Comments on: New Product Category: IBM announces a “mashup server”

By: Ben Watson

Ben Watson — Tue, 27 Jun 2006 23:28:20 +0000

The question is who needs it…for those who are already mashing up on their own server, a new product may not be required. However, let’s assume mashups moveup, shakeup, and wakeup the big bad million-liner Java or C++ camps. They would have room for the mashup rack in the budget, and they would be looking for someone to point the fix IT finger at, so they would download, install (whew! that was hopefully easy) and they could be mashing up later that day, happily still invested in the IBM “we got you covered” approach.

fyi - Flex has the tooling right, and anyday now…

By: Matt MacKenzie

Matt MacKenzie — Fri, 23 Jun 2006 23:38:54 +0000

I posted about this over on my blog. See: http://www.mac-kenzie.net/blog/2006/06/20/mashup

By: Dennis Howlett

Dennis Howlett — Thu, 22 Jun 2006 08:03:21 +0000

Have you seen any of this stuff James? From reading the release, it sounds like stand alone apps.

In the real world, value may well be derived from that style of app but the power comes from being able to act on that information in the context of the job at hand. I’ve not seen any of THOSE types of example.

By: Scott Dietzen

Scott Dietzen — Wed, 21 Jun 2006 21:57:41 +0000

I’m not quite sold on the notion of a generic mash-up server, mostly because I think mash-ups are more powerful in the context of an application, but I think adding generic mash-up capabilities to applications servers (e.g., .NET, WebLogic, WebSphere, JBoss, Tomcat) and Ajax Toolkits (e.g., Kabuki, Dojo, Google Web Toolkit) makes sense, assuming we can get the abstractions right (see below). Doing so would make it easier for the applications hosted on such containers to mash up other applications. For example, if more mash-up capabilities were provided by the underlying technologies, that could have simplified the work we had to at Zimbra.

I’m still worried, however, about how much of the mash-up infrastructure we can make generic—much of what is needed seems inherently application specific to me. Fundamentally, I think the mash-up is more likely to remain a design pattern than become pluggable infrastructure.

By: Andre Charland

Andre Charland — Mon, 19 Jun 2006 00:29:13 +0000

speaking of marketing 1.0…are these mash up servers anything like the infamous “universal business connectors” of the late 90s early 2000s? the pipe dream of IT heads everywhere. afterall a true “mashup server” should enable many to many mashups where salesforces enables integration with their one app to many others…not sure that deal with zimbra is just yet. (although I’m interested to hear your experience with it James…I’ve had it with ms exchange!!!!)

completely agree with the business modeller pitch, i don’t many programmers who can even do that:P

By: Tim Matthews

Tim Matthews — Sat, 17 Jun 2006 00:49:55 +0000

While we’re not calling our product a “mashup server,” this is something we’ve been doing for a while now, and have been calling “data mashups,” which I wrote about in a recent post http://blogs.ipedo.com/integration_insider/2006/04/data_mashups.html