the stars have somehow aligned correctly for it not to happen this time.
If you double tap on the column of text to zoom in and then scroll past the fixed element it will trigger it in this case. It happens to me on G+ and Techcrunch a lot. (iPad)
What's especially interesting is that your iphone layout is actually better than G+ on a desktop webbrowser, where the lines are insanely long and do not wrap when the window width is reduced.
Agreed. I actually find it one of the nicer reading experiences on the iPhone considering so many other sites don't wrap properly and make me scroll sideways to read each sentence.
I also don't get any of the blur mentioned when scrolling.
Could you post or email me a screenshot of this? I don't have my iPhone handy to reproduce it but I'll point the right people to this if I can explain it well enough. My handle at google.com.
I tried to take the screenshot, but when it's doing the blurry-loading thing it doesn't let me take the screen shot, and when it does take it, it's right after the page loads. So I guess it's "freezing" while it's trying to adjust the page.
While trying to do this, however, it just crashed Safari twice. I tried the same on my wife's 3GS, crashed Safari as well trying to read the Google Plus post.
Unfortunately I don't think there's anything we can do about Safari crashing while taking a screenshot at an inopportune time. Perhaps we can try to find someone on HN who works on at Apple?
I read the article just fine on my iPhone. But when I view a G+ entry on my computer I get a huge "Suspended" notification because I used a slash in my name when I registered.
> It's the worst reading experience ever on the iPhone
Yup, but it's probably the best writing experience for most bloggers, and also people who don't have blogs. Even if you have one with a decent readership, the engagement you get from Google+ is far higher than anything you can get when people need to use RSS and a reader to follow you.
Yeah, but patio11 would have a field day on streeteasy writing a popular blog post and giving Google all the the link juice. Unless Search Your World means G+ optimization is more important than SEO now?
And there you've highlighted the issue with Google+. For some reason, people are so determined to see it work that they attempt to shoehorn it to be something that it isn't. Google+ isn't a blogging platform and shouldn't be treated as such. Google themselves own a blogging platform.
I've repeatedly considered building a HN Google+ scraper that grabs the text from Google Plus stuff and throws it onto a more readable cache that won't be blocked by work filters as social networking.
However I'm a little worried about stealing what may be Google's content.
What causes that text blurring on iOS devices? I've encountered it on TechCrunch as well.
By the way, Hacker News is fairly unreadable on an iPhone because the text is so small text width so great. I haven't found a mobile interface that allows me to log in and comment.
mentioned in an earlier Hacker News thread, and it works well. I knew the submitted article was from plus.google.com and was able to anticipate its content accordingly.
As much as I'm with you on this. This is not the place to discuss over HN's features or GooglePlus experience over hand held devices. Sometimes I wish we could move these discussions to some other place. I'm here to read and discuss about the article. But I'm welcomed with tens of comments unrelated to it. This is annoying.
Protip: hovering over the link will display the full URL in the status bar. You should really spend more than a few milliseconds before deciding to get excited about something.
The Public Suffix List (publicsuffix.org) might be useful for this, although I doubt plus.google.com is listed in there (no same-origin security considerations). It's still the only systematic approach I know of that even remotely addresses the problem you're describing.
> Protip: hovering over the link will display the full URL in the status bar. You should really spend more than a few milliseconds before deciding to get excited about something.
You're assuming he's using a device with a mouse. And why not add a simple feature that makes usability better on all platforms? Special case google plus here, just because this is such a common complaint. I've wanted it for a while, too.
You can also trigger the status bar prompt via the tab key, link search, and probably assistive technologies have their own approach.
I'm amazed that offering up a potential general solution that already exists, or suggesting someone measures their response before jumping at headlines, would result in downvotes here.
> I'm amazed that offering up a potential general solution that already exists, or suggesting someone measures their response before jumping at headlines, would result in downvotes here.
I think it's tone and missing the point. Starting a response with protip is a bit condescending, and the point isn't actually about not being able to use the UI to get the desired information, but about how the usability would be improved (for some users at least) by changing how information is presented.
When those prices were announced, I was really expecting that they were just for show, and really everybody would end up paying way less than list price. It's bizarre that this isn't happening. Making a real and profitable business out of providing a geo API is great. Pricing it out of just about everyone's range and driving the potential customers away, less so. I can come up with a lot of theories, but none that is really believable:
1. So many companies really paying nearly list price that it's worth losing the other business.
2. Using different pricing schemes on different customers to figure out what the pain threshold really is, to maximize profit.
3. The sales people have a bad set of incentives, and are too strongly encouraged to try to make sales with low discounts.
The most telling part of this experience was the discovery of how good open source mapping has become lately.
