[Cross-posted from http://blogs.jackbe.com]
[As a preface to this blog, I want to give my apologies for being away from blogging for the last few months. I’ve been busy creating Enterprise Mashup software here at JackBe. Now that our software is production-ready, I’m looking forward to getting back to posting regularly and getting your feedback on my development efforts. Deepak.]
I was at the Mashup Ecosystem Summit organized by IBM at their offices in San Francisco last week. Our CTO, John Crupi, and our Chief Architect, Raj Krishnamurthy, also attended with me. It was an interesting mix of people from different backgrounds and companies all converging on the concept of Mashups. Jeff Nolan (ex-Teqlo, ex-SAP) gave an interesting talk about his experiences in a starting up a mashup company. Some notable points were: (lack of) availability of APIs; Do-it-yourself Data Formats; Performance can be a challenge; Need for strong visual composition tools; Lack of Standards. I think these are questions that this group will be able to tackle over time. (At least, I hope!)
At one point during the end of the meeting, someone asked: “What really is a Mashup?”. This led to a brief but inconclusive discussion. Which goes to show how nascent this field is that even among the experts in the industry, there is still some uncertainty about how to define and qualify a mashup. (Coincidentally, JackBe has a webcast coming up on May 23 on this very topic: ‘Enterprise Mashup Bootcamp: What, Why and How’. You can register for that webcast here.)
I managed to slip in a question at the Mashup Summit about how the group feels about Composite Applications which were a hot topic a year or more ago and how they relate (or not) to mashups. Predictably, there wasn’t a big desire for this discussion. However, my own personal take is as follows. We in the SOA software industry have been busy implementing SOA in the enterprise over the last few years in our architecture and IT infrastructure. This effort in SOA has largely ignored the end users, and mostly focused on the IT and business stakeholders. Composite Applications are, in a way, IT efforts to provide integration of data, services and processes. While there has been a proliferation of services around the internet/intranet, users have no good tools to use them to do their job more effectively. Therefore, end users do not see much direct benefit of SOA or services or even composite applications.
These SOA/services/composite applications efforts are now undergoing an (r)evolutionary transformation. Enter mashups. And enter the users! The integration is now happening at a higher level in the application stack, much closer to the user. (This is also what Dion Hinchcliffe points out in his blog, where he talks about the 5 styles of mashups.)
There was also some question/opinions about security and how to govern mashups. I think there is no need to be overly paranoid about this. However, service access control is still important in an enterprise world of mashups. Consider that today users are already doing mashups, whether you like it or not. They are doing this mostly in their heads or in a spreadsheet, mostly manually and mostly in a tedious and time consuming way, and in a not-easy-to-replicate-or-teach manner. If the users have access to the data, you don’t really have control on what they do with it in their heads or spreadsheets. So, why not make it easy for the users to do this, faster, more efficiently, collaboratively (so they can share their analysis/knowledge), and to do all this with the least amount of programming skills required (leaning towards no programming required). We are aiming for this at JackBe, as I am sure other vendors in the mashup space are.
Other notable topics of discussion at the Mashup Summit were:
- Microformats. My favorite topic. I think Microformats (www.microformats.org) are currently under rated/under used in an enterprise. Expect to see a lot of interesting things in this space in the coming months. We have a thing or two in the works about this as well at JackBe. (Also see Jeff Nolan’s comments on this topic.)
- Mashup of Data vs. Processes. It is easy to confuse the two and it’s important to remember they are not the same.
- Offline & Synchronization: Ross Mayfield of SocialText asked how important this feature is and whether there really are good use cases that demand this feature. He blogs on this (and related) topics quite a bit and has a interesting perspective.
Bottomline, we at JackBe believe that a mashup has to be user-focused and user driven (not IT driven). Let’s stop searching for the killer app. The killer app is here and it is the User. Let’s enable and empower them to do what they really can do with mashups.