Yeah, I built the front-end of that site. Lots of long hours. Lots of gray hair. And now, I get to brag about it.

Indulge me, I rarely get to do this.

…

Okay, I’m done. Time to move on.

I downloaded XCode 4.3.2 (I think) from Apple and installed it, then I downloaded and installed RVM, then I tried to install Ruby 1.9.3 via RVM. That’s when things went south.

Hell began with errors about missing build tools. I checked, and–WHO KNEW–out of the 3+ GB of XCode that I downloaded, no command line utilities were included. Really? Okay, then WTF is in there, Steve’s junk email and browser cache files from the last 10 years?

Whatever, you can install command line tools from within XCode or download them separately through Apple’s developer website and install them without XCode. Installing the command line tools took care of the first set of OMG-missing-everything errors I was getting when building Ruby.

The second set of errors was harder to solve.

You’ll see this if you try to build Ruby 1.9.3 using RVM on a Mac: openssl_missing.h will throw all kinds of type conflicts during the build. Lots of websites out there say to move files around, get rid of MacPorts, get rid of Homebrew, etc. I have MacPorts installed, and OpenSSL was built/installed from it, so I removed OpenSSL from MacPorts (see MacPorts docs for that. You’ll have to remove lots of other packages that depend on OpenSSL before OpenSSL will leave) and then I reset the Ruby build within RVM so I’d start with fresh build configuration files:

Eriks-MacBook-Pro:~ erat$ rvm remove 1.9.3
Eriks-MacBook-Pro:~ erat$ rvm install 1.9.3

That fixed it. For cryin’ out loud…

Note: some folks say this is new to Ruby 1.9.3, so if you don’t need it, maybe sticking with Ruby 1.9.2 will work better for you. Your call.

When browser vendors want to support CSS attributes that are either proprietary or experimental, they frequently isolate their implementations by adding what’s called a vendor prefix to the attribute. This allows vendors to experiment, and it’s generally understood that if/when the attribute becomes fully cooked and part of a CSS standard, the vendor prefix will be dropped.

CSS3 is all the rage right now (if you hear folks buzzing about HTML5, I can almost guarantee they are actually referring to CSS3), but CSS3 is not a complete standard yet so pretty much all of it is experimental. An example of an experimental CSS3 attribute is transition, which ‘animates’ the changing of numerical data such as margins and heights. Vendors that are still working out CSS3 transitions will offer support for it using prefixed attributes. Web developers that want to utilize transitions need to specify not only the standard name for the attribute but also all variations that use prefixes, like so (these examples transition all numerical attributes over a span of one second):

-moz-transition: all 1s;
-webkit-transition: all 1s;
-ms-transition: all 1s;
-o-transition: all 1s;
transition: all 1s;

In order, we have prefixed attributes for Mozilla/Gecko browsers like Firefox, Webkit browsers like Chrome and Safari, Microsoft’s Internet Explorer, and Opera. The last attribute is the standard attribute without prefixes. To future-proof the CSS, this should always be last in the list.

Okay, end of back-story.

On mobile devices, the king of browser tech is Webkit. It’s the basis of Mobile Safari which is used almost exclusively on iOS devices like iPhones and iPads; it’s also the basis for browsers that ship with Android-based phones and tablets. There are mobile versions of Firefox and Opera, and Windows phones use a mobile version of Internet Explorer, but web developers’ love of Webkit is a force to be reckoned with. It’s completely understandable: Webkit is feature-rich and available almost everywhere on mobile and desktop devices. If your website works well on Webkit browsers, you’ll be covered on most mobile devices and a growing chunk of desktops.

However…

It turns out that the fore-mentioned affair between web developers and Webkit is becoming monogamous. Websites are popping up left and right that only offer support for Webkit, even to the point of sniffing browsers and turning non-Webkit users away. One of the main motivations for doing this: developers are tired of typing five lines of CSS code for every experimental attribute. Recall the chunk of CSS from above, the one that had a line for each vendor-prefixed attribute. Multiply that times the number of experimental attributes in a stylesheet and you’ll see why this is an irritant. I know it bugs the hell out of me. Still, I do it. Like many developers, I wish (almost daily) that there was only one web browser, that it worked on every O/S, that it wasn’t proprietary in any way, that it rocked, and it was free for anyone to use. That day may come. Until then, monogamy isn’t a viable option.

Get to the point already

So, yeah, you wanted to know what’s so pesky about -webkit- vendor prefixes. Here it is: Mozilla, Opera, and Microsoft–all the vendors that produce browsers that are not based on Webkit–are now planning to support -webkit- vendor prefixes.

Silly, yes? Not really. They all have mobile browsers that support most of the same experimental CSS as Webkit, but lazy web developers aren’t adding the full list of vendor-prefixed CSS attributes. If Mozilla, Microsoft, and Opera add support for the -webkit- prefix, websites that previously looked broken will suddenly work fine, or at least that’s what they hope will happen.

This is potentially bad.

To make matters worse, it seems that Apple (clearly basking in the dominance that Webkit–primarily an Apple product–enjoys) is planning to never drop the -webkit- prefixes. POOF! There goes the single most effective way to stop Webkit monogamy. If this happens and other browser vendors add -webkit- prefix support, kiss early access to experimental CSS goodbye. Essentially, we’ll find ourselves in a new only-works-in-IE6 era of web hell, only this time it’ll be an only-works-in-Webkit hell. Sorta.

Okay, maybe hell is too strong a word. It won’t be ideal.

Having said that, I do think it’s possible for a Webkit-dominated web to work. A few things would need to happen:

• Apple and Google need to share all of their extensions to Webkit so they can be shared as web standards,
• Apple and the rest of the Webkit maintainers need to welcome participation from folks they perceive as competitors, and
• Microsoft needs to make Internet Explorer render sites exactly like Webkit

Don’t hold your breath on that last item. You’ll turn blue.

Bottom line: if you’re a web developer, use all vendor prefixes, not just Webkit’s. If that chafes you, use SASS or some other CSS pre-processor to do the duplication work for you. One way or another, just do it. If you’re already doing it, thanks.

Peace out.

The change: I just accepted a full-time position at tenfour agency. My title won’t fit easily on a business card, especially the tiny ones they use at tenfour: Senior Front-end Web Developer/Designer. Inhale.

The not-so-happy part of this deal: I need to put Modal to rest. My attention needs to be focused on my new role at tenfour, and working full-time and freelancing don’t mix well (I tried it. It doesn’t work), so one has to go. I’ll miss the loose schedule, the ability to meet people during the day, not having to request vacation time…but I’m trading it all for a rockin’ job with an awesome crew, and for that I’m grateful. I’ve admired tenfour from a distance for some time now so it’s a real kick to be taken under their wing. Good times, totally.

To all my clients here in Oregon and beyond, thank you for helping me get to this point. I appreciate the work, and I appreciate the mix of folks that I’ve worked with. I wish you all the best, and I hope you find someone out there to take my place that’s almost as good as I am. Okay, as good if not better. But we all know “almost” is right, don’t we? Yes, we do.

I’m keeping this website around for a least a little while. Maybe I’ll keep it forever. My LLC is going to be killed within days, but the domain is mine so I may as well continue with it. At least two people read it and I don’t want to make them mad.

Happy new year, folks! It’s already a doozy.

E.

What makes Apple’s CalDAV incomplete? It only syncs “pop-up” (gCal) and “message” (iCal) reminders. Anything else is dropped. Best I can tell, even combination reminders (example: “message and sound” in iCal) are dropped. Folks that like getting eMail and SMS reminders are SOL.

It’s nice to get any reminder from a calendar app, especially if you blindly depend on these things to get from one end of the day to the next. For a long time, I have watched in dismay as I set alarms on iCal events only to see them disappear when the gCal sync is completed. Ditto in the other direction. iPhone, ditto again. Now I know to avoid certain kinds of reminders. All is (almost) right in the world.

I only found out about this Apple CalDAV issue today. It’s not earth-shattering or even approaching a nuisance, but still, I hope Apple fleshes out their CalDAV support soon. Seems like an easy enhancement to me, but what do I know.

Exercise: Using Chrome 16, go to the main page on this website and hover over one of the three navigation links that you have not visited. You should see the link color transition from orange to black. Now, click on the link, then after the new page loads go back to the main page. Hover over the same link. No transition.

Setting up CSS transitions for a:visited doesn’t fix this. Safari, “The Other Webkit Browser,” doesn’t have this issue. Smells like a bug to me.

Hey, I’m just trying to help.

The first step is fairly obvious: remove the hard disk from the ED Mini and put it in an a USB drive enclosure that supports eSATA 3 drives. I used a Rocketfish USB3 (compatible with USB2) enclosure, about $80 at Best Buy.

Here’s the sticky bit: you need to use Linux to get your stuff off of the drive. It’s formatted with a number of Linux filesystems and one “XFS” filesystem, the latter used for the actual storage. You’re not going to get to the XFS filesystem through Windows or Mac OS. Don’t even try, it’ll only frustrate you.

I didn’t have a Linux system handy but I did have a PC laptop, so I downloaded the latest Ubuntu CD image (available at http://www.ubuntu.com/download/ubuntu/download), burnt a CD from the CD image, then booted the PC laptop from the CD. If you love Ubuntu feel free to install it, but I just ran it from the CD.

When Ubuntu was fully booted, I popped over to a virtual terminal by hitting CTRL-ALT-F3 (aaaah, the dreaded command prompt!), then I ran the following command:

sudo rmmod uas

This “uas” module was preventing the computer from mounting the partitions on the external USB drive. I don’t know what it is and I don’t care. It was bad, and I killed it. I then hit CTRL-ALT-F7 to get back to the graphic desktop.

Once that module was removed, I believe the partitions on the external USB drive automatically mounted. That, or I unplugged the USB cable and re-plugged it. Either way, holy moly, file browser windows started to pop up everywhere from the former-EDMini drive, including the partition formatted with XFS (the data partition).

Awesome! The files were all there and appeared to be intact. Then what?

I copied them off, of course. I had a spare 320GB USB drive sitting around so I plugged it into a spare USB port and copied files from the ED Mini disk onto it. Note that because of permissions issues, I couldn’t drag/drop files from the ED Mini disk to my spare USB drive. I had to go back to the CTRL-ALT-F3 virtual terminal and manually copy the files over. If you need assistance with this, feel free to leave a comment and hopefully I’ll see it in time to help. It’s more than I’d like to type up here.

So, there you go. May you never have to deal with this, but if you do, hopefully this blog post will help you get your files back. It only took me about a week to get this figured out. Ouch.

My main computer for work and personal use is a 1st generation MacBook. This is not what I’d call a weak machine: 2GHz Core Duo processor, 2GB of RAM, a 500GB 7200rpm drive that I installed a year or so ago when I ran out of room…I could go on.

No, it’s not a MacBook Pro. Some folks chase Apple hardware annually; I, however, use my Macs for years before considering upgrading. I paid a premium for Apple quality so BY GOD I’m going to use the damn things as if they aren’t disposable (like Dell PCs. Don’t go there).

The curse of these 1st generation MacBooks is that they’re 32bit machines. Don’t worry if you don’t know what that means. All that matters is the newer Macs with Core 2 Duo processors (note the new “2″ in the name) are 64bit so they can chew more metaphorical food in their larger metaphorical mouths than a 32bit machine. That said, these 64bit machines run 32bit apps just fine. The reverse, however, is not true; if you build an application for a 64bit architecture it’ll only run on a 64bit machine.

Logic (at least to me) says you should build applications for the widest reasonable audience. Adobe does that: it builds 32bit applications that run well on my MacBook. If a piggy application suite like CS5 can run on this baby, I have to guess smaller/lesser applications should be able to run as well.

Silly me.

Clearly I’m mistaken, because it seems like every week I run across an application that either no longer supports 32bit Macs or may drop 32bit support soon. I’m not talking about CS5-level applications, either. Oh no, these applications are much less demanding and should not require 64bit machines. Applications like TweetDeck and Github for Mac are two examples. Neither application is so intensive that it requires a 64bit machine, yet I can’t run either on this still-plenty-powerful MacBook.

So why are developers catering to the bleeding edge? Is it because of users? Have Mac users taken a shine to disposable hardware, similar to the PC world?

What’s the point in paying for extra quality if you’re going to need new hardware every couple of years? I can pay around $1,300 for a new MacBook Pro or I can pay a few $hundred for a new PC laptop with similar specs. Sure, it wouldn’t be a Mac. However, if this trend continues, I can see the PC laptop becoming my main machine and the MacBook being used to test Mac browsers. Tables can turn rapidly, and although I’m a Mac guy, I’m not so married to them that I can’t function almost as well in Windows 7. The majority of my work happens in text editors and web browsers, most of which are platform agnostic. My editor of choice, Vim, is available on just about every platform in existence, and the only web browser that I work with that is stuck on one platform is Internet Explorer. And besides, so far I haven’t heard of many/any Windows apps that absolutely require 64bit hardware. That, my friends, looks like a Mac exclusive.

I’ll note here that Apple is not the problem. Yes, Lion will only run on 64bit hardware, but applications that run within Lion don’t need to be 64bit. No, folks that build their software for the latest and greatest hardware are at fault here. I’m not sure what their motivation is, but I’ll thank them for knocking it off.

UPDATE: It has come to my attention that Apple is actively encouraging developers to abandon 32-bit hardware. Smooth move, gents.

I couldn’t agree less.

There is a saying that I throw around liberally, one that I adapted to apply to designers: ask 10 designers for their opinions and you’ll end up with 11 opinions. After folks do a little math and figure out that 11 people are actually involved here, they generally disagree.

But if you think about it, this really is the core of the issue: everyone thinks they’re great at design even if they’re mostly great within their own minds. Design, for all its history, practice, religion, and celebrity applications, is subjective. Sure, we can say a poorly kerned headline is distracting, but can we say all well-kerned headlines are awesome? Show me a set and ask me which ones work and why they work, then do the same with other designers, then see how many of us agree 100%. I’m going to guess none will. There’s a valuation here that can’t be summed up with kerning or other measurements, and that’s the level at which we all, as designers, begin to diverge.

Before digging into this issue, let me point out the two problems from which I think licensing committees will suffer:

• They’ll establish guidelines that will be disagreeable to the majority of designers, either because the guidelines are too picky or because they’re not picky enough.
• They’ll license their favorite designers and friends, turning the licensing into a contemporary Algonquin Round Table where only the elite (or perceived elite, or maybe just the well-connected) will be included. Insert Groucho Marx quote about club membership here.

As soon as you distill a field down to a set of guidelines, you’ll need metrics for assessing the candidates. However, when the field is subjective, direct metrics seem impossible to define. The only usable metrics seem to be peripheral: checking web analytics for web design, consensus/nomination by peers for visual design, ROI reports for print design, etc. How can you accurately rank anything that is subjective using these kinds of yardsticks? You can’t, that’s how.

And as far as Algonquin Round Tables go, I could argue that the current slew of invitation-only, curated-membership design websites are indications that the haves- and have-nots are already being defined subjectively. As I stated earlier, 10 designers asked for opinions will result in 11 opinions. It all comes back to that. Members can curate all they want, but regardless of their convictions, many people will merrily disagree with their choices.

Question: would you license Paul Rand? He’s done lots of great work, but (I’m prepared to duck) in my opinion he’s also done lots of really lame/bad work. Ditto for Saul Bass (the self-cannibalizer), Michael Beirut, Stefan Sagmeister, Jan Tschichold, and even a few Bauhaus designers. The list goes on and on. You can’t say one is worthy while others are not because all of them have produced awesomeness as well as drek. Just my opinion, of course. I’m sure folks will disagree.

And I think that’s okay, the differences of opinion as well as the awesomeness mixed with drek. We all started with no experience, no portfolio, no design sense, nothing to fall back on. The folks that insist you should “fail harder”—horrible advice in my opinion—all started off fearing failure just as much as you and I did. We all do awesome and shitty work, so who’s to judge what’s great and what isn’t? Glass house, anyone? Expecting the elite to equitably determine criteria for licensing that does not limit it to other elites…I don’t know. It seems impossible to me, or at least highly unlikely.

Oh, and I lied earlier: there’s actually a third possible outcome to licensing, something that’s seen in some IT professional certification programs, the result of industries putting too much emphasis on credentials:

• So many people get licensed or certified that the credentials become worthless as differentiators.

It happens. Don’t believe me? Even when Novell networking was hot, there were so many CNEs in the workforce that the certification meant very little. It became like a high school diploma. That’s the curse of a popular, necessary credential: if it becomes ubiquitous, it becomes a minimum entry point for all professionals, the very thing that could prevent folks from becoming professionals at all. If that isn’t a chicken-or-egg situation, slap me.

Bottom line: licensing sounds good in heated designer discussions but it doesn’t make sense in practice. A designer’s work should speak for his/her abilities, not a license, a certification, or even a college degree. There needs to be a way for someone to get started, build a collection of work, hone skills, and enjoy the benefits of experience without having to start with a credential. Licensing, if implemented, could cut people off before they even begin. That may make some elite designers happy, but as the Utah Jazz saw when its few elite players retired or otherwise left the team earlier this decade, there’s a price to pay for not nurturing the next generation.

Think about it. I don’t know if I made any convincing arguments, but think about it anyway. And tell me I’m up in the night if necessary. Note: opinions are also subjective.

I can explain in detail why I made this decision, but I’ll let you ask if you really want to know. It’s nothing bad, just a time and availability thing.

Thanks for understanding. I’ll announce when I’m back in freelance action, don’t worry. In the meantime, stay dry, okay?