Hacker News new | comments | show | ask | jobs | submit login
Predicting Google Product Shutdowns (gwern.net)
261 points by kiba on May 4, 2013 | hide | past | web | favorite | 139 comments



I have some thoughts about the spikes on the death dates.

September: all of the interns go back to school. These people who exist on the fringes of the system manage to get a lot of work done, possibly because they are free of most of the overhead facing real employees. Once they leave, it's up to the FTEs to own whatever was created, and that doesn't always work. I wish I could have kept some of them and swapped them for some of the full-timers.

March/April: Annual bonus time? That's what it used to be, at least, and I say this as someone who quit in May, and that was no accident. Same thing: people leave, and that dooms whatever they left.


September: Googlers return from Burning Man with added appreciation of impermanence and the value of cleansing fire to clear the way for new creation.


The September one is a cute idea, but how many shutdown things immediately had a bunch of interns? For example, with Knol, there weren't any site updates or functionality changes for something at least a year before the final shutdown announcement went out, so how could that have had anything to do with interns? The spike in September is so huge I'm not sure interns explains it.

Maybe people are going on vacations (not necessarily Burning Man) and returning with renewed perspective and distance from things and are more willing to close down things. Not sure how one could test this...


It could also line up with Perf season.

I'm happy to report that I have successfully forgotten exactly when perf happens. There was one supposedly optional perf and one supposedly non-optional perf per year. I'm not sure exactly how that would affect cancellations of projects, though.

One thing I do remember seeing was a LOT of half-baked crap being foisted upon people (especially internal customers) because the end of quarter was approaching. Someone had to make their OKRs, and they didn't care how...


March/April: Annual bonus time?

I'm not sure if it's good or bad that I somehow never managed to work for an employer that offered any kind of bonus, ever. It granted a perspective of "this is what you get, and no more, unless you fight for it," which... again... can be good or bad depending on how you look at it.

I wonder how many 20- and 30-somethings are now or have recently been in a position where bonuses are reasonably common, and how that correlates with the size/type of company they're working for.


Funny, I have the opposite opinion. Without a bonus, it just seems like "this is the work you need to do, and no reason to make any effort to do more". With bonuses that scale with your contribution, I have far more incentive to put in effort above and beyond.


I agree, but it's pretty rare to have individual performance based bonuses unfortunately. It also makes people feel entitled, so they will never go above and beyond without the bonus in the future.


Why should they go above and beyond for free?


Bonuses are the norm in the finance industry. I worked at a financial research company for five years and got a bonus of roughly 10-20% of my salary each year, depending on that year's profits.


At some companies, signing bonuses are starting to become standard.


>September: all of the interns go back to school.

This isn't true. Internships aren't only during the summer for many schools. I'll be starting a four month internship in September as will thousands of my peers.


It's true that Google has interns year-round, but I'm willing to bet a sizable majority of them are over the summer.


Or it is twice a year company wide review and financial planning, not major product decisions driven by temporary employees.

I love how people say Google is awful because it is a corporate hell where engineers have no autonomy, and then claim that most products are launched and run by rogue indivuals.


> Gmail seems like a service for the ages; but once upon a time, so did Hotmail (even in 2012, 8 years after Gmail’s launch, still the #2 email service) - yet, RIP Hotmail (1996-2013).

That's silly. Hotmail didn't go away; it was rebranded. All the email and accounts were transparently migrated - you could even use both front-ends at the same time for a short period. The important part is, Microsoft still runs email and nobody lost data. Comparing it to the Reader shutdown (or a shutdown of Gmail in a similar fashion) is an apples-to-oranges comparison of the first degree.


>The important part is, Microsoft still runs email and nobody lost data.

Unless you fail to log in at least once every 270 days, at which point your entire account and everything in it is wiped.


That's probably a safety precaution but if it is like you describe it is also a risk. It could be a safety precaution to make sure that accounts do not eventually become compromised in the case of deceased person.

But if the account is deleted entirely that could indicate the address can be re-used and in that case any password reset mails and other important email might land in the new owner of the address' inbox.


Wasn't that true before the Hotmail -> "Outlook" transformation? I.e. a Microsoft free email policy that as jacquesm notes could have good reasons behind it.


It's always been that way. In fact, I'm almost certain the time was shorter in the early days.


> That's silly.

