Your input is wanted: ‘What is the future of enterprise mashups?’

Hi Chris,

I'll happily start the ball rolling if you like. I suppose I'd have to preface my comments with the rider that my total experience of mashups is around using emml and Presto, so bear in mind that this may skew some of my opinions, both good and bad.

I believe "Enterprise Mashups" as a business tool have a huge potential to add benefit to business productivity over the next few years. Just based on what I've been able to do with Presto in our own organisation speaks volumes about how you can use the technology to add value to any business. Being able to pluck data from diverse sources, format and re-purpose it, and then deliver new "information" to users, is a very powerful proposition.

However, as I've worked with Presto I've come to realise that there are some impediments to wider adoption of mashups within the enterprise;

• Much of what we initially see in mashups externally (and some may wish to disagree) seems somewhat less business focused than I think is needed to encourage wider adoption of mashups in the enterprise. To get the attention of business users we need to grab their attention in a way that they go "Wow, I want that working for me too". How we achieve that is the question. I'm personally taking the approach of just creating business mashups, putting them in front of the users, and then hoping at some point, they'll realise just how powerful the technology can be. I see it as a "Quiet Revolution".
•  
• My other concern is the dependence on XPath and XQuery. I came in to writing mashups using emml with virtually no knowledge of XPath and XQuery but with a lot of effort and perserverance and many long hours wrestling to get my XPath expressions right I've manage to produce some pretty cool code. However, I have this nagging feeling that the "art of XPath" could be an inhibitor for many non-technical users who may well give up in frustration long before they reap the benefits of mashups.
•  
• I've spent a lot of time hand coding my mashups with emml because some have turned in to quite complex exercises, but my early attempts at using a GUI to build mashups easily outstripped the capabilities of the tool. I think the mashup vendor that builds a comprehensive GUI front-end that makes even moderately complex mashups possible will attract more attention. I love the "drag and drop" style of Wires but the lack of a comprehensive set of actions limits what is achievable, and that I think this makes mashups less accessible for a broader range of business users.
•  
• As I commented in a recent blog, one of the problems I face now is accessibility to raw data. I can solve the problem using mashups and datasources to expose corporate data to users who can then utilise it in their mashups. I think a key to greater use of mashups in the enterprise will be a comprehensive storehouse of pre-exposed data feeds. That's one of the great things about emml, it doesn't take long to make "locked up" data easier to access, but the more pre-exposed data available the better.
•  

I'm personally optimistic about enterprise mashups and what they can do for any business, big or small, but there's more that can be done to promote and expand their use as a truly indispensible business tool.

I hope these comments help in some way. I love what Jackbe has done with emml and mashup technology and I'll continue to be a strong advocate for using it in our organisation and talking about it with others in the industry.

Cheers, Innes

Here's my take on the question:  http://jitlife.blogspot.com/2010/03/what-should-be-next-for-enterprise.html

I hope you all enjoy!

Mike