People Are Crazy...

This is a follow-up to my post from a few days ago about hardware and software. When I wrote that post, I brought up a contrast between Apple and Google's design philosophies. This just happened to be convenient example, as I could snap screenshots of the two apps for the purpose of illustration. Unfortunately, I realize after reading the post that the point I was trying to make with the comparison was muddled (that's what happens when you try to hammer out a post as quickly as possible), but I've also realized that it was a bad comparison in general. This is because people are crazy.

First, some background: I've been a pretty committed podcast consumer for ~5 years now: I listen to them while jogging, at the gym, doing repetitive benchwork, commuting, etc. I like to have a mix of 'news' casts to stay informed (e.g., NPR) as well as educational casts to learn interesting factoids (e.g., Econtalk, Stuff You Should Know, Caustic Soda). Recently, I've also started listening to a few different 'technology' podcasts focusing on news and reviews of the most recent products in computers, gadgets, and media. I was reccomended a few new casts, specifically Tech News Today and Build and Analyze, so I listened to a couple episodes of each (I'd reccomend the former and eschew the latter, by the way).

I've noticed a common thread among all of the tech podcasts that, when you step back and think about it, is very weird: Almost every time that certain companies or products are brought up in any context, discussion is either preceded or followed by extensive apologies to the listeners for having to bring up said company coupled to defenses against 'bias' and 'selling out'. I'm serious: think of how mind-blowing it is that when a show's hosts discuss a new Android phone, pointing out its relative merits and flaws, they literally have to spend a minute or more apologizing for their opinions and defending them as being based on 'facts' and not because they 'love Apple'. Alternatively, discussion of the new iPad announcement is preceded by an apology to all of the people who 'hate to hear about Apple' as well as the suggestion that they 'scrub forward' a few minutes into the show to get back to the rest of the tech news.

Take 5 seconds to think about what this image means.

Seriously, people are crazy. Admittedly, why they get 'upset' about the products that they don't own is obvious: Investing money into a brand is a reflection of your personal opinion and thus a statement of your 'values' (or whatever). Criticism of the brand that you chose (or, more ludicrously, praise of this brand's competitors) becomes internalized as direct criticism of your own choices and/or values. This is insane. Are we so thin-skinned and criticism-averse that we get upset when someone suggests that what we bought isn't perfect? Is it reasonable to throw tantrums when people simply point out facts1?

Now, I'm not completely unsympathetic to these umm, passions. I can understand the behavior in kids, for example - I certainly fell into the trap myself. This is because purchasing decisions are arguably more permanent among kids who only receive big-ticket items on rare occasions. If I decided to ask for a particular brand of PC for my birthday, only to find out a little later that a much better competitor was released, it's likely that I'm stuck with my decision for some time. Every time some tech podcast host points out how much better the new model is, it may cause me child-self to get 'defensive'.

Therefore, it strikes me that these podcast hosts are essentially wasting time and effort trying to appease 12 year olds2. The alternative possibility, that grown adults are 'losing their s$%t' over comparisons of the number of pixels between two random cellular phones, is too terrifying to contemplate.

1I'm dead serious. You should listen to letters written to tech blogs in response to their listing of sales figures. Clearly, pointing out that iPads outsell Android tablets by some margin is 'bias'.

2There's another problematic issue here that could be an entire post on its own and can be thought of as the 'vocal minority problem' in tech and entertainment. The people who write in to complain about 'bias against Android/Apple' are a very specific crowd. They're certainly not representative of the market at large, and it's not even clear that they're even representative of the total listeners of the podcasts that they're writing in to complain about. Therefore, podcasts that cater their discussion to the wants of this crowd become strangely 'out of touch' with what's happening with the market at large. It's fine to have a podcast that discusses HP products exclusively (if that's your thing) but no one would expect to get insight into braoder market trends from such a limited focus. The people who only ever want to hear good things said about their chosen products should find appropriately skewed shows.

Picnics with Smelly Hipsters...

If you're ever visiting the city of San Francisco, I'd recommend taking a walk through Mission-Dolores Park, located between the Mission and Castro districts (a short walk from the 24th- or 16th & Mission BART stations).

I first visited the area when I arrived last fall and immediately noticed the beautiful views of the city (the park is located on a hill above the downtown area). However, it was rather chilly and the place was pretty empty. Things were quite different there last weekend when the weather was gorgeous and the place was packed... with smelly hipsters.

I hope you're able to see just how many people there are in this half of the park. There wasn't an event or anything - this is just how it is. The only reason that the center areas aren't packed with people is that they're fenced off so that folks can use them to play sports.

