Please. Fix. That.

Web: status.twitter.com

A "status" page for a high-traffic web service such as Twitter is meant to provide users with service interruption news when the main site is offline or malfunctioning in some way. What it should not do, is be offline itself.

Twitter uses a Tumblr.com blog for reporting service interruptions, and I cannot fathom why. Tumblr is a good life-blogging service, but they have had a lot of service problems over the years. I would not select Tumblr for a mission-critical aspect of my business, were I to have a business that served millions of people.

Inevitably, when Twitter has a sustained period of downtime, their status page goes down as well. This defeats the purpose of having a status page. A flood of users are directed at Tumblr when these outages occur, and Tumblr doesn't seem to be able to handle this kind of load. It seems unfair of Twitter to abuse Tumblr in this way. Tumblr is a free service— I don't know if Twitter pays them a dime for barraging them this way periodically, but I bet they don't enjoy it, one bit.

Plain and simple, Twitter should manage their own status site. They have the resources to do it. Such a site should be static in nature and should be backed by enough redundancy to keep it operating even if every Twitter user were accessing it (the resources would be a small fraction of what is necessary to operate the full Twitter service). It shouldn't be Tumblr's responsibility to report service interruptions for Twitter.

Additionally, their downtime service should include an API as well. This way, if an application is getting an error, it can ping the downtime API to let the user know what's going on. This is particularly important when Twitter has decided to bring their services back up incrementally, making their web site work again, but leaving their API disabled until later. The downtime API can be fully static as well, returning JSON data instead of HTML, that's all. It need not require any dynamic Ruby/Scala/whatever backend to do that job.

Twitter, please fix that.

Pandora: Enough with the live recordings

Look, I love music. I listen to as much music as I can. But live recordings (particularly of a live venue) just aren't my thing. I don't care for the crowd cheers, the sing-alongs, the improvisation, the occasional rant-about-something-in-the-middle-of-my-famous-song. It just doesn't do it for me. When I hear a live recording of a studio track I love and is permanently etched into my brain, it borders on being painful.

I really wish Pandora would catch on to this -- I skip them every time. I can't thumb these tracks down, for fear that it will negatively affect general play for that artist. I'd love a simple preference in Pandora that would remove these from play for me.

Pandora, please fix that.

Twitter: Who to Follow

Update: Appears to be fixed! Thanks, Twitter.

I love the little things. Little things are easy to fix, too. Here's a little thing I found on Twitter's "Who to Follow" suggestions page:

 

(And I fixed the bug of not following @ylnt.)

Google: Search For What I Asked For

Google has adopted an annoying habit of changing what I have asked to search for with something similar but different to what I want. Perhaps this doesn't affect you, but for a software developer, it happens all too often.

Here's a recent example. I search Google for the phrase: "document insertImage" (without quotes). Google then goes and uses an algorithm of some kind to see what I really meant to search for. It determines that what I probably meant to search for was "document insert Image" (without quotes). Which is not what I wanted. Google explains this to me on the results page: "Here are the results for 'document insert Image' ... Search instead for 'document insertImage'", giving me the opportunity to correct their mistake.

But what I really want is for Google to search for what I asked for to begin with. Inevitably when this happens, I wind up scanning the list of the results and about half-way down the page I realize that none of these results are useful because Google didn't search for what I asked for.

I know Google does this to be helpful — a lot of people write searches with typos and this is Google's way of improving the search experience for most of their users. But for people like me that are pretty good at typing and need to search for programming phrases ("insertImage"), it just gets in the way. I would love a preference for disabling this feature altogether. I'd also settle for Google just doing the requested search and including a block at the top suggesting their alternative search if one is found relevant.

Google, please fix that.

Apple Developer Login: Stop preventing paste

UPDATE: Apple has changed their login form quite a bit and it no longer prevents pasting into the password field. But, other sites out there still employ this tactic for some reason. So to counter that, you can install a Safari browser extension that fixes it for them too.

When you go to sign in to the Apple Developer web site, and navigate to login. You get this little innocuous login form. But unlike most login forms on the web, this one hates you. It won't let you paste anything into the password field. Some developer at Apple, in their infinite wisdom has decided to prevent such a thing.

You can see in the screenshot here taken from the HTML source for the page, they capture and prevent the "onpaste" event (and needlessly blocks cut and copy, which the browser already prevents) for that field.

I'm at a loss as to why they do this. Is there a security benefit I'm missing? If so, why does NO ONE ELSE IN THE WORLD do this? I'm sorry, but some of my passwords are lengthy and generated and I can't remember them. So, I often copy the value from my password management tool (1Password) and paste it into the password field. Except here, I cannot¹. Is Apple trying to encourage short, memorable passwords that can be typed easily? Are they trying to discourage the user from copying passwords to the system clipboard? Cause they didn't prevent that; they prevented populating the value already on the clipboard into the password field. And that's not helping anybody.

Please fix that.

 

¹ YES. I realize the 1Password Safari extension can be used to populate the fields properly, but that still doesn't excuse breaking the expected and accepted behavior of a simple login form.

Google Reader: Broken next and interestingness

Update: Will never be fixed, since Google Reader is no more.

This post concerns Google Reader. In it, I will cleanly side-step:

• the recent redesign
• the loss of certain features
• the encroachment of Google+
• the issues around data portability and difficulties surrounding the transfer of former Google Reader social contacts to Google+

No, I'd like to draw a bit of attention to two specific things that Google seemed to have missed with all the recent attention their engineers have been giving to our beloved Reader.

Here's number one: there's no easy way to navigate to the next unread item after marking all items in a folder as read.

I don't know about you, but I use folders for grouping common types of feeds together and I tend to read feeds by folder. For instance, I have a folder for technology news, another for humor, another for world news, and so forth. This allows me to read more important categories of news before I get to the more frivolous stuff in the humor folder.

So, after I finish scanning through a folder for stories to read or star to read later, I will typically hit "Shift+A" to mark all items left in that folder as read. Having done that, I would LIKE to hit the spacebar at this point to take a look at the next unread item that is in a different folder. But, that never works. This is because in this state, Reader refuses to navigate any further. It all of a sudden has become very dumb about navigation and has no clue as to how to get to the next unread item.

Now, check this out. Instead of using "Shift+A" to mark all those items as read, if I were to hit my spacebar to read through each item individually and then reach the end of the items in the folder, hitting the spacebar again WOULD take me to the next folder that contained unread items.

This is how Reader has worked for as long as I remember.

Number two: interestingness.

Something I've been wanting for a long time now is a better way to bubble all the important, interesting things to the top, so I can spend less time sifting through the endless stream of Internet consciousness for the scattered pearls of content. You'd think with all the aspects of a hosted, widely used service like Reader, they'd have the data necessary to pull this off. But no, they apparently don't. (And "Sort by: Magic" doesn't cut it.)

But, they went and acquired a company this past June called PostRank, whose business it was to do exactly this thing. So far, nothing has come of that acquisition. But I'm hopeful that it's coming very, very soon.

So, those are the two things I really would like Google to fix with Reader.

foo.example.com: Don't forget to link to example.com

Congratulations! A visitor has found your web site. They came across it while doing a Google search that landed them at a weird address (something like "http://blog.bigcorp.com/archives/2011/blah-blah-blah.aspx"). After spending all of 5 seconds on that page, they now want to see what your company is about.

So, they click on the big logo in the upper-left corner of the page but it just takes them to "http://blog.bigcorp.com/". That's not where they expected to go, so they leave, thinking that you  didn't go all out and buy a "www" address to go with your "blog".

This scenario is how a typical person behaves on a site that cares so little about site navigation that they can't bother to put a link back to their primary home page on each and every page of their site.

Your average visitor isn't going to think to fiddle with the address bar and delete the "blog." and the "archives/...." bits. They just won't. If they type something up there, it's more likely going to be a different address, or a new Google search.

Today's example of this terrible oversight that is so oft-repeated around the entire web comes from microsoft.com. They're all excited about Windows 8 and how it's coming along and have set up a bunch of pages in the MSDN corner of their web site, found here:

http://msdn.microsoft.com/en-us/windows/home/

It's a nice site. Lots of videos and information. Even a free beta copy of Windows 8 to download and play with. I'll wait here while you try to find a link back to "www.microsoft.com" on that page. They even have a Microsoft logo in the footer area of the entire subdomain. But they didn't bother to link it to "microsoft.com". Tsk, tsk.

Their "buildwindows.com" promotional site is another example. That home page also has no link back to "microsoft.com"; again, not even the "Microsoft" in the copyright shown in the footer (they also have many company logos on that page, also unlinked which is sort of a dis, frankly).

As I said, this happens all over the web. Apple is just as bad: their developer.apple.com site has no link back to "apple.com". Oh, they have a link in the footer to their online store there, but not to the home page of their site. For shame.

If you manage a web site that suffers from this problem, please fix that. It's easy to do and it makes for better navigation across your web site. And hey — it could improve page views and time-on-site (if you're into that)!

So, a rule of thumb: if you start at the home page of your site and click around long enough, you should never reach a spot where you can't return to the home page without resorting to the "Back" button.

Simpler rule of thumb: always include a link to your primary home page — preferably an obvious one — on every page of your site.

apple.com: We'll be back soon.

I'm going to pick on Apple for this one, but it happens all over the web. Lets say you're trying to view a web page and for whatever reason, it's unavailable. Perhaps there is a server error, or perhaps there is scheduled downtime, or the site has been taken offline. So instead of seeing the page you were trying to view, you see a "Try again later" page. Okay.

Time passes. You try again. Still down. So you wait some more and try again. Still down. You may be viewing a page such as this one. Note the address:

Clicking the refresh button in your browser should cause that original page to reload. But when a web site does this to you, it never will. Because you're not reloading the page you were trying to view— you're reloading their "We'll be back soon." page, which will never change.

This is broken. There is no good reason for redirecting a user's request in this way. If you're doing this to your users, please stop doing that.

Web video players: stay in full-screen, please

Our family cut the cable TV cord years ago. Not that we don't watch TV anymore today. We just leverage our Internet service for it now, shown on the big screen via a Mac mini. This means using a combination of services (some paid, some free): Netflix, Hulu and iTunes mostly, but we also watch some TV shows on the network's site. See, some networks are doing their own thing instead of allowing Hulu to host their shows. This fragmentation is an annoyance, something that apps like Boxee are trying to fix (but so far, not quite there).

One of the networks doing their own thing is CBS. There's really only one show we watch on CBS today: CSI (original series only). The CBS video player is better than most self-run network video players — video quality is good, the CBS logo is relatively small and not obnoxious. But the commercials are the problem.

It's not the quantity of commercials — typically just one or two during a break — but some escape out of the full screen mode when playing the show. This happens with ads that have a mix of images and interactive elements. And when those ads end, the episode you're viewing resumes, but it's no longer shown full-screen.

I realize that there are limitations in place for Flash that prevents a Flash object from switching to full-screen mode programmatically. A user has to deliberately click on the full-screen button to cause that transition to occur. And if you think pop-ups and pop-unders are bad, imagine not having that protection... web ads would be taking over your screen all the time (so thanks for that, Adobe).

CBS should take a closer look at how Hulu does this. Hulu will sometimes show an interactive ad during a commercial break, but you never leave full-screen mode. That's because the ad is still all-flash, not a web and Flash hybrid ad. So there is no transition out of full-screen. It makes for a much better viewing experience. Remember folks, TV viewing is supposed to be as passive as possible.

So CBS (and many others, such as the player at cwtv.com, which deserves a much longer post all to itself), please fix that.

 

Flickr and self-favoriting

I've been a Flickr user for years and love the service, but if I could be the Flickr product manager for a day, I'd fix this:

I can't favorite my own photos. And you can't favorite your own photos. There is no technical reason for this. Flickr has simply decided that a user cannot favorite their own pictures, and that is that.

"Oh, silly user -- you can just make a set and put your favorite pictures in that!" No. It's not the same thing. When other people view my favorite pictures under my profile, I want them to see all of my favorite pictures on Flickr, even if some of them happen to be my own pictures. No, I don't want to futz around with photo set management to corral my own pictures into a "Favorites" set. I want to click that little star button and mark it as a favorite.

Another problem with the "just use a set" solution is that Flickr doesn't let me put other people's photos into a set. So that won't let me combine favorites from my photos and from others into a common place either.