It is. The rebranding and new interface are a huge improvement and a sign Microsoft is taking there email service more seriously. I don't really get how someone could interpret it as the loss of a service.


In the near future someone may say: that's silly. Gmail didn't go away; it was rebranded. It is called Google+ now.


Except Google+ and Gmail are very different things. Hotmail and Outlook.com are exactly the same thing with a different UI.


For anyone else that is wondering which Google product lived for one day ("Days: Min: 1"):

http://en.wikipedia.org/wiki/Google_X


Great find. It must have been total chaos - good chaos - in those days.


Seems more like a Doodle (homage to OS X and its dock) then a product that lived for one day.

Although the March 15 2005 date was a bit off any OS X anniversary I knowm


I think this actually fits with what I think about Google overall. Search is great, everything else is just a mediocre tie in to search dollars.

Google is attracting more "nesters" and less "innovators" from what I see as an outsider, which is likely good for long term products but bad for everything else.

Glass is already a bad word (RE: glassholes), Plus is a wasteland, Voice is only popular due to branding, as is Calendar and Docs.

I've seen announcements of products that appear to be more functional than Voice/Cal/Docs on HN with better security and performance, but I think the likelihood of those taking over is probably low until Google gets out of those markets. The products are better, but the mindshare isn't there yet.


I don't know why people keep insisting that G+ is a waste land. If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

Personally, I get far more value from G+ than I do/ever have done from twitter, Facebook, MySpace, or any of the many social networks over the years. I have a use for it, which it fulfills better than any of the other services out there, and so for me (and the thousand or so people who I'm connected to at various levels through G+), it's far from a waste land.

G+ isn't for everyone, and that's fine. It is a product which is now highly polished, and provides exactly what I and many other people have wanted from a social network. The exact numbers of users who actually actively use it may not be clear, but if it was a waste land, Google wouldn't be integrating all their services so tightly into it.

Google is not a stupid company, and they're not about to "break" all of their services to drive adoption of a service which can't stand on its own. G+ is a good service, and is providing massive amounts of value to Google. Reader was not creating new value for them; wave wasn't getting adopted; Google had reasons why continuing to support those products was a losing bet.

They are integrating so many of their services into G+, that you can pretty much guarantee that it's a winner (for them, at least).

</rant>


> I don't know why people keep insisting that G+ is a waste land. If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

It's more that nobody I know uses it. All of my friends are on Facebook, and everyone is talking about Facebook. If I try finding my friends on Google+, it looks like a wasteland.

> They are integrating so many of their services into G+, that you can pretty much guarantee that it's a winner (for them, at least).

Why should I care if it's a winner for them? Google has a huge reach, and they are leveraging that reach to push Google+, but people are still uploading photos and updating their status on Facebook. Google+ may have many unique users, but their use is far more shallow than their use of Facebook.


>If you personally don't use it, or see a use for it, that's fine, but there are many people in the world who are not you.

I wholly agree, but G+ is barely better than tribe.net (for me and my peer group) when you compare it to what Facebook has in terms of content and users. There's also the whole "Nym" issue. I also can't understand why people are excited by giving Google MORE data, the "don't be evil" days are clearly over with.


They are integrating search so tightly with G+ that websites/authors have to get a G+ profile to make authorship data visible in search.

A fairer way would have been to use microdata alone to achieve the same thing. This would have been acceptable to other search engines too.


> G+ isn't for everyone, and that's fine. It is a product which is now highly polished, and provides exactly what I and many other people have wanted from a social network.

Is it possible yet to change the visibility of old posts? The absence of small privacy features like this—features that FB does have—always was very off-putting for me. When Google+ was launched they made a huge deal out of Circles, but it strikes me that since then Google has pushed Google+ more as an alternative to Twitter than as an alternative to Facebook (and I don't like Twitter much).

Anyway, I used Google+ so little that I deleted my account in a petty rage when they shut down Google Reader. It's actually nice to have one less social network to worry about.


> Is it possible yet to change the visibility of old posts? The absence of small privacy features like this—features that FB does have...

Are you sure you have thought this through? Allowing retro-active changes to visibility (unless it's only in a more restrictive direction) can be a privacy violation for the commenters, who may have commented based on the original visibility list. And allowing visibility changes towards a more restrictive set will confuse the people who could see the post till yesterday, might even have commented on it, and suddenly can't see it anymore.