Now the Mission District is technically the hipster zone of SF, so it's probably no big surprise that they'd be congregating in the park. But the heavy clouds of marijuana and cans of Pabst Blue Ribbon1 interfere with the relaxation of picnic-ing. 

In all seriousness, I've got nothing against hipsters - except the one's who are actively trying to 'be nonconformist together' (if you get my meaning) - and in fact, there were a lot of strange people of all ilks in the park. For instance, there were people walking up to each group of park goers selling 'magic cookies' or cold beer, both of which are illegal in the park (the former also being illegal everywhere, of course).

Here's a shot looking North towards the less densely packed sporting area. There's a group of young guys playing bocce - a fine sport!

There was also an activity going on that I've never, ever associated with public parks: tightrope walking. Apparently, when you're in a place with palm trees, it's common for people to tie ropes between these trees and make fools of themselves trying to perform circus acts. One guy was really talented, I'll admit, but the rest of it was just sorta weird.

Anyways, as usual I've made a Picasa Web Album with a few photos snapped with my iPhone. A park's a park, so there isn't exactly a ton of scenery to write home about. Next time I'll do a better job of photographing weirdos.

1This is not a stereotype, hipsters actually do drink PBR.

Soft- vs. Hard(ware)...

I've now been a Mac user for 3 years. Before buying in to the Apple camp, I used Linux for computational work and Windows at home. My experiences with Apple products had typically been negative (why can't I properly maximize my windows?!?!?!?), and I had a philosophical opposition to their 'closed' ecosystem1. Then one day, based partially on the recommendations of co-workers, but also on somewhat of a whim, I bought a MacBook and became an Apple fan.

I still have a PC for playing the occasional, umm... videogame, but all of my work and most of my 'productivity' applications (such as writing this blog post) are done on my Mac. The reason is simple: I really enjoy using Apple's software. What all of my Mac-hating, Windows-using friends have always referred to as 'dumbed-down' interfaces, I now see as intuitive. There's a certain elegance and consistency of design among software on the Mac, such that you can immediately figure out how to do things on an application that you've never used before. As an example, take these screen grabs from two relatively similar pieces of software:

The first program is Apple's iPhoto.

The second is Google's Picasa 3.

Just looking at these two screen grabs, it's pretty clear to me that iPhoto is more sparse: There's less clutter on the screen, fewer information boxes, and fewer buttons sliders and toggles. Picasa has more functionality, but (and all of this is just my opinion, of course) that functionality is realized by continuously bolting more features (i.e., buttons and sliders) on top of an already clunky interface. iPhoto is also arguably more 'aesthetically pleasing', but again that's (common) personal opinion. 

There are arguments for both views - Shiny interfaces are often undesirable when functionality is at a premium, after all. But I think that this specific example illustrates a larger problem: why is there so often a disconnect between the quality of the hardware we use as compared to the software running it? Or, for that matter, why do companies that produce very functional software (Picasa) often put little stock into developing their user exprience and aesthetics?

Another example: I got an HP printer-scanner combo thing last year with my new laptop and while it scans great images, the software that came with it is ABYSMAL. There are freeware programs that are so much better. You could repeat this example ad nauseum with various products that have come out over the years: digital cameras, cell phones, MP3 players, etc. Why does their in-box software suck so badly2?

Part of Apple's recent success is undoubtedly due to their concerted effort to improve both the quality of their hardware and their software. This is particularly evident on iOS devices, where the ecosystem encourages uniformity of design such that even very different apps 'feel' the same.

So I guess the real question comes down to that philosophical opposition discussed at the beginning of this post: 'freedom' seems intuitively better in some moral sense, but what if the 'walled garden' approach actually produces better and more stable systems? I'd rather have both options available, but now that computing is mainstream rather than the domain of basement-dwelling nerds, is expecting everyone to 'figure every new program out' reasonable? Can Apple be thought of as simply setting 'standards', something that can enhance a market in situations where too much competition leads to consumer confusion? 

1It's debatable how closed the ecosystem of software has been on the Mac/Macbook side in the past, but it's become more closed with the inclusion of the App store on desktops. The new version of the OS coming out this year may close it off a bit more. On the mobile side (iPhone/iPad), the ecosystem is pretty much as closed as can be: You can't install anything not authorized by Apple without hacking the device. 

2We can always say that these are often hardware companies and that software is a secondary consideration, I suppose. But someone, somewhere must've looked at some of these programs and asked 'are you serious?' The software for my old Sony mini-disc player felt like it had been made in a high-school programing course. 

El Corte de Maderas...

