Splash screens are useless: they don’t solve any user’s problem, they shamelessly invade the user’s world, they draw additional attention to the fact that the user is waiting. There is no point in advertising your product to the person trying to actually use it.

In mobile apps splash screens are particularly out of place: they occupy the whole screen and the user has nothing else to look at. The phone is being slow, and a nice picture (which I’m undoubtedly looking at) tells me whom to blame. Isn’t it a huge marketing success?

iPhone’s standard app launch image is its screenshot. While the app is launching, the user has a couple of seconds to look at the interface and prepare for action. It appears that the app has launched instantly. According to iPhone’s Human Interface Guidelines all apps should behave like this, it’s even stated explicitly that you should not use splash screens. Unfortunately, this is the rule which Apple does not enforce, and so sometimes stupid apps with splash screens make it into the App Store.

If you find yourself drawing a splash screen, better do something useful, e. g. meet the programmer and discuss the potential design changes to make startup time shorter.

Sometimes the sudo command stops working for reasons beyond human understanding. When you try to sudo anything, it says: “no valid sudoers sources found”.

It happened to me twice. The first time it happened to a MacBook. I tried to make some changes to the sudoers file, but it didn’t help, and I gave up. After adding an SSD drive to my Mac Pro and moving the system to it, it happened again. Mac Pro is my main working machine, so I had to fix it this time.

Turned out the problem was with the permissions on the root directory. Presumably I’ve somehow broken them while moving my system.

I found a working solution on MacRumors forum. Here’s what you do:

1. Restart your Mac is a single-user mode (press and hold ⌘+S during boot until you see command prompt).
2. /sbin/fsck -fy
   check the filesystem integrity (not sure it’s necessary, but I did it anyway).
3. /sbin/mount -wu /
   mount the root filesystem.
4. /bin/chmod 1775 /
   fix the permissions.
5. /bin/sync
   commit the changes to filesystem.
6. exit

After booting to the normal Mac environment, you may also need to repair permissions with the Disk Utility.

It’s always a pleasure to talk to Kostya Gorsky, lead designer from Yandex and my good friend.

We’ve been discussing the future of browsers, apps and the Internet lately and have formed a vision of how all this stuff should work. I’ve been convinced that the address bar is not necessary and will die some day, and that there won’t be any meaningful difference between apps and websites. This simplification raises some questions and causes some problems at first, but they are all technical. People won’t care. We can make it.

Today one can find a site with Google from the Spotlight page on an iPad and add it to their home screen — all without knowing anything about web addresses. Even if you can explain the difference between an app and a website to your grandpa, this will not be of any value to him. Why bother, it’s all just coloured rectangles.

Kostya have recapped it all well: “You should have a screen with the coloured rectangles and a way to add them to it”.

And so I’ve coined a new design for iPad’s home screen. See the App Store icon. Before:

After:

Of course, manual updates will die, too.

We’ve went into much more detail with this, having talked for about an hour. And I wanted to publish this picture.

Life does not present you with a choice to be or not to be a designer.

Everyone always has to make design decisions. There’s no escape: whatever you do, other people will interact with your product or service. Your design decisions can be good or bad, but they cannot not exist. You have to make them, or you are stuck.

To make these decisions, you have to be competent. That’s why design is a basic discipline, as language is. Saying you don’t want to learn design is like saying you don’t want to learn to write. It’s bullshit. Kids should start learning design in primary school: it’s for everyone.

You can’t just make crap and say “hey, I’m not a designer”. There’s no such thing as “not a designer”. If you make crap, you’re just a crappy designer. Fix it.

As I’m sure many designers do, I have an Assets folder with a collection of small things you always need (mouse cursors, spinners, checkboxes etc.). There’s one thing that’s so easy to create from scratch that I don’t even bother to store it anywhere.

Let’s say you want to add a popup calendar to a date input. Here’s how you make a calendar icon for it.

Start with a 4×4 black rectangle (showing at 1200% zoom):

Make three of them in a row, with overlapping borders, like this:

Clone this whole thing and move 3 pixels down, like this:

Now clone this whole thing and move 3 pixels down and 3 pixels to the left, like this:

Done (don’t try to click, it’s just a picture):

Ben Brown asks if it is time for password-less login. Well, yes it is. It’s actually long been.

I think an even better solution would be to remove the password completely, allowing users to login with only an email address. Each time a user needs to login, they enter their email address and receive a login link via email.

That’s part of our usual process of removing mandatory signup. We try to convince all our clients to do this (see also: Interface is Evil). The thinking goes: password is not really needed to sign in: you can always click the “I forgot” link. So why not just assume that the user has forgotten their password?

This flip makes sense for almost every website. If users visit a site infrequently, signing in via email is not a big deal. If users visit a site regularly, then chances are high that they are already signed in and won’t see the form anyway.

In Bureausphere (it’s a kind of online designers club available only in Russian), we go even further. There’s no sign in or sign out at all. On the “me” page, if we haven’t recognized the user by cookies, we just ask for an email (I’ve translated this bit for this screenshot):

If we know this user, we just send a link to their page, and the link logs them in behind the scenes. There’s no way to log out. And if we don’t know the person, we still do the same: the page will be created on the fly as the user clicks the link in the e-mail.

There is one problem with this approach. While no user likes messing with passwords, people at least understand how the system works. So if we change it, even for the better, we must take into account that some users will be confused.