After an unsolicited and random tour of Dublin by my cab driver (Brahm Stocker's house, school where the guy who wrote the screenplay for the Commitments taught, etc) I arrived at mashup camp held in the Guinness Storehouse, fresh from an early morning flight out of London. Amazingly enough, mashup camp is held in the very place they brew Guinness beer. This will be a very interesting few days.
The first day and half of camp is dedicated to what is called "Mashup University". These are presentations by various contributors to the mashup ecosystem. Today they included several presentations from IBM on QEDWiki and the Mashup Hub, which together make their Mashup Starter Kit, Microsoft on their Popfly site showcasing mashups with Silverlight, John Musser from Programmable Web, and presentations from Kapow, AOL and others. (sorry if I missed anybody, I was a little late).
QEDWiki from IBM
I have had some experience with QEDWiki already. I saw a demo of it at AjaxWorld in SantaClara last year and later had the opportunity to work with Dan Gisolfi, who presented today, with doing a proof of concept with it around a project at Reuters. QED has come a long way in the past year, and I think that packaging it with MashupHub was a good idea. MashupHub is feed aggregator much along the lines of Yahoo Pipes (even with a similar slick GUI for linking feeds together). One major difference that is beneficial to potential enterprise users, is that the MashupHub can run on a private server, unlike Pipes which is all in the public domain. The combined kit from IBM attempts to enable the assembly of Mashups by non-technical domain experts who would work with widgets and feeds already built by the mashup enabler. This is an ambitious and important goal, since so many situational applications are never built for the simple reason that the domain expert who conceptualized the app is not able to create it directly. And engaging IT to build something from scratch is too expensive and/or too slow. QED covers a lot of ground, but we are not out of the woods yet. While it is true that you can build an application in QED without doing any coding, the assembling itself requires quite a bit of expertise of a technical sort, especially when it comes to resolving aggregation and filtering of disparate data feeds. These are problems that no one has the solution for yet, but IBM has done a lot of heavy lifting with QED and MashupHub.
Dan touched on many of the partnerships IBM is forming around the project: wiith Dapper, StrikeIron, SMILE and OpenSearch. He also summarized the Enterprise Mashup Challenge that IBM is hosting here for the best Enterprise Mashup built on QED.
One of the big problems with QED is aesthetics. Its clunky look detracts from its real utility, and risks undermining the confidence non-technical people will have in using it to build mashup applications. It’s a shame because if it just got polished, it would really shine. Ironically, the mashup hub looks far more finished, and this is the backend tool that feeds QED.
Hi Ho Silverlight and away
An interesting contrast to IBM's demo was the demo by Microsoft of the the Popfly community for Silverlight mashups. Here it was all style with a shortage of substance. The widgets in Silverlight looked awesome, zooming around the page in 3D and Martha (Rotter?), who was doing the demo was able to easily link the widgets to feeds, visualizing data in different formats effortlessly (barring network problems). However, most of the lush graphics have no meaning for me and in fact are a negative. What use could I have for the whack-a-mole visualization of data when I am trying to solve the workflow problems of high volume financial traders? Microsoft seems to be preoccupied with out-flashing flash here, in a bid which I don't think plays to Microsoft's strengths.
social networking
John Musser delivered what was an excellent summary of the evolution of web mashups he has been tracking with Programmable Web in the past 2 1/2 years. He focused largely (for good reason) on the rise of Social Networking platforms and discussed Google's OpenSocial project. John also gave some entertaining examples of mashups. My favorite was a mashup that actually used the Amazon.com wish list service to identify political subversives and locate them on a Google map. Oddly enough, they are mostly on the east and west costs. Check it out here.
Gregory Cypes' summary of AOL's AIM API project tied into John's presentation very well. Gregory also talked about AOL's use of the OpenID standard.
The day was wrapped up perfectly with the freshest pint of Guinness possible from the bar at the top of the storehouse. Tomorrow should be even more exciting as the gloves come off and the geekiness is unleashed.
No comments:
Post a Comment