Mashup FAQ - Frequently Asked Questions - Java

What are some of the ways we can address the security issues and concerns related to enterprise mashups that leverage external data sources?
Are you putting Excel in the cloud, when you say repository? Where is this Repository? Google Docs, SharePoint?
What are the differences between Presto Wires and Presto Mashup Studio? Do we need both at the same time or is Presto Wires sufficient for our needs?
Can I use unstructured data and a type of input for a mashup?
What type of effort does mashup creation usually entail?
Is the data that is output by this system, universal? Can I take the data and make more mashups on a variety of systems or does it only work with your software?
So the only practical limits to what can be used in a mashup setting are the limits that the enterprise attaches to accessing its data, right?
Are there ample number of adaptors for connecting to legacy systems?
Can you only pull live data in via XLS files, or from SOA, JDBC, API, etc. and if you can, how?
So, the mashups are a way for an individual in an enterprise to get what he wants from IT which he couldn't get because building an app for a single instance isn't cost effective for IT. But then you say that once the mashlet is built, thousands of users can use it. Sound like circular justification.
Process was mentioned; are there flow objects in the wire tool for workflow type conditions between data steps?
Are mashlets output as JSF Components?
What language does the mashlet use (flash, Ajax, etc)?
So the Wires env is the runtime and each user will need to have it installed ? Is there a runtime env that runs in a browser ?
Does JackBe integrate with enterprise solutions like LDAP,AD etc ?
Do all capabilities need to be a service? Can you bring in capabilties from an application (i.e. Sametime)
Would like to see it listening to events via JMS.
JSR-168 is nice, but it doesn't help me as an ASP.NET developer. Is there support for displaying mashlets in ASP.NET apps?
I guess the Silverlight question is really: do you have an API that allows me to do that binding? Or do I need to do it myself at this point?
Can we place these mashups in portal pages/pda, if so, how much customization do i need to do?
Can I display a mashlet within a Silverlight control/application?
Can services be prepared from pretty much any data source including fixed length text files?
What if I need to do a string manipulation on Job_ID, in order to match the IDs? e.g., if one source was JobID=00011 and another field was a taxonomical derivative of JobID such as PositionID=CsCI00011, where I need to strip off CsCI in order to make it match. Can I call a java method to strip this?
Can we do mashups on process?
Are there any limitations to the amount of services that you can join together?
Can you show database integration?
What is the deliverable to the client? A mashup? Mashlet? I'm trying to understand how the application dev process changes in a mashup architecture.
Can you please review what the outcome are of the mashup? RSS / GRID / CHART anything else? .. HTML?
Is a Visual Studio plugin also available?
How the Mashups are being handled for different devices display capacity (PC, iPhone, Blackberries, etc)?
Can widgets only be viewed on web browsers or is there support for programmatically accessing widget data from .NET or Java?
How does the protocol (transport or security) conversion happen? Is the developer responsible for it or does the mashup platform take care of it? What i meant was when we talk about using the same mashup on different platforms/channels.
Does this technology integrate with any of the commercial ESB vendors (TIBCO, IBM, Oracle Fusion etc)..?
Could you try and differentiate the mashups from traditional portlets?
You said this works with the existing security. Does it work with LDAP? Does it work with Active Directory?
How do you assess the accuracy or reliability of the source data. How do you know you aren't being fed false data? What ESB security are you using.
Transactional cases: I looked into the links you gave about usage of mashups inside eGovernment applications but I haven't seen any transactional ones. Could you confirm that most of the mashups applications are 99.9% non-transactional ones?
How do we calculate the ROI of doing Mashups? Most of time, it's difficult to sell this concept to CIO.
Have you seen successful mashup deployment on Force.com platform?
You briefly mentioned governance, how critical is it that the organization has a cross Biz Unit Enterprise Architecture body to enable (publish) access to a data silo(s) somewhere within the enterprise?
How would you compare or link J2EE patterns with Mashup patterns?
In the examples on competitive analysis, were APIs used to get data from different sources or was data obtained via HTTP and then parsed? I'm guessing a lot of retail sites don't necessarily provide APIs to access their price data.
Regarding ‘api enabler’...are you using the browser source or more of a screen scrap for creating an external website API?
Say we have a legacy app that is written in VB code that wraps multiple web browsers. The web browsers actually call different java ear files; plus the VB code invokes java web services via SOAP. Could we "rewrite" this app as mashlets using the basic underlying java services & publish the "package" to a browser? Or basically simplify this legacy app plus make it customizable by adding new mashups?
The value (and potential risk) of disparate information sources may be larger than the sum of the parts. (your competitive analysis pattern) Are there any design patterns (or tools) for service providers to understand how their information is being used (in perhaps anticipated ways).
Are their best practices for a template on how to organize your mashups? (Not patterns but a template to be used for enterprise mashups)
It seems mashups won't grow to its potential until tools are available for the business users to easily create the data combinations they value. What are the best business-user tools available today?
If custom 'mashlets' have to be developed, what's the benefit of developing a mashup versus developing a web application using traditional tooling? Are there productivity benefits you've seen?
