Justin hasn’t allowed comments on his blog, as far as I can tell, so I’m stuck cluttering up my own blog posting responses to his ah, misguided analysis of mashups.
Six reasons why Justin needs to rethink Mashups (it’s okay, we’re brothers. Once you’ve taken a baseball bat to a guy’s head, mere words are but the ephemeral fluttering of so many butterfly wings):
- First, the easy point: sorry Justin, mashups are quite old in internet time (Wikipedia’s article goes back to September 05). They’re just getting a lot of renewed attention due to the Yahoo! Pipes release.
- Mashups are unreliable? That seems like a dubious critique, since people are mashing on services like say… oh, I dunno… Google Maps? Unreliable? Cmon. Remember that time Google went down? Neither do I.
- Changing feeds will crash the mashup? First off, if the service changes the feed or API, then sure, there’s some down time to rewrite the feed parser or whatever, but weigh that against free data and applications. And don’t you think that Google and Craigslist are aware that their services are used in Mashups? Don’t you think they’d give advance notices of updates? This is also why people are standardizing their feeds as RSS or Atom, so they can be confident that their feed schema won’t need to be refactored. Is there a legitimate reason to expect this is a persistent problem?
- Mashups are transitory? Sure, and so’s any application. All code is perpetually refactored and applications relaunched. Version what dot what? Mashups, like Extreme Programming, plan for change by minimizing the cost to build and update applications. So your mashup is only good for a year before it has to be updated – compare that effort to the savings associated w using Google Maps instead of building your own mapping app.
- Mashups are a travesty or a toy? Tell that to people who use housingmaps.com. I know a guy who planned much of his move from Connecticut to LA with this tool. It allowed him to understand the neighborhoods being advertised and narrow his search well in advance of an expensive trip to finalize his first rental arrangement. A friend at work used the same tool to understand hyperbolic real estate agents’ descriptions of New York hoods when he moved here from Syracuse.
- Mashups provide free advertising? By that logic, the credits at the end of a movie are “free advertising” for the people who made the movie possible.
I hate to say it, but the Madirish post reeks of so many engineers’ build-don’t-buy mentality. Sure, given enough time and resources an engineer could theoretically recreate any mashup, but in reality APIs and data feeds allow projects to achieve usable results with far less effort. Perhaps you say “yeah, but a mashup only provides 90% of the desired functionality,” to which I reply “at 10% of the cost; and really, if we can get by with 90%, let’s get real and move on.” Mashups are not PB&J, they are the European Union – by removing borders, they enable more efficient economies.
What surprises me is that this post was written after we did hours of unpaid work to create the George School campus map when we could’ve probably achieved similar results with a simple Google or Yahoo! Map mashup, at a fraction of the cost. And your concern is that a Google logo would appear on the map? I think that’s a fair trade.
Second, I find your lack of faith disturbing. Are there a ton of great mashups out there today? Perhaps not. But this doesn’t exclude the potential for greatness. We still have some work to do figuring out standards and connectivity, but I think mashups will eventually be recognized as yet another step in the current trend away from atomic applications and data sources towards a truly integrated universe where any engineer can take advantage of everyone else’s brains and sweat to create wealth for all users. Here are other components of this trend:
- Widgets (mon amor)
- Pirated music, Creative commons and open source web design
- Drupal, Rails, Scriptaculous, and anything else on Sourceforge
- Offshoring
- OpenId
- Mechawalruses (more on this later)
This is a real trend, and part of the evolution is letting go of the need to control the source code and trusting other engineers to share the load.
Update: A couple more interesting stories about the value of exposing services to Mashups and Widgets:
- Widgify, from Snipperoo:
As the web continues to become disaggregated, there will be a burgeoning demand for tools that can help users effectively leverage these “information atoms,” together in a meaningful manner. Not only will there be a need for tools that help users aggregate widgets, but also tools that enable widgets to work together.
- Social Networking Sites Open Up, from Businessweek:
Social-networking sites are realizing that if they want to grow their user base—and build a sustainable business model—they need to attract third-party developers. “Social networks have reached a point of maturity, and opening APIs will help them grow,” explains Adam Trachtenberg, a senior manager for eBay.
And that, Justin, is why the data sources will prioritize uptime and consistency: because they benefit from Mashups as much as the engineers do.
Updates: Justin has replied (with some good points and some lamerz), and I’ve fixed the oblique star wars reference.