"Mashups" are a new form of creative collaboration. The word refers to certain music and web applications.
According to Wikipedia, a mashup is "a website or application that combines content and/or functionalities from more than one source into an integrated experience."
Many sites combine material from multiple sources. The distinction with mashups is essentially that they do so dynamically. For instance, the mashup may update soon after the originating site updates. And the mashup may enable the user to select the output, such as for a speficic ZIP code.
Programmers work with an application programmer interface (API) of existing services and products to create new hybrid applications. Software providers are often often eager to make their APIs public, in hopes of wooing developers to make new mashups, thereby increasing the popularity and value of the original product.
EXAMPLES
For example, a popular mashup is Virtual Places which "mashes" (meshes, combines, blends):
Amazon + Alexa + FeedMap + Flickr + MapPoint + WindowsLiveSearch + Virtual Earth.
See the mashup listings at Programmable Web.
http://www.programmableweb.com/mashups
Web 2.0 Mashup Matrix: hover your cursor over points in the grid to see a drop-down menu of known hybrid applications in development.
http://www.programmableweb.com/matrix
Some of the most popular types of mashups include:
Some sites help enable people to make mashups, typically with maps. Yahoo Pipes is more comprehensive but also more difficult to learn. Another example is Grazr.
ARTICLES
Business Week "Mix, Match, and Mutate" article explains how mashups are changing the internet from a static library to a participatory workspace and interactive playground.
http://www.businessweek.com/magazine/content/05_30/b3944108_mz063.htm
The Economist "Mashing the Web" article (paid subscription only):
http://www.economist.com/displayStory.cfm?Story_id=4368150
Business week wrote an interesting article on music community mashups. The idea behind music mashups is to combine the instrumentation of one song with the lyrics from another, or creatively blending two entire songs into one new work.
While recording companies have been resistant to people using intellectual property they own, and remixing or refactoring it without permission, this attitude is basically anti-consumer and self-defeating. Mashups are generally distributed free, so there is no commercial gain sought. In addition, mashups can help boost sales of the original artist recordings, igniting new interest in product that may be experiencing declining sales.
Wikis could promote new applications of existing but "high activation energy" software programs by lowering the barrier to error-free input.
When wikis are deployed as the "friendly front porch", wisdom from mass opinion can be quickly compiled as if taking a poll. This can be done on topics such as the technical opinion about whether a new product would be difficult to produce, or if a new drug might have side effects based on its parent chemistry.
The output from the wiki need not have an embeded link to the other software since this output could be simply typed into the second program by a skilled user. This avoids the need for those providing their expert technical opinions to be required to operate more complex and hierachical systems.
Let's now refer to this non-hard linked melding of a wiki (Network) with secondary Enterprise or Functional software as a "mashboard", giving due credit to Andrew McAfee for distinguishing these three types of corporate software. While the "mash" part is the coupling of two softwares, the "board" is as in "dashboard", a grid or matrix that can be filled out and then studied for errors and edits on a single table or page by a pool of experts.
A very close-in example is using a wiki with SAP, and let's now provide a simple notation for the software synergy, N->E, meaning, Network software used to provide inputs to Enterprise software without having these electronically linked but sharing a table or matrix as the output of one into the other. SAP requires that input into data fields be "Six Sigma" (one defect in 10,000) free of errors because errors will instantaneously be propagated throughout the interdependencies of the underlying matrices, making corrections arduous.
Before such data entry, a corporate wiki could be used to "bleach" the data free of error by exposing it to the "sunlight" of the whole organization. The search features of the SAP wiki and its e-mail alert features would pull expert eyes to the proposed data table for some appropriate duration before entry into SAP, this greatly reducing the potential for erroneous entry.
Let's now provide a bit more obscure example, but perhaps a more innovative one. Since the advent of the computer in the early Sixties, statisticians with techncial and marketing organizations have attempted to apply Bayes Theorem to predict future success or failure. Before that, the computational power required to crunch through the probabalistic calculations was intractable.
When properly constructed and continuously updated, Bayesian Belief Networks (BBN) can diagnose diseases, predict the safety of a new drug, and generally improve the risk analysis for any new venture with less frequent and costly-to-accumulate data.
Perhaps the greatest hinderance to creating a BBN is the need for "priors", or put another way, knowledge before fact. Of course, the best source for this is expert opinion, and now the reader can mentally apply the wiki to collect these priors from an appropriate "crowd", and our newly stated notation for this approach would be N->F, where N is our wiki and F is the Functional, Bayesian software of which there are many commercial and Open sourced products. Another useful part of using wikis is to also "compile from crowds" the set of possible outcomes for each node (Example: Compile from crowds) in the Influence Diagram, for the prior probability of every possible outcome is required before the model can be completed.
Finally, the real power of BBNs is that they can be "taught", given updated data, and by this refine their predictive precision. As time goes on, and casual events pass, the models simplify themselves because some possible events did not in fact take place. Thus, a wiki running concurrently with the Bayesian simulation would become the "basket" for updating the model. One obvious application would be a test market for a consumer product where the chances of successful commercialization would be revealed when the wiki contributing test group collectively achieved the needed threshold to expand or terminate the product introduction plan.
The author apologizes for any expected but inadvertant errors with respect to Bayesian statistics, and welcomes corrections. However, to use this insensitivity to illustrate the proposed application, those adept at constructing and presenting the wiki to the "crowd" need only understand the specific fields of inputs needed by the BBN and its technicians to make this "mashboard".
contributed by kaboboom on Feb 25 6:44am
Page Last Updated: Apr 29 3:46am by maurreenskowran