The End is Nigh

Okay, not really - but I wanted to make a brief note that soycode.com proper will be getting a facelift of one sort or another shortly. It is one of the (perhaps few) pages around still powered by Google Pages, which is being replaced by Sites (which of course powers the separate Soycode Wiki), and is one of the even fewer sites that uses Pages to host files that won't work in Sites (custom JavaScript/CSS/even Java applets).

So, I'm looking into replacing it with some sort of App Engine powered CMS. There's a few out there, so I'll likely just grab one and do minimal customization just to get it suitable for my purpose. If I end up doing anything interesting with it I'll of course note so here.

That is all, thanks for reading you nonexistent reader you.

Yes, Bing has no bananas

Fear not, person incapable of forming coherent search queries - the decider is here.

At least, that seems to be the message of Bing. Well, the more sensible of the several messages - Bing advertisements also claim that search engines are somehow responsible for the recent economic collapse, inexplicably linking what is arguably the most significant advance of the past ten years to all of the bad things that happened in the last ten years.

But never mind that - let's take Bing on its merits and not its marketing. Consider a simple test - you want to know how much a banana weighs. A simple property of a common object, admittedly it's not a common question but a search engine ought to be able to handle it nonetheless.

Well, compare and decide.

To my eyes at least (and as of the time of this writing), the only search results that clearly and quickly answer the question at hand are those from Google. It's clear from the snippet that the resulting page contains the answer. This is true if you form the query a few other ways as well - if you try 'weight of a banana' the wiki.answers.com result that actually has the answer peeks into Bing at #10, but still #1 at Google.

What does Bing (and to a lesser degree Yahoo) give you instead? Mostly dietary results - it seems to assume that you want to lose weight, and that you're thinking bananas may somehow facilitate this. Admittedly I could imagine a less savvy searcher having this in mind and conceiving of a search like "banana weight" rather than "banana diet" (or "banana weight loss" or "banana weight control" or something), and it seems to be those users that Bing is targeting. A lot of people probably look for information about diets online (more than who care about the weight of produce), and so Bing has just boosted it up for everyone whether you happen to want it or not.

However, I think this would be a step backwards in search technology. To me, search is a tool - it requires proper use, but the payoff is a great deal of power (and really, it's not *that* hard to use). If I want to search about a diet with bananas, great - but if I want to do a stranger search for the weight of a banana, then that should work too and yield different results. Frankly, I don't want something to make my decisions - I do that. I want something to provide me with information that helps me guide my decisions.

If we start weighting all searches heavily towards the most popular queries, the results may be disastrous. Obscure information will become even more difficult to find, while the great morass that is popular culture will expand to consume even more. Sure, most users will be happy most of the time, but that's not the only thing search is about - search is about indexing all knowledge, even the stuff that is obscure or even not clearly "monetizable" (*gasp*).

Of course, this is just one example, but combined with their marketing and the fact that they generally are trying to "tune" top/commercial queries (things related to travel and shopping), Bing seems to be aiming at the lowest common denominator here. Will it pay off? Maybe - I'm sure there's some proportion of users who will prefer it. I just hope that it's not a substantial enough portion of the market to cause Google to "dumb down" in order to compete.

Oh, and by the way - Wolfram Alpha gets this particular query right on.

(Coming next week - when you give a Bing a cookie. Okay, not really.)

Thoughts on Rhomobile

I've spent a wee bit of time getting my feet wet with Rhomobile these past few days, as I mentioned in the previous post. Specifically, I got an account at Rhohub, a "hosted development environment" for the Rhomobile platform.

Put simply, Rhomobile is an attempt to make it easy to develop for all the major smartphone platforms (iPhone, Android, Symbian, Blackberry, Windows Mobile). You write it once, doing the design/layout in HTML/CSS and the logic in Ruby (all pretty accessible stuff), and then create natively installable packages that can use advanced hardware functionality (camera, accelerometer, etc.). The idea is that you get the ease of a "webapp" (e.g. a webpage tailored for mobile use) with the power (and maybe speed) of a native application.

The main catch? Every Rhomobile app needs to be packed in with requisite Ruby libraries, meaning that even a tiny program (a few kilobytes of code) will end up needing 2mb or more of phone space. Not a huge deal for iPhone and higher-end Nokia's, but definitely a significant limitation for the current T-Mobile G1 and many other smartphones that have limited space for applications.

But really that's not a horrible catch, and I think we can expect phone storage space to increase at a faster pace than Ruby library size (or at least I'd hope so). So, I'm quite enthusiastic about the possibilities of native cross-platform smartphone development, though as of now the documentation (particularly for Rhohub) is somewhat lacking and so I haven't actually done anything substantial. I may at least show a "Hello World" type app here later, so stay tuned and thanks for reading!

Addendum - Adam from Rhomobile let me know that the minimal app size is actually not as severe an issue as I thought: "Regarding app size, for what its worth, the most recent Rhodes-based AppStore apps that I've seen is On My Means which is 1.4MB. The "minimal size" is significantly smaller than 2MB. But I'm not sure if I've ever seen a Rhodes app larger than 4MB."

This blog is not dead!

(I'm so glad this blog is not dead, etc.)

Okay, a quick post to the nonexistent readers out there, really more for posterity and a self-reminder than anything else - this blog/site is not dead. It's stale and inactive, but not dead.

The site proper is powered by Google Pages of all things, which is perhaps even more defunct than the content, so I'm going to replace it with an App Engine powered CMS (I'm looking at this one). This will hopefully grant some more flexibility, which will in turn lead me to actually put more stuff up.

I intend to work more on Soy Statistics - I actually have a v0.11 sitting around (few minor bug fixes and tweaks), and want to push ahead with my planned roadmap for 0.2, 0.3, etc. However, I'm also thinking of trying to document/write my actual statistical thoughts a bit more generally (more on that perhaps in a future post), though the projects may complement each other nicely.

Other things - I've toyed a bit with a variety of mobile phone development ideas, and currently am looking mostly at Rhomobile. Of course, I haven't actually *built* anything yet, but stay tuned (and if you don't know about Rhomobile go check it out, it's pretty cool - multi-platform smartphone development in Ruby).

That's it for now - expect a bit of a site cleanup first, followed hopefully by some updates to Soy Statistics and maybe some mobile phone stuff if I get some inspiration. Until then, keep on not existing you nonexistent reader you.