Privacy decisions are not as simple as they might appear on a cursory glance, and overall I find that G+ takes privacy very seriously. Seen from the above light, the current policy of not allowing visibility changes, but allowing the author to delete the post, makes a lot more sense.


I disagree.

Allowing more people to see a post is already possible by adding people to the circles that you shared the post with.


Social is inevitably a winner take all market, so in this case the ignoramuses who never use G+ who declare it dead are probably right. Its not about great features, just ask FriendFeed.


Why do you believe that social is a winner take all market?


Metcalfe's Law.


I hope you're being facetious. While (for instance) Facebook's ubiquity will insulate it for many years, come what may, raw connection numbers aren't nearly as important as local connection numbers, and "local" can mean many things, including distance on an interest graph. For instance, I joined Twitter (and Google+, for that matter) because there were interesting conversations happening there and I wanted to take part. Those networks continue to sustain interest for me, so I contribute, and that brings interest to other people, etc.

There might not be a steady state (either becoming abandoned over time, growing to monstrous proportions of popularity, or changing character enough to drive old timers away, or some combination of the three), but all the evidence for the entire history of social groupings suggest that other social groupings will pop up to take the place of any that fail and to fill in niches no longer served. It's certainly not zero sum.

Metcalfe's law was also only ever meant to be an observation (and really just boils down to a description of any network, with "value" being some nebulous term), and certainly was never meant to describe humans' social connections, if only because they cannot sustain an arbitrary number of them. Applying it to social networks for predictive purposes is just sloppy sociology, which is about as low as you can get.


I wasn't being facetious, I just hadn't thought it through as thoroughly as you have. Good points.


You're nuts.

Google certainly has some mediocre "products," but they have many great products as well besides search.

Gmail, of course, essentially sets the bar for online mail, and is still leaps and bounds better than any of its competitors.

Google maps is very, very, good, vastly better than most of the competition (a point made starkly clear by Apple's recent maps C-F).

Google+ is, technically, absolutely excellent, much better than FB. If all your old high-school buddies are only on FB, it's a shame, but there's not much Google can do about that.

Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

...etc.

Google has its downsides (e.g. their online support sucks), and you may not like them for other reasons, but they're very, very, good at writing software...


> Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

For me, I feel like I have 2 hands when using iOS but 17 hands when using Android. It asked me to pick one of 3 power-savings modes for Wi-Fi when in sleep mode. I don't want to have to care about that (nor do I want to pick one of the three options without thinking, because then I'll not have an optimal configuration). I have enough things to tweak on my computer.


Regarding the power-savings mode selection issue, that is most definitely not a part of stock Android. Google releases Android without any OEM or carrier modifications. If the OEMs/carriers decide to do something, Google is (generally) not involved.


This fragmentation issue is always presented when a flaw like this is incorrectly attributed to Google, but somehow that has nothing to do with Android? That seems like a silly thing to say. Yes, iOS suffers some problems but so does Android, as a fragmented beast that is still stuck on jelly bean as the majority version (which IMO is pretty lacking). I do look forward to Android's future though.


It's really pretty easy to avoid this by buying a Nexus branded device, though.


Agreed, but explain that to a regular consumer in the Verizon store.


I don't think anyone is suggesting that we explain anything to regular consumers in Verizon stores. The point is that an HN member should be able to understand and appreciate the fact that stock Android is very different from what most Android phones have, and that those who want stock Android should get a Nexus.


It's been part of stock Android at least since 2.0. But it's also hidden 4 layers deep within Settings (Wifi / Menu / Advanced / Keep Wifi On During Sleep / ...). Complaining about being "asked to pick" a setting when it's hidden that well is pretty disingenuous.


The setting may be part of stock Android, but being asked to pick one of the options is most definitely not part of stock Android.


I'm not making this up. It was a Sony tablet with 3.x (much later there was an update for ICS). It asked me to choose a value for this during initial setup. It didn't even have a recommended setting iirc.

In any case, it's just one of many examples where I feel Android requires more maintenance than iOS. Which is perfectly great for those who want something to tweak and play with.


> In any case, it's just one of many examples where I feel Android requires more maintenance than iOS.

Once again, this is not part of stock Android, it's part of "stock Android + some modifications Sony made." There's a big difference.


That is true. Unfortunately, that's the only first-hand experience with Android I've had so far. It wasn't by choice. If I had to get an Android device, I would without a doubt try a stock Google one. I don't expect a huge difference, but it should be more pleasant.


