ClipperHouse

Design is information

mwsherman — Wed, 23 Mar 2011 06:48:11 -0700

It seems to me the companies with the greatest customer loyalty are those where design has a strong presence in the corporate culture. I don’t mean this in a touchy-feely way. I mean that passionate customer loyalty is worth billions of dollars, and most companies see fit to leave that on the table.

Design is product. Fonts and colors are not design – they are decoration. Design is the hard work of understanding everything a customer might do, or want to do, how they will do it, and how they will be rewarded. It is an evaluation of the customer’s cost/benefit.

Design should report into the c-suite, or be in the c-suite. Two companies where this is the case: Apple (natch) and financial newcomers Square. I see raves about Square all the time. Those raves are customers that will tell ten friends or 1000 Twitter followers. Try purchasing that goodwill via traditional marketing.

More than anything, design is information. It is the conduit by which users needs are described and communicated. It requires you to articulate what you are selling.

Everything we do as programmers is based on our own tacit assumptions about what the product should be. We think of ourselves as logical creatures. But our premise – “this is what the customer wants” – is often so unempirical and unexplored as to make the whole enterprise a bit, um, onanistic. It is reason is pursuit of folly.

Every programming problem…

mwsherman — Sat, 19 Mar 2011 20:46:00 -0700

…can be entirely described in these four dimensions:

Cache: The browser is a cache for a web page, whose canonical version lives on a server. A web page is a cache for structured data, whose canonical version lives in (say) SQL. Structured data is a cache for bits that live on a spinning piece of rust. The questions for the programmer are, how similar is a piece of data to its canonical source, and how similar does it have to be?
Abstractions: You’ve hidden some words behind a smaller number of words. What words are you really saying, when you say DoThing()? It’s turtles all the way down.
Scope: What does this variable (symbol) mean at this moment? Will it mean the same thing a moment later, after we’ve gone off and done other stuff? There is a symbol over there which seems to be the same guy. Is it? And if I change it here, what happens elsewhere?
Delimiters: Aka encoding. Nothing is understandable without rules describing where meaning begins and ends. White space is a delimiter between commands in your programs. Slashes are delimiters in URLs. HTML, JSON, GZIP: defined by delimiters and encoding.

Every bug fix is an exercise in asking one of these questions until the answer is satisfactory.

Fixing the jQuery “jump” on slide

mwsherman — Thu, 10 Feb 2011 08:59:00 -0700

You may notice in some situations, when using jQuery’s slideDown(), that the last little bit of the animation “jumps”. It feel un-smooth and frustrates those who like to dot our i’s. (Here an explanation.)

The main issue is that the browser (and thus jQuery) is uncertain about the height of the element that is sliding. This, in turn, is based on uncertainty about the width – the width determines how lines wrap, and dictates the height.

If your width is defined in % or auto, it remains uncertain. So, to smooth the animation, let’s make it certain. Here’s a one-liner:

$elementToSlide.width($elementToSlide.width());

All it does is assign a pixel value to the width of the element, based on the … width of the element. Worked for me.

How Microsoft can make Visual Studio cool among the cool kids

mwsherman — Tue, 01 Feb 2011 20:34:00 -0700

I am quite partial to ASP.net MVC and C#. I’ve played with other platforms, esp Ruby, which have appeal as well.

While many aspects of the platform are, imho, technologically ahead of its competition (LINQ, Intellisense and Razor), I can’t shake the feeling that MS still appeals primarily to those with an enterprise background. Which is to say, the cool kids are choosing Rails and doing their editing on a Mac.

It needn’t be this way. Microsoft can do a small handful of things to make Visual Studio relevant to these trends, by reducing the unresolved pain points one experiences across all platforms. My ideas:

Best-of-breed IDE for dynamic frameworks, especially Ruby on Rails.

I haven’t seen anyone crack the nut of code completion for methods and properties that are likely to exist at runtime. Rails’ conventions are pretty clear, and it seems to me that Intellisense should be able to use this information to offer hints and save keystrokes.

At the very least, VS should do basic syntax coloring for Ruby and Python out of box, with Intellisense for core keywords – just as they did with jQuery.

Workflow and GUI support for the major source control systems: Git, SVN and Mercurial.

Command lines suck. There, I said it. I consider them a failure of imagination. Which is to say, I think of workflows visually, and I wish the tools were with me.

Nobody has done a great job with source control GUIs yet. The Tortoise-* and Visual-* folks have done some nice things, but they are not nearly where they could be. They are Windows 3.1, design-wise.

Imagine a clear visual picture of your source built into the IDE and Windows Explorer. Imagine your “git status” files staged in real time in a sidebar. Imagine a system that pings the main repo and gives you a hint of what others have done lately.

The simple tasks of add/commit/push/pull should be two clicks, max. No one is doing this yet, I see a huge opportunity for the one who gets it right.

PS: I recognize that Microsoft can’t *ship* any of these source control systems. But they can obviously detect a .git folder.

PPS: I know that source control workflows often require command-line precision. But the 80% case should be achievable in a GUI.

Chrome breaks 10% market share; modern browsers at 80%

mwsherman — Tue, 01 Feb 2011 11:19:00 -0700

The headline say most of it. Google’s Chrome browser popped up over 10% market share for January. Modern browsers* are creeping up to around 80% market share, which means the user experience of the web is moving forward.

Something I hadn’t noticed previous: Internet Explorer 7 has substantially less market share than its predecessor, IE 6. So, when a site decides not to support IE 6, their justification for supporting IE 7 is even less. (IE 8 is the market leader overall with 34%.)

The beta versions of IE 9 and Firefox 4 are neck-and-neck at 0.5% share.

* My casual definition of “modern browser” is any Chrome version, any Firefox version, and and IE version 8 or higher. I exclude browsers with less than 1% market share.

This is what a non-neutral net looks like

mwsherman — Tue, 01 Feb 2011 07:06:51 -0700

Gogo, which offers in-flight Wi-fi on a bunch of airlines, is now offering Facebook at no charge – but access to other sites will cost $$.

This violates net neutrality, no? Obviously, certain content is getting special treatment by an ISP, while other content is being blocked unless you pay up.

Am I being facetious? Not really. Discriminatory pricing is generally pro-consumer when you look into it. Different prices for different products usually translates to discounts, not price increases. ISP’s can raise their prices any time, today, so one must assume that they are already charging as much as they can get away with.

My aunt is paying full price for Internet access, somewhere around $50/mo. If she were offered a price of $15, with access only to (say) a dozen sites, she would happily take it, and use that extra $35 to further spoil her grandkids.

But that wouldn’t be neutral.

A call for industry-wide pixel-doubling

mwsherman — Tue, 18 Jan 2011 16:31:28 -0700

Computer screens have not changed much over the last 20 years, relative to other technologies like processing power and bandwidth. Yes, panels are better in many ways in terms of color accuracy and “resolution”. But most applications are designed around an assumption of 72 or 96 pixels per inch.

I put “resolution” in quotes because, while screens have inched up in absolute pixels-per-inch, we are not using those pixels effectively.

Each increase of 10% is not enough to justify a change in assumptions about how applications interact with the screen. The incrementalism is the problem.

Apple has had a rare opportunity with its iPhone. When switching over to the Retina Display of the iPhone 4, with exactly twice the resolution of its predecessor, Apple was able to dictate to its captive application vendors, deal with it. It’s time for the computer industry as a whole to make the same move.

How did Apple manage this transition? They supported previous-generation applications by simply stretching pixels. The user experience was not dramatically different – older apps might look a little blurry – while allowing new applications to take advantage of the hardware. Nothing broke, while new apps raised the quality bar.

We can do the same with the PC. By approaching this transition as an industry, though a standards body, we can make a quantum step forward in terms of on-screen quality and user experience.

While I don’t have much to say about native apps, but here’s how it might work in web browsers.

Typography

Many web sites specify their typography in pixels, despite good advice to the contrary. A new “pixel-doubled” browser would simply double CSS pixel sizes, unless otherwise instructed. Old sites would not break.

Then, as time permits, a site publisher could:

Improve the CSS to use true typographical concepts like em, or,
If they must, add a newly-defined META tag to instruct the browser to respect the CSS pixels (though this would be a bit silly)

Images & video (ie, raster images)

Similarly, browsers would simply double the pixel size of images in the browser by default. They might appears stretched on close inspection, but really they shouldn’t look worse than they did on a traditional-resolution screen.

(When you hit full-screen mode on that Hulu video, your computer is already doing something similar.)

Site publishers could then update their IMG tags with a new src2 attribute – this file would be handled in literal pixels, in most cases backed with a higher-res asset.

Dynamic & vector graphics

True vector images would not struggle with this, as they are already are resolution-independent. The element in the DOM in which they reside would follow the rules as above – rendered at double the pixels unless otherwise instructed.

Elements such as <canvas> would follow the same rules as raster images above – doubled by default – while adding an attribute or meta tag which communicates “I recognize that we are at double-resolution, my code understands that, render exactly the pixels I specify”.

Even better, for programmatic images, expose the presence of pixel-doubling via a Javascript environment property. Then the script could could test for capabilities and do its own math. (Tedious, I realize, but such accommodations are necessary with every new capability.)

If this isn’t done by the industry, some PC vendor (hmm, wonder who) might go ahead and do it anyway. I’d be OK with that. But it would be nice for everyone to be on board.

The Internet is not a public commons, and it is not Beetlejuice

mwsherman — Fri, 24 Dec 2010 10:34:29 -0700

There are some stories making the rounds that “the Internet” is a public commons built on private property. This is wrong.

Firstly, is nothing public about it. “The Internet”, as we know is, exists and is useful because it is private. Or, more precisely, it is a lot of private entities working together because they choose to.

Those private entities include the corporations that provide the infrastructure and services (ISP’s like Comcast and services like Google). They are not and have never been public property. Instead, they are self-interested organizations that choose to offer their services to the public.

Further, we are these sharing private entities, every time to write a blog post or upload a picture to Facebook. We do this voluntarily, and in your own self-interest. (Where self-interest means something that benefits us in some way – entertainment, money, or a feeling of making the world a better place.)

“Private entities, sharing” is what the Internet is, at its essence. Were it not organized this way, it would not be “the Internet”. It would not be at all.

“The Internet”, like “the market” or “the public”, does not exist. Rather, it is a shorthand for a lot of interactions, with amazing emergent properties, that we try to put our arms around.

Declaring it a public commons is a blithe exercise in semantics. It is like Beetlejuice: say it enough times, and it exists.

Now, I understand why we talk about it in this way. The arrangement works so well that it’s easy to think that it’s just a big, public thing that’s out there and does what we want. The cloud.

But here is the thing: if there is to be “the Internet”, it can be no other way.

It can either be multivariate and privately owed, or monolithic and publicly-owned – where “publicly” should be read as “state”.

Remember, there are two organizing principles by which citizens can participate with one another. One is plural (private), the other is majority-driven (government).

The implication of the above-linked article is that, post-Wikileaks, we should be concerned. After all, a bunch of private companies (such as Amazon) have refused to support Wikileaks with their infrastructure. Thus, we are at the will of self-interested, profit-seeking corporations in the public sphere.

It fails to mention that the private Internet is what made Wikileaks possible in the first place. Of course I would prefer that Amazon stood their ground and continued to publish the material. But if the Internet were run like a public commons, then the exposed organization (the state) would be the same one running the infrastructure.

(I dare speculate that Amazon withdrew its support for fear of legal trouble from same.)

To support Wikileaks-style transparency is to support a private, plural Internet.

The scourge of paid prioritization

mwsherman — Wed, 22 Dec 2010 07:35:34 -0700

I received a pair of earmuffs today via Amazon Prime. Great service! They got to New York in just two days, all the way from a Nevada warehouse.

I feel bad, though. Clearly, by my paying extra for this service, I’ve moved others into Amazon’s slow lane. For that, I apologize.

I downloaded a new ~~Justin Bieber~~ Gnarls Barkley track on iTunes just now. It took seconds! Being curious sort, I discovered that Apple is paying a company called Akamai to store the file nearby so I can get it quickly. Akamai in turn pays my ISP for the privilege of being close to their customers, network-wise.

Those other services – the ones who don’t have the ability to put servers at ISP’s – are two, three, ten hops away. They must feel like they are in the slow lane.

I am looking forward to flying home for the holidays, too. See you soon, Mom! Luckily, I have the means to fly across the country instead of taking the bus, or driving.

I just feel bad for people on the highways, since my flying has clearly slowed them down. Sorry gang! I hope you will forgive me.

What Ricky Gervais doesn’t understand about God

mwsherman — Sun, 19 Dec 2010 20:45:00 -0700

I am not a religious person. I also happen to believe that religious people bear the brunt of more everyday bigotry than most groups.

I have heard maybe two or three clearly racists sentiments in my life. By which I mean, uttered in my presence. I hear a similar number of slights against religious people every week, probably.

So upon reading Ricky Gervais’ explanation of his atheism, I had to cringe at his understanding of religion. It was, charitably, a (pedantic) swing and a (bigoted) miss.

What he fails to grasp is that faith is an aesthetic choice. It is not rational, and does not claim to be. When people are lifted up by a symphony, or saddened by a tragic play, they are undertaking the same, non-rational choice as people of faith. It is a decision to transcend, and be moved.

To invoke science in this is to dance about architecture. Gervais says that,

Science is humble. It knows what it knows and it knows what it doesn’t know.

Yes, good science is humble. A good philosopher (of which the scientist is a species) learns humility very quickly.

A person of faith begins with this humility. A belief in God is fundamentally an admission of not knowing.

One does not invoke science in describing how they are moved by a piece of music. One can apply science to this experience (in terms of psychology and such), but your decision to fire up a Yeah Yeah Yeahs track is not a rational one. One might ask if Gervais’ comic inspiration is “rational”.

Why do I mention bigotry? Sentiments like these:

I would just rather you didn’t kill people who believe in a different god, say. Or stone someone to death because your rulebook says their sexuality is immoral.

Does he believe this to be the mainstream of religion? If he were living under a repressive theocracy, and observed these sorts of things on a regular basis, I might be sympathetic.

I rather suspect that he is using headlines generated by a few to impugn any others he includes in the category. A person that harrumphs “See?!” while watching a black guy being arrested on the evening news would be an intellectual compatriot of Gervais in this regard.

I might also point out that the great killers of the last 100 years were atheists. And not simply on a personal level – Hitler, Stalin and Mao considered the elimination of religion to be a fundamental part of their ideologies. Between them, we are looking at, conservatively, 20 million murders.

(We’d need 500 deaths by stoning, every day for the next 100 years, to approach that. I can be pedantic too!)

So believing that religion is a force for conflict and oppression, when compared to other belief systems, is simply…irrational.