While it's the money that's forcing developers to look for alternative mapping solutions, it's also unleashing a wave of interest in open source maps. Those will undoubtedly get better as time passes.
I do believe Google will look back on the decision to start charging for their map products at the levels chosen as a bad one for themselves, but a great one for open source.
Not to sound like a downer, because I really appreciate open source mapping, but that's the part where I think the costs, long term, will swing back.
Google maps is deceptively simple, in that the javascript library and tiling is a relatively solved problem. Google has the best datasets, and the best large-data know-how in the business, so outsourcing things like an up-to-date transit location and route database is built into that seemingly large cost.
What if you want to offer transit directions from your location to a rental?
Again, sorry if this sounds like I'm poo-poo'ing your effort; that's not my intention. I think there's another side to this story, though.
EDIT: Forgot to mention my own personal white whale. Google's geocoder is far from perfect, but it's the best I've seen. I worked for years with them, and none of the open source geocoders (or expensive ESRI products) could deal with addresses like Google's. Geocoding is like web WYSIWYG editors; they all suck, but some suck less than others. And everyone thinks they can do it better.
I'm a big fan of OpenStreetMap in theory as well, but once Google launched vector maps with really solid offline caching for their Android app it made it a lot harder for me to consider switching. I know OSM will let you cache tiles for offline viewing, but the level of caching you can do with vector data is a whole different level.
OsmAnd nightly build (0.7.0 branch) is GPL'd, with offline vector OSM maps and it's mind-blowing. Navigation works, with voice. There is now a native rendering option on some devices, it's pretty spiffy. It can also record GPX tracks, with optional background service. The vector map of Croatia is 20 MB compressed, with a huge amount of POI data. Daily, an average of 50000 points are added to the map of Croatia. I've actually stopped using Google Maps, OSMAnd is the best map app and it's completely offline.
I did a test with different Geocoders (Google, MapQuest, Bing, Yahoo, Nokia). Google was lightyears ahead of the rest in my test. It was for the Netherlands, though.
The netherlands is a different beast when it comes to Geocoding since each postcode only corresponds with around 10 buildings all very close to each other. Most other systems cover larger delivery areas. My complete amateur comment is that you need different algorithms for different countries.
I frequently have to deliver the bad news to bewildered drivers that the federal building they are looking for is in the middle of Seattle, not on a suburban street Kirkland which just happens to have some similar street numbers - they are always using Google maps directions...
Yes, but the data can be pretty bad, and software to access that data could use some improvement. For geocoding, there is only Nominatim which is a major PITA to get setup. I am still trying to get a full planet import and it has been 30 days on an 8 core 16gb machine with 6 drives in RAID10. Without 64gb of ram and some SSD's, it's going to take a month. Additionally, the results can really really suck. For example, zipcodes are often returned as "97006:97015", a range, rather than as a point. On top of that, sometimes you get the range delimited by a semicolon rather than a colon, etc. The price you pay for user inputted data - there's lots of it, but it's not all quality. Still though, better than paying an arm and a leg!
But they mention that in the text. Sure it's not something that happens now, but since now they generate the maps, they can add them or anything else they want to add to the maps.
I've been using Streeteasy.com a lot over the past few weeks. Sometimes I get Google.com maps, and sometimes I get the Open ones. I always think that the Google ones look and function much better. I hope the open ones improve, but it's not there yet.
Google charging heavily for maps is a good thing, in a way - more alternatives will come up, which are (hopefully) cheaper, and not locked to one company's whims
The take away I got is that Google maps is a very good way to implement a map tile on your pages if you are a startup, one less thing to worry about, but if you succeed, be prepared to "optimize" the cost of that portion later. The per-unit cost is a steep slope, once you start to get volume.
Fair enough. No way would I try to make my own map at the beginning of a project.
Does anyone have any pointers to tutorials on how to get started building OSM maps? For example, if I have my own "map" of the US (a simple graphic image), how do I go about connecting that up to state boundaries and plotting lat/long on it, and then breaking it all up into tiles? I feel like that part of the puzzle is always missing from blog posts like these (though the post is helpful for the bigger infrastructure picture.)
While notaddicted's reply can also be useful to you, I think the part of the puzzle that you are missing is "georeferencing". From your example, with a raster (bitmap, image) map of the US: the part where you start connecting that to lat/lon is georeferencing. Essentially, you want to tie a spatial reference to your image, so that the raster's state boundaries will correspond to state boundaries on a "real" map. You can think of it as pinning an image you know nothing about to a map you do know something about, and then copying the information from known to unknown.
There are plenty of ways to georeference an image. The easiest would be to use something like QGIS or GRASS (both opensource GIS software). You will need existing spatial data to base your (as of now, non-spatial) map on: a shapefile of the state boundaries would work great.