It actually is a pretty big difference. I had a really bad first impression of Android from using early HTC stuff, but my current job is Android development so I figured I should check out the alternatives--and I still had my iPhone, so I could swap the SIM. The AT&T store sold me a Motorola Atrix HD, running either 4.0 or 4.1, and it was awful. It was ugly, Motoblur stuck a bunch of crap I didn't want everywhere...just bad. I took it back the day I got it and bought a Galaxy Nexus online, with 4.1 on it, and it was remarkably impressive; I've been using Android ever since.

I've seriously considered a Galaxy Note 2, but TouchWiz's existence offends me. :) If there was a decent stock-ish build that supported the pen I'd be all over that.


Yeah I get that, you might like Google maps more if say you cared about the actual maps.


That'd be Bing maps then - which I use pretty much exclusively these days.


Bing is great… unless you travel. It fails for many countries. Even for major cities in China it will only show the main roads.


> Android is vastly better than iOS (I use both, and using iOS always makes it feel like my hands are tied behind my back).

It's really in no sense "vastly better" the market pretty much bears this out.

From a developer/technical standpoint it can be a pain that you have less access to device features on iOS and the AppStore review process can be a nightmare. But developing on XCode isn't bad, as is the UIKit architecture. It's unfortunate that iOS isn't open source like Android, but I can see no way from a user or technical perspective that it's "vastly better".


> It's really in no sense "vastly better" the market pretty much bears this out.

Some markets. iOS is big in the US, Android's big in the EU. It's clearly network-effect.

Because of that, "the market" doesn't bear out anything regarding quality.


I'd like to see citations for that. But one of the reasons could be that low end smartphones and off contract phones are far more popular in Europe. That drives the market towards cheaper devices which tend to run Android.

In any case it was the word "vastly" I was objecting to. It's possible to subjectively think Android or iOS is better but I don't see how to can make a reasoned argument that Android is objectively vastly superior (aside from the fact that it's open source which could be considered critical).


Here's some data for Japan; Apple has about 1/3 of the smartphone market, almost all of the rest is Android. There's generally little difference in device capability or pricing between iphones and non-iphones in Japan, so it's not a low-end vs. high-end thing, and two out of the three major carriers heavily push the iphone, so it's certainly not marketing....

http://www.comscore.com/Insights/Press_Releases/2012/4/1_in_...

> I don't see how to can make a reasoned argument that Android is objectively vastly superior

It's an opinion... :]


> There's generally little difference in device capability or pricing between iphones and non-iphones in Japan

Actually Japan is probably one of the few places you could say Android phones are more capable as a lot of them have 1Seg tuners. :)

The iPhone plans in Japan also tend to me more expensive (partly due to LTE on the iPhone 5).

The situation is also slightly odd in Japan as the historically largest provider (docomo) doesn't distribute the iPhone. They are however rapidly losing customers possibly because of this:

http://japandailypress.com/docomo-bleeds-subscribers-as-ipho...


> Gmail, of course, essentially sets the bar for online mail, and is still leaps and bounds better than any of its competitors.

If Gmail keeps lowering the bar instead of raising it, that's not going to be true for much longer.


There about two features that matter in consumer mail: storage quota, multi-location/mobile availability (not a fixed single client machine), and spam.


Storage is storage, no worry catching up there. Multi-location availability is just IMAP. That leaves spam as the only hard problem, though admittedly it's a problem squarely in Google's wheelhouse.


The days of Google Maps being vastly better than Apple Maps are fast disappearing. It is about 95% parity for Australia and UK at least and any differences are being eliminated rapidly. And the touted advantages of Street View aren't there since the data is at least 4 years out of date in some countries.

The fact is that Maps aren't something you can dramatically keep improving. You reach the "everything is mapped" point and then it becomes a keeping everything up to date scenario.


I don't know about the rest of the world, but it's still a lot worse than Google Maps in Japan. I'm sure it will come in time, but it's a difficult problem and one that google has a better handle on with it's background in "big data".


> The days of Google Maps being vastly better than Apple Maps are fast disappearing. It is about 95% parity [...]

Not where I live (Tokyo), it isn't.

