Usage: Hover the cursor over any cell in the matrix. A small box gives details on mashups for that API combination.
Top links in hover box bring you to that API's reference page.
Links in body of hover box take you directly to the mashup.
Not all combinations have mashups & only those with the 'º' indicator currently have entries.
Cells at the intersection of same API (ex: Amazon+Amazon) list any other examples for that API.
Note that there are two views into the matrix: the default view shows only those APIs for which mashups have been added to the
database. The second view shows all APIs regardless of whether there's currently a mashup registered. It's big.
Definitions: What is a mashup anyway? As always, it's good to check
Wikipedia's definition,
but essentially a "mashup" is a web-based
application built through (creative) combination of data from multiple sources. Often, but by no means always,
this data is retrieved by using a vendor's API such as those listed here.
(An API? Also at Wikipedia.)
Some recent press may also help explain:
BusinessWeek's "Mix, Match and Mutate",
The Economist's "Mashing the Web".
Background: This is an experiment. It is intended to be both a reference point and also a visualization.
What you see here today will change both in content and form shortly.
I am quite interested in seeing the 'space' in which mashups exist.
Clearly, some APIs such as Google Maps, appear to be more widely used than others.
UI Issues:
Cross-browser support is good but not complete.
Sometimes it can be tricky to 'catch' the hover window (but you can also click on the current cell to go to that intersection).
When using the 'crosshairs' feature the performance can lag. The HTML and JavaScript can be optimized.
Other Issues or 'points of debate':
Completeness: this list is not comprehensive (but feel free to add more).
Defintions: The definition of what constitutes a mashup is not 'concrete' and since it's so new will certainly evolve.
Gray areas: Sometimes a mashup uses APIs, sometimes RSS, or XML, screen-scraped HTML, or 'internal' data that's not publicly accessible. How should these be represented?
Also the one-offs or one-hit wonders, they fall under the 'Other' bucket for now.
Site distinctions: I'm not entirely happy with my initial division between an Example and a Mashup (the idea being that an Example was someone's use of a single API and a Mashup was more than one). But this doesn't entirely work.
Future: Since this is v1, there's plenty of room for improvement (presuming it's useful).
Needs to address some of the issues above. Can there be a meta-mashup?
From a visualization perspective, I think there are more interesting things that can be done (ala Tufte): quantities per cell, sized graphic per cell, popularity, classification, age, utility.
Other Notes or Trivia:
The etymology of the term mashup is assumed to derive from its
musical/DJ community usage when two or more songs are mixed into a single track. People don't yet agree: is it "mashup" or "mash-up"?
Technical nitty-gritty:
The matrix is derived from this site's API database.
The HTML is generated nightly by a process that extracts the most current data: APIs, Examples & Mashups.
The dynamic table hover is a based on some ugly modifications to Walter Zorn's excellent JavaScript library.
Disclaimer:
Everything here is 'best effort' and is not guaranteed to be 100% accurate. Please refer to respective parties to verify.
Particpation:
As always you can contribute examples and mashups here, and are welcome to provide your feedback and ideas.