Once you've wrapped your head around georeferencing and the idea of spatial vs nonspatial data, you will probably get more out of the mapsfromscratch tutorial.
This isn't 100% what you're asking about, but http://mapsfromscratch.com/ has instructions for how to render and server map tiles, where you basically control the whole stack.
Could somebody explain to me how google charges for maps? I'm using it for http://lanmarks.com/ -- and it's not like I'm throwing an API key at them with the javascript include...
Start using an API key - you can then use their API console to track your usage, and you'll actually get notifications of overages, impending shutoff etc. Much better than flying blind.
Fortunately Google Maps' geocoding API limits still appear to be based on IP address. So if you do your API requests via client-side AJAX, you should be fine. That's because it's unlikely any individual user will make more than 25,000 requests in a day.
It's a very overview of the current open source map technologies out there. People may be talking a lot about the price, but my feeling is that the next batch of map servers etc are catering to a major set of folks who wish to diverge from the stock standard Google Maps look and functionality.
The technology is simplifying the whole process of generating and rendering maps quite a bit - comes with it's trade-offs of course. For those requiring a step or two above a simple basic map yet wish to have more customization over what's in the map tile, it's perfect - processing times are getting less, and the technology is becoming less arcane to understand
This might just be a matter of taste or just what I am used to, but I have yet to see another online map, that looks as good as Google Maps. Do others have a different option about this?
Btw, when Google announced it's ridiculous new pricing scheme for Google Maps, I though maybe they actually just wanted to force people to become Google Maps Premier costumers. I mean, no way they really except people to pay that much. So I am really curious if their Premier program is actually less expensive. Did you actually contact Google Sales about it?
Thanks for all the excellent pointers, some of which I knew but most I did not. I'm happy things turned out so well for you.
We recently had to navigate the Google Maps licensing obstacle course at $work. I tried getting quotes from Bing, but it was just as hard to get a number since we could only go through resellers. In the end we negotiated down a bit and reduced our map usage. But the next go around I'm seriously considering taking some time to investigate doing the maps ourselves. TileMill looks gorgeous!
This a great overview of the open source mapping options out there. I just posted some notes the other day on how MapBox pricing works for hosting custom map tiles: http://ds.io/A12Drf and a comparison of overage fees for MapBox premium accounts vs Google premier.
It sounds like it was going to be around $200-300,000 a year after a discount was applied from the google sales staff. With the other numbers bandied about in the text it sounds like that represents a 50% discount based on their current and projected usage.
It can't be too cheap to serve that many views reliably and quickly. OTOH, they are just NYC and they were already doing some things (like geocoding) themselves, while the google maps price included features they would never use (even implicitly, eg the cost of acquiring street data in other countries has to be figured into google's cost structure).
Sounds like you're on the current nightly. The bug was introduced this morning by Chrome, and affects a large class of webkit-transform CSS. Like other Chrome bugs, I expect it to soon be replaced by a new bug.
Another option, though still expensive, is to use the Google Earth Enterprise tools. They basically let you build your own private Google Maps though this too runs into some problems.
1.) Imagery, 2.) Cost for GEE ($10k I think), and the cost of maintaining your own servers to host it on.
I think it's still cheaper than going over the 20,000 free, but not optimal either.
Anyway, good for google for charging for something. I don't see why it should be free. But it was a blunder to charge this much from the start. It seems like they want to get rid of customers.
By shapes do you mean borders? Or do you mean buildings and things... The borders come from http://www.openstreetmap.org/ which is editable/exportable in many formats.
Interesting. After years of paying through the nose for a Google search appliance that actually wasn't very good, my company decided to replace it with Solr, and it was a really fun project. We hired the Solr project's main committer to consult with us and in the process submitted patches for bugs that we ran into.
The whole thing cost less than one year of the GSA license fee, we can search all of our content (the GSA had a hard limit on the number of documents it would index), and the search results are way better.
Charging is good since it keeps the product around, but Google's prices seem ramp up suddenly from free to way too high suddenly. First Appspot and now this, looks like their pricing is too out of line for small to medium companies. They need a flatter pricing curve.
I don't think they really care. In the current landscape, Google Maps is a competitive advantage for them and the less people using it the better. They leverage it across the Android ecosystem, their new social network, etc.. and it creates more work for their competitors to play catch up.
"TL;DR: We at StreetEasy decided to build our own maps using, among other tools, OpenStreetMap, TileMill, MapBox and Leaflet, instead of paying hundreds of thousands of dollars per year to Google. And yes, the money pushed us into doing it, but we're happier with the result because we now control the contents of our maps."
>Good bye, Google Maps… thanks for all the fish (google.com)
is simply misleading.