I have an iPad mini (got it for free!! :] ), which has a nice large display, and I'd really like to be able to use a mapping app on it when I'm at home, but using Apple Maps is painful compared to Google Maps (which I have on my phone). The presentation in Apple Maps is very pretty (well, as long as you don't turn on satellite view and notice the super low-resolution images...), but the data/integration just isn't there.

If you live in Cupertino, of course, Apple Maps probably has fewer issues...


I'm not quite sure that this is true. I've been noticing many more house addresses in captchas recently, indicating that Google is working to make sure that home addresses in Maps accurately correspond to those in the real world. While improvements of this nature are far less noticeable, they demonstrate that many of the remaining improvements to the product are being made 'under the hood'.


Good luck with that. BTW aren't you going to point out once again that for example in Australia all the data comes from Tele Atlas and Whereis?


Glass is piquing more interest than any consumer electronics device since the iPhone, Plus is thriving, and Docs is by far the best cloud office suite I've seen. I am by no means saying that Google is the best in every market they enter--some of their offerings are pathetic compared to the competition-- but your claims of across the board mediocrity in their non-search products is severely overblown.


The concept behind Glass is fantastic. For all the talk of "cyborgs" and "glassholes", it's clear that it struck a nerve. What we don't have enough information to say yet is if Glass is the new iPad or the new Newton.


Google+ doesn't appear to "thriving" to me. Here is one indicator: http://www.google.com/trends/explore#q=google%20%2B%2C%20goo... (flat line). For comparison, add "imgur" or "hotmail" to the input. Here is another where it looks a bit better: http://siteanalytics.compete.com/plus.google.com/. However, considering the amount of integration with their other products, I can't see it as anything more than a disappointment for Google.


That's looking at search queries though, right? Why would any one Google for g+, when there's a link to it on the top left corner of Google's search page?


As soon as I read "Plus is a wasteland", I knew you had no idea what you were talking about, and stopped reading.


Indeed. Communities has made G+ something I read multiple times a day, at the expense of Twitter


Google Maps/Earth, Android, I wouldn't call that mediocre and not innovative...


Maps/Earth are cool, but they've stagnated, Barely changed in the last few years.

Android is neat, but its not innovative either. The only reason I use one is because of the wide variety of HW available.


There's been a ton of new features on maps over the past few years, including indoor maps, indoor streetview, transit info and all kinds of new aerial views. I'm not sure what you'd want from them... after all, they are still a map service.

http://google-latlong.blogspot.com/


> Maps/Earth are cool, but they've stagnated, Barely changed in the last few years.

That's a feature, not a bug.

I actually like the way maps works right now, I would not want change just for its own sake.


If you think android is not innovative, then I guess google search isn't innovative either: search engine existed before google.


And yet, I have to resort to Garmin software and maps to do real trip planning (cycling, hiking, and even driving).


At least in my experience (in the US) the maps data on Google is much better for cycling routes, partially because users can go in and submit fixes, which I don't believe you can do for the Garmin data.


I skipped ahead to the "10 most risky and likely to be shutdown products", read it, then did a double take "wasn't I supposed to be reading about Google?" Never heard of any of those.

Interesting poems. I was certain Ozymandias would show up though.


That's because he's using search hits as a primary measure of how much services are being used, which (maybe) kind of works for consumer services but doesn't work at well at all for services used by developers (at the least, they should be judged on their own scale). Which is why you see Cloud Messaging popping up in that list - it's just not something many people talk about, but that doesn't mean it's not heavily used.


> Never heard of any of those.

Yes, and that'll be reflected in the hits data and be one reason they'll be rated so risky...

I love 'Ozymandias', but unfortunately it's become so overused it's cliche. So I picked poems that I was reasonably sure almost no reader would ever have seen (the exceptions: the _Julius Caesar_ quote, which I parodied; and the Basho haiku, which I gave my own rendering of) but were hopefully still good.


http://www.google.com/ig/directory

Ozumandius was never shut down, it was left up in ruins.


Interesting post, but oh my god, the styling on this site is hurting my brain.


I know I designed the styling mostly to appeal to myself, but serious question: what's wrong with black text on white background? That's basically how books have been printed for hundreds of years.


Well, there are several problems.

First, to address your point. Paper is not pure white, and ink is not pure black. Books are also not backlit. If you want to recreate the experience of reading a book, use a slightly off-white color and a dark colored ink (for example, a very dark blue or brown).

Second, most everything on the page is way too cramped.

Third, you need to constrain text width. It's a lot of work to scan 1000 pixels for each line.

