On Robots, Automation, APIs Opencaching and Groundspeak

I Am Not The Droid U R Looking 4

Groundspeak is shooting itself in the foot, and with the new opencaching.com website, sponsored by Garmin, it’s time they stop being selfish and open up their API … and they better do it quickly.

dor.ky blog, by Scott Wilcox, has a great explanation and a few good comments on why groundspeak needs to open their API.

The forums at groundspeak on the topic pretty much ended with a comment from Jeremy (doubtlessly employed at Groundspeak) stating:

There is a private API and it is available for trusted partners. We have no plans to offer an open public API at this time. It was considered in the past but the current decison is to keep it private.

Sounds short-sighted, if you ask me – especially considering that this comment was made August 2008, nearly two and a half years ago and they still haven’t opened their API.

Why does this interest me? Support tools by fans would be an enormous benefit to Groudspeak. I would like to develop an app or even a web-page that helped geocachers manage and automate their virtual caches. Imagine, as a geocache owner, that you are able to provide a form for people to simply fill out. Once done, it allows the geocacher to log that they found the virtual. (Better yet, it even automates their log post if they have the right answers!)

Imagine, then, the service growing from that idea, allowing earth caches and history caches along with the virtual caches … all automated.

Currently, Groundspeak not only doesn’t provide an API, but their site license prohibits the use of spiders, robots, scraping, etc. That’s understandable, but they should create an API. Google’s done it and see how they’ve grown. Flickr’s done it, too… they’re going strong despite the competition. It appears that any strong data-driven company has an open API. People want to automate process, they want to enhance the experience.

As @williamtm commented on the dor.ky blogpost Why GeoCaching.com needs an API

If an API was made available, and a company/developer like atebits, TapBots, or any number of the excellent developers out there decided to make a Geocaching app, doubtless it would be brilliant. For starters, you can dump the ugly custom UI, improve the speed of it, and, as you said, improve the reputability of it over 3G/EDGE/etc, and bam, you might have an app that’s actually worth £5.99.

Fact is, the app as it is now is just terrible. It rarely seems to work when I’m out and about and want to find nearby caches, and often simply visiting the website using Safari and finding caches that way is faster.

At such a critical time when new geocachers are nudged into other offerings, it’s the combination of a variety of good tools and access to data that will win the race. Opencaching.com already has a public API out and ready for anyone who wants to develop a kick-a$$ iPhone app for them (and I would be one of the first ones to buy it)… Is geocaching.com up to the challenge?

Not Yet Flickr Favorites

Every once in a while I come up with what I think would be a totally cool idea. Maybe one out of twenty of those times do I actually spring for it.

Flickr is amazing. It’s API is moderately impressive, too. I thought: “What if I could go through all the tags of all my favorite photos up on flickr, ordered them by popularity, then did a search for photos that matched the top X tags in that list?”

I expected to discover some incredible art that would be right in line with what I already enjoyed. BTW – what is it that I enjoyed?

First, I discovered that flickr can only handle searches up to 20 tags. Any more than that and you get back zip – zilch – nada. Doesn’t matter if any photo in their database actually contains all 21 tags – you get nothing back.

Second, I discovered that some people out there have over 150 tags on the photos of which only 2 tags might apply. As a result, when searching by relevance you get these over-tagged photos in the list along with the worthy ones.

Blame the upload tools. They don’t ask the user to tag each photo, but rather to list out all the tags used for the batch of photos being uploaded. Non-savvy users might batch upload a photograph of a donkey and another of a stop sign. Given only one box for all photos to list their tags they would enter something that places stop-sign related tags on the donkey and vice-versa.

All the same – try the tool out, it actually does a shot-gun result of what I hoped for. Try sorting by “interestingness, descending” for the most polished works first, then try the others for works that are “sleepers” (fantastic photos that are uploaded while people are dozing off, and therefore rarely discovered).

Not Yet Flickr Favorites