Last weekend the gf and I decided to try out another local hiking destination: El Corte de Madera Creek (cutting lumber creek) open space preserve in Woodside, CA:

The preserve is up in the hills separating the San Francisco Bay from the Pacific Ocean, and can be accessed by driving up a steep. winding road. There were a ton of hardcore road cyclists going both up and down this street, in a way that felt more than slightly dangerous... I love biking, but I'm not sure that I'm quite at that point, yet.

There are a lot of trails in the park including a variety of different inclines and difficulties to suit different interests. Unfortunately, there are also a few drawbacks, the most significant of which is that it's not very scenic:

I hope you like this type of scenery, because you're going to be seeing it a lot.

The density of the forest has its own beauty, of course, but serves to block any of the scenic views that would otherwise greet hikers. The redwoods and associated flora are pretty, but not nearly as impressive as other local hiking spots like Muir Woods.

There are a few things to see, such as this sandstone formation. And a small creek...

Another significant negative is that the place is teeming with mountain bikers who just happen to be sharing the same narrow trails that you're navigating. We heard 'coming through!' from behind several times as we walked, forcing us to move to the side and wait for a group of bikers to zip by. Given the number of blind turns and hills as well as the speed at which some of these people are motoring along, it's not exactly a perfect arrangement. Also, you're not allowed to walk dogs in the creek, but you are allowed to ride horses... What gives?

Hey look, a mountain biker zipping by in an area where the trail actually widens! What a nice change...

There are a lot of nice places to hike in the Bay Area. When it comes to this place, as one person put it, 'maybe it's best left to the mountain bikers.' I've put a few pictures into a Picasa Web Album for anyone who wants to see  a series of very similar nature scenes. :-) 

Supplementary Materials...

As the biological sciences are getting more complex, involving ever more intricate analytical methods for dealing with ever larger datasets, reproducing other people's results is becoming more challenging. This is unfortunate, as reproducibility is one of the bedrocks upon which science is built. It doesn't help that certain journals (that shall not be named) have incredibly restrictive page counts and reference limits. At the very least, we can go online and download reams of 'supplementary materials' where the authors can, in principle, describe their work in more detail.

That being said, given that we do have the ability to upload supplementary materials and files, wouldn't it be nice to require authors to upload the formatted data tables that they used to do their analyses (in addition to simply pointing to databases containing the 'raw' data)? I've always provided such tables as supplements to my publications as a matter of principle, but there are a couple of reasons that I think it should be standard.

To begin with, if only the raw data are provided (e.g. microarrays, next-generation sequencing reads, etc.) subsequent authors essentially have to repeat the entire experiment from scratch if they want to verify or check any particular part of it. One could say that that's the 'best' way to ensure that it was done properly, but it's not very practical. Not counting the time it would take to redo everything, not every lab has access to the resources and expertise required to perform every step: For instance, I can download the publically available software used by the original authors, but I don't have all of their scripts, or home-written code to parse and modify the files for each step in the analysis.

More importantly, it's very rare that I'll want to recheck a lab's previous work and far more likely that I'll want to use their results in order to inform my own analyses (this is how science progresses, of course). It's nice for authors to report that 'X number of genes fell into this category', but what would be useful to me are actual lists of said genes. Most of the time such lists are provided in the supplements1, but I've been at this long enough to find that many times, they are not. Is it reasonable for me to have to redo an entire paper just to get a list of loci that were describe and analyzed in the original manuscript?

There's one last specific little thing that I'd like to bring up concerning the raw data themselves. I've noticed that sometimes people will upload a series of experiments to say, GEO, but only use a subset of these for the actual analysis performed in their publication. For example, they may have profiled gene expression levels for 10 different environmental conditions, but only analyzed 8 of them in the paper (for perfectly legitimate reasons). However, because of the particular idiosyncrasies of data labelling, the manuscript will point to the entire dataset, while the dataset itself will not make clear what samples were actually used in the publication2. Again, because we're forced to go back to the raw data instead of formatted tables, this can become a huge headache.

I don't think that these sorts of things are done maliciously - there are better ways of not making you data 'public' immediately. Rather I think it's simply because there's so much to do when getting a paper ready for publication that making nicely formatted supplementary tables simply adds an additional, tedious step. That being said, I think that it would benefit all of us to have to do so.

1The worst is when such lists are provided as enormous PDF files. How is this better than providing a simple tab-delimited file?

2In full disclosure, I've been involved in such issues myself. If the data were uploaded prior to publication of the manuscript it's not easy to go back and indicate what samples were used. It then becomes a game of answering a lot of emails from people trying to use your work.