I made some quick changes in CSS to address the basic readability problems, and uploaded a screenshot:

http://new.tinygrab.com/d5ac58cc253b078fa42a97b3c35417d28563...

(I was playing around with the links before I decided I was getting too far into the weeds, but the links in #content should have the underline. The links on the side shouldn't.)

CSS changes:

    html {
      background-color: #FCFCFC;
    }
    
    html, a {
      color: #303C3C;
    }
    
    blockquote {
      font-style: italic;
    }
    
    #header {
      margin-top: 50px;
      margin-bottom: 30px;
    }
    
    #sidebar {
      width: 150px;
      margin-right: 50px;
    }
    
    #content {
      width: 550px;
      margin-left: 200px;
      padding-left: 0;
      line-height: 150%;
    }
    
    #TOC {
      width: 200px;
      margin-right: -235px;
    }
    
    #TOC li {
      margin-top: 0;
    }


> Third, you need to constrain text width. It's a lot of work to scan 1000 pixels for each line.

It's funny you criticize that, because it's one of the only parts of my site design that I actually have A/B tested: http://www.gwern.net/a-b-testing#max-width The smaller widths did not do well.

> Books are also not backlit. If you want to recreate the experience of reading a book, use a slightly off-white color and a dark colored ink (for example, a very dark blue or brown).

Hm, this sounds plausible.

    html {
      background-color: #FCFCFC;
    }
    
    html, a {
      color: #303C3C;
    }
Done.

    blockquote {
      font-style: italic;
    }
No; I use blockquotes too much and they're too long to make them italics. The background shading seems like enough for now.

    #header {
      margin-top: 50px;
      margin-bottom: 30px;
    }
I'll try it.

    #content {
      width: 550px;
      margin-left: 200px;
      padding-left: 0;
      line-height: 150%;
    }
Duly stolen. And the TOC changes look like an improvement too.


The spacing (or lack thereof) on the site is very distracting. The left menu has zero pixel spacing from the content, and so the side menu basically runs right into the content. The table of contents on the right isn't obvious that it's a table of contents for the article, it just looks like another menu of some sort on the right.

I did some quick restyling just to satisfy my curiosity with some notes. I know it kind of makes it looks like wikipedia, but wikipedia is easy to read.

http://imgur.com/G5QfOua


I've added a 'padding: 1em;' to the sidebar so hopefully that deals with your left menu issue.

Looking at your screenshot, I think I like what you did with the table of contents. Could you post the CSS you used to do that? (Maybe it's obvious to you how to do that, but I know very little CSS so it's a mystery to me.)


Happy to:

Change the div#TOC rule the following:

  div#TOC { 
    float: left;
    width: 25%;
    font-size: 10pt;
    background: #eee;
    margin-right: 15px;
    margin-bottom: 15px;
    line-height: 15pt;
  }
I hope that helps!


Yes, it does. I've tweaked it a bit and added it to the site CSS.


It's not really that bad, just a few quirks:

- Left sidebar has no right padding

- The right sidebar loses its left padding in narrow viewports

If you want to go hardcore, there are a couple of other things, based off of James Felici's Complete Manual of Typography :

- The measure (line length) at its max width is about double what it should be for the font size (which is inheriting the browser default, usually 16px). Measure in picas (1/6 inch) should be 2-2.5 times the font size in points (1/72 inch, but a straight conversion of 16px = 12pt is usually fine).

- Currently, the line height is set to 130%. If using your current measure, the leading should be about 5 points, for a line height of about 141%. If you halve the measure, the leading would be about 1.5, for a line height of 109%. However, I'd still set the line height to be ~140% because you have a lot of superscript and your links have underlines that go below the descenders. Leading in points should be Measure in picas / font size in points, rounded to the nearest half point. Increase if you have things that go beyond the descenders (like you do) or decrease if you don't have any descenders (e.g., all caps).


> - Left sidebar has no right padding...- The measure (line length) at its max width is about double what it should be for the font size (which is inheriting the browser default, usually 16px).

Fixed or wontfix, see https://news.ycombinator.com/item?id=5654011

Fixed.

> - The right sidebar loses its left padding in narrow viewports

Don't know how to fix this.

> - Currently, the line height is set to 130%. If using your current measure, the leading should be about 5 points, for a line height of about 141%. If you halve the measure, the leading would be about 1.5, for a line height of 109%. However, I'd still set the line height to be ~140%

Done.


