Application Wikis: Powering the 'Big Red Button'?

Bob Glushko, director of UC Berkeley's Center for Document Engineering, has a saying that goes something like, "Just take your application's XML schema, give it to the Platform, and press the big red button. Done, application created." That was the vision behind the Center's "XML Application Platform" project I led last year. It's model-based application nirvana, where the data model is explicit in the Schema, and through annotations and maybe some additional metadata, an entire web-based application is created automatically. And it's hard to do. We built a prototype on Orbeon's Presentation Server, which can be thought of as a next-generation Cocoon, but alas a 2 semester research project can only get so far...

I've finally starting digging deeper into JotSpot, the new-wiki-kid-on-the-block. They're positioning as an "application wiki", and I'm now seeing what that actually means. From what I see so far they've done a nice job of making it very easy to have 2-way integration with enterprise data and applications, whether