Essay: What We Can Learn from Geocities

I’ve been pondering on the now-defunct Geocities, a free web-hosting service that was ubiquitous 15 years ago. It might seem like a failure, at least in light of it shutting down last year (Japan being the exception), but its existence was integral to the popularity of the Internet. Four commodities current users take for granted are free emails, domain names (to a certain extent), storage space, and bandwidth. Without Geocities, the last three wouldn’t be free to this day (which isn’t to say Geocities was the only free web-hosting service; there were several but you can’t argue that Geocities was one of the most popular).

Perhaps one of the reasons Geocities eventually shut down is not because it didn’t innovate (which I’ll get to later), but because it failed in becoming simple and accessible to lay people. The success of sites like Blogger or Flickr for example isn’t that they provided free domain names, storage space, and bandwidth, but that they were easy to use. I could create a coherent and aesthetically-pleasing blog in under five minutes with Blogger. Same goes for photo albums with Flickr. Geocities was too general and too customizable for the non-techie, at least when it competed with the more focused and specialized services. Want a photo blog? Go to Tumblr. Want pure storage space? There’s Megaupload.

Despite that, Geocities did learn from some of its shortcomings. One element I’d like to emphasize is Geocities domain names. Originally, when you signed up for their service, you didn’t get a website like www.geocities.com/charlesatan (this was their eventual epiphany). Instead, what you got was something like www.geocities.com/Tokyo/Towers/0001. In retrospect, Geocities’s naming scheme seems counter-intuitive (I’d like to note that it’s working for Japan though) but it’s understandable why the designers originally planned Geocities to work that way (and why they’re called Geocities in the first place). And it’s a mistake many people make in light of new technologies, such as eBooks:

Creating a platform that’s identical to it’s real-world counterpart (even when it’s not suited for it).

Now in the case of Geocities, they thought that URLs, or web addresses, should be similar to the way we have addresses in the real world. You get a “city” name (Tokyo for all things Japan, Area51 for all things science fiction, etc.), a “district” name (in the case of “Tokyo”, towers, flats, temple, ginza, garden, pagoda, etc.), and a numbered address (0001~9999).

To be fair, this isn’t without its benefits. Just looking at the URL for example, you can tell a lot about the website. Similarly, finding related websites is easy by just changing the variables (the numbered address, district, etc.). In fact, this is how Call Numbers work in a library.

Unfortunately, that’s not how users interact with the Internet. People navigate either through URLs directly or through links. Someone doesn’t think, hey, if I change one of the letters in Microsoft.com, maybe I’d arrive at Apple’s website.  Nor is the inevitable long URL of Geocities memorable (sure, they might remember “Tokyo” and “Towers” if they’re into anime for example, but remembering a four-digit address for your website specifically is asking a bit too much; how does anime fan 1813 distinguish themselves from manga fan 1764?).

So one of the eventual changes they made was that your website url was that of your login name (i.e. www.geocities.com/charlesatan if your login name was charlesatan). This practice has been appropriated by most, if not all, free web services today.

The lesson here is that while real-world analogs can be helpful at times, one shouldn’t be constrained by them. For example, a podcast isn’t simply radio on the Internet (for one thing, you can pause, rewind, and replay podcasts upon demand). Neither are eBooks simply print books uploaded to an electronic device.