The issue is not black text on white background, but the layout choices.

This is what it looks like when I load the page: http://i.imgur.com/kmMsyEg.png

And here's how it looks after some fiddlin': http://i.imgur.com/JhwwjfY.png


Made an imgur album: http://imgur.com/a/8XUDP

First is with AdBlock on, the second is AdBlock off.

It's better once you've scrolled down, but everything is poorly mashed together at the top. Multiple words in the left column are badly wrapped.


Nothing's wrong for black on white for books. The thing is I don't read your site on a piece of paper but on a screen (and I'm probably not the only one).


I actually like it quite a bit . . . The lack of triple-nested faux windows is nice and I usually set things to display compactly anyway, so there's that.

It looks like some people are having font issues, though. I'm not seeing any overlap between the columns, but there's a lot visible in some of the linked images.


Just checking: is it supposed to look like this? http://i.imgur.com/PWbZHW5.png

The color is totally fine, it's really just that sidebar that's odd.


Interesting, on Chrome 26 for OS X 10.8, it looks like this:

http://i.imgur.com/OnRL0h1.png

I agree, the colors are alright, but the layout could use some TLC. The sidebars could be wider and need some padding, the line height should be increased and uniform for all the different styles so that the blocks line up, headlines shouldn’t to be uppercase as that decreases readability, and the indents for quotes and unordered lists could be smaller.


The font difference probably comes from the A/B testing Gwern is doing: http://www.gwern.net/a-b-testing#fonts.


Monitors are not paper. They're much brighter.


Great article and analysis.

One answer to your question, "what's wrong with black text on white background?":

http://ianstormtaylor.com/design-tip-never-use-black/


That answer would be a more convincing if the page itself wasn't a great demonstration of the opposite. I'm sure #554 with a spindly font looks great on a retina MacBook, but it's just not all that readable on most systems, especially for people with less than perfect vision.

I just don't get what issue designers have with contrast. You'd kind of think that the point of text is for it to be readable, not for it to look soothing.


Eh, sometimes you don't want things to look natural for the sake of looking natural. Plus, given that there's basically just black and white on Gwern's site, a dark blue would have the same 'overpowering' effect as a black, since there is nothing else that is similar in color to it.


I usually try and use Readability to read long and interesting articles... but it did not like Gwern's site.

http://www.readability.com/articles/5jqxbits


Hah, what's going on there? I don't understand how it can take nice regular Pandoc markdown-generated HTML and crop it down to nothing but the source code. Maybe you should file a bug on Readability.


Its definately a bit hard to read as it is, its a bit more academic doc vs blog.

For lulz I fired it up in Lynx and wow does it look bad there too...


We just launched our Google FeedBurner replacement in anticipation of its upcoming doom: http://feedsnap.com (release info http://neosmart.net/blog/2013/worried-google-will-kill-your-... )

We're planning on "launching it" on Monday.


Picasa seems unnaturally safe in this model. It's also stopped iterating generally and Google's acquired a totally separate mobile app for photos (Snapseed). And Google really doesn't make any material amount of money off Picasa either.


Picasa is slowly getting replaced with Google+ photos. I don't expect it to be around all that much longer.


And Linux support (which was only through Wine) was officially dropped.

Even before the recent spate of services being shut-down, shortly after the Linux support was dropped, I moved Picasa to a kill list and started migrating what I was using it for to something else (Dropbox for storage and sharing).

I presume it's going to be killed, to me that's what the product state signals.


Dropbox also has photos now, when you log in on the bottom left there was an opt-in for the beta testing. It's really nice compared to Picasa/G+'s "2 different photo services but not really" design.


Yes, Picasa's rating seems wrong enough to me that I called it out in the discussion as probably wrong. (Not sure how one would improve the model... how do you objectively rate or quantify something like 'stopped iterating generally'?) And it make not make 'any material amount of money', but it is still selling advertising and whatnot on Picasa: https://support.google.com/picasa/answer/166073?hl=en So the binary variable for 'profit' gets flipped...

Snapseed is not doing any better, though. IIRC Google already shut down an app for Snapseed.

I think Picasa might wind up being merged totally into Google+, although by my established criteria that would not count as a shutdown.


You mentioned in passing that you have a feeling that Picasa isn't surviving for long due to it's "old-fashionedness".