What are the best practices to mitigate website saturation if one or several mashups are hitting the same source website?
What about how reliable an external resource can be if it's not under your control? What if Google suddenly changes their api or interface or eliminates it?
Standards are important to enterprises as well. Are there efforts underway on this front w.r.t. mashups?
What is the future/scope of "Semantic Mashups" w.r.t Semantic Web?
Will you get into how Mashups and Workflow and BPM work together, risks, issues, etc...?
Some “mashups” have ability to support some business logic (eg: scripting languages)… are these considered mashups?
How is the challenge of data sanitation and denormalization addressed in JackBe?
We cant use XML, a programming language that originated on the internet to give the information metadata, and then use RSS and other languages to stream the data via syndication into one data aggregator.
Are enterprise Mashups portable to different web/application/mashup servers?
What is the difference between a mashup and a Portal? Don't they both just collect disparate data and present it in one central area?
Can I display the mashlets in a Silverlight application?
Does Presto integrate into other portal platoforms, e.g. Webcenter?
The mashlet UI support OOTB is simply but has limited features, right? e.g., it's not possible if the users want to define some drag and drop features with JavaScript?
How do you compare the Enterprise Mashups with Portals? Do Mashups replace Portals?
Say we have a legacy app that is written in VB code that wraps multiple web browsers. The web browsers actually call different java ear files; plus the VB code invokes java web services via SOAP. Could we "rewrite" this app as mashlets using the basic underlying java services and publish the "package" to a browser? Or basically simplify this legacy app plus make it customizable by adding new mashups?
Is Mashup driven by AJAX?
How does this compare in construct and operation to the Portal/Portlet style of web page composition.
What charting package(s) do you use? Is there an available list of the specific chart types available? Is it possible to extend the existing chart set with 3rd party charting packages?
Is it required to compile the mashup separately for its deployment into .Net and J2EE containers? Does it require separate code base?
Are the Presto services exposed to other systems such as OpenESB?
Can Mashups pull specific portions from Seibel eCRM, both from Seibel portal and/or data inside Seibel (oracle database)?
Are you using .NET FRAMEWORK 3.5 with AJAX?
Can we mashup two different databases...like SQLServer and Siebel ? I mean is it dependent on a particular technology?
Will it cost a whole bunch to connect to SAP systems?
Excel "apps" that is scalable, maintainable and secure could be built with MOSS (Microsoft Office SharePoint Server) deployment and give the power to end user. Why should we embrace Mashup?
What’s the difference btw Enterprise Mashups and an ETL solution?
Do you have connectors to sources such as OLAP cubes?
How are mashups different from dashboards that have been around for a while?
Can you have input fields in the Mashboard that feeds multiple mashlets on the dashboard?
How much does the use of mashups increase storage need and capacity? It is obviously inherently dependent on how many... and how long they hang around or are for re-use... but is the impact huge?
For real-time reports, how is Presto handling the high volumes of information flow in terms of pagination of detailed reports or long waiting time-outs to generate a chart? Does it have timing / summarization parameters?
Our firm is not going to go out and scrape info from the web. How about an example where internally we can mashup mainframe data, existing java webservice calls via SOAP, and data from an excel spreadsheet! Is this possible with the mashup platform?
Mashups seem well suited for information centric / lightweight BI applications, but how do these solutions evolve over time or provide enterprise capabilities for inclusion of more complex business logic, or support for business processes, rules, event processing, etc.
I don't understand the BI, we know that excel is THE solution used by the business users to do their job. I have difficulties to see mashups helping in any way?!?
How does a mashup compare to business intelligence?
Will the future of BI be mashups rather than the traditional data warehouse, marts, and access tools like cognos.
Can mashups access historical data to show trending, or do aggregations for large volumes of data?
Can you only pull live data in via XLS files, or from SOA, JDBC, API, etc. and if you can, how?
Most organizations are currently in the ‘wild’ SOA stage; is it required to implement governance and SOA policies before adopting Mashups successfully?
Sounds like MashUp is yet another well-tooled layer. Layers get born (or just get named) when the popular layer de-jour fails to be the golden solution and is becoming silo-ed. Mashups, as I see it, are subject to the same trends and, possibly, the same fate of being over-ed by yet another layer. An environment with a reasonable SOA implementation may claim MashUP to be just another massive SOA service.
I don't see any difference, so far, between SOA and mashups. Many SOA vendors also provide Portals to work with SOA services. (and business users who know the data are responsible for their services).
What level of SOA maturity is required as a prerequisite for effective enterprise mashups?
Hi, Presently I make use of Mashup Studio(Eclipse Plugin) and write my EMML scripts. Later I publish these scripts as Mashups that shows up in PRESTO Services/WIRES as Mashup. Assume that some time in future when our application goes LIVE(to production) and while working with PRESTO SERVICES if someone accidentally remove the service (mashup), then do we need to use the same methodology of publishing the Mashup from Eclipse or are there any alternate way of publishing mashups to Presto? Note: The other possibility could be to do migration of data from say staging repository to production repository of Jackbe. Regards, Guruprasad