Considering how important "usefulness" and "usability" are by now, maybe add a criteria along those lines - how large is the "your mom uses it and likes it" factor, how "contemporary" are the UIs, how much effort puts Google into polishing it (it does a lot with G+ and Gmail for example) - something like this.

Another criteria could be "level of integration" - how much can you use a Google product as a standalone project or how deeply connected is a product into another one (e.g. Picasa used outside of G+) - which might in the end indicate not a direct shutdown but a product's dissolution into another one.


I'm not my mom, so I can't do that! Such subjectivity is something to avoid in an analysis like this. And 'level of integration' seems just as hard to assess.


"Picasa" the desktop app maybe, but I think the reason Picasa web albums/API lives on is that it provides the backend for Google+ Photos and they don't want to break existing apps using that API just to rename it, so continue to build stuff on it. Most recent example would be the Photosphere viewer API they just released: https://developers.google.com/panorama/web/


Am I the only one who sees Google Groups at risk? It could be seen as a competitor to Google Plus the update it's getting is awful and looks like something designed in 2009, the updated before that happened a decade ago, it's not very popular with the general public... it sounds a lot like Google Reader to me.


You'd think so, but the comparison to Google+ doesn't seem very apt - it shouldn't be a threat any more than Gmail is (but something to be leveraged).


I don't think that Google Voice will get shutdown. Google needs a unified communication product, tying Google Talk, Voice, and Hangouts together.


This is said to be coming in "Babel."


This is the first time I heard about changes to Voice and I really really hope they just leave Voice alone. This is the one Google product that is deeply integrated into my life and I rely on it continuing to work (it's also the only Google product I pay for). Sure it could become better, but lately when Google makes things better they usually get worse (at first, or forever).


I fat fingered a down vote for you on my iPad, sorry. I hope two passers by give you up votes to make up for it.

Google Voice is quietly used by a high percentage of colleagues as a "Line 2" to keep work comms under control and out of their personal SMS and VM, and by others as a way to have a truly portable personal number.

The only downside is occasional but serious VOIP quality issues even when landline to landline (routed through a GV number). If Google fixed that, sure seems to me GV could become a defacto identity endpoint for the mobile device generation.


Regardless of the actual subject, it looks like an excellent analysis. I haven't verified the maths, but I like that he includes every possible calculation, source code, and external reference that was used. A very academic and data-driven approach.


If anyone is interested in helping out, I have created a timeline for browsing such services: http://jensenbox.github.io/timeline/


That's a nice start. (It's a good idea - I was actually trying to do a similar 'stacked lines' plot to show the lifespans of all entries in my dataset, but I gave up at the time because it looked too hard and grotty to do in R.)


It's been a few years since my 3 stats courses, so pardon if I'm wrong, but aren't his p-values for the logistic regression kind of... important?

Intercept and log(hits) sure, but the rest have rather high p values indicating that it's likely to be present even if it has no effect (lower p-value, more robust, < 0.05 => < 5% chance of incorrectly stating it has an effect (when it does not)).

Or am I missing something basic here?


Definitely agree with Feedburner and Alerts being likely targets based on how well (or not well) they're supported currently.


You do realize that the 30% that Google takes from Android Developers from App sales doesn't go to Google right?


No, I don't.

But even if they don't, Google is still serving ads to Android users AFAIK. While browsing on exactly that topic (pondering whether to flip the profit bit on the Android entry), I ran across an estimate of $1 billion: http://www.forbes.com/sites/ericsavitz/2011/02/08/googles-an... This made the decision an easy one.


where does it go?


To the respective mobile carriers.


What makes you think that?


[Citation needed]


You do realize it does right?


Ah, a study based on something Charles Arthur said?

I once had an email conversation with Charles Arthur, it didn't go well...

http://en.wikipedia.org/wiki/User:Ta_bu_shi_da_yu/Charles_Ar...


More like, 'a study prompted by noting serious flaws in an Arthur column and deciding to do better, while borrowing perhaps <1/10th of the final dataset from Arthur's data' (he had ~30, I wound up with 350).


When will blogger be killed and rolled into Google plus?


If you believe the model, roughly 'when hell freezes over'. I find it a little hard, but I guess we'll know who to believe during the next 5 years if Blogger is indeed killed.


Google Finance is definitely not being maintained and I wouldn't be surprised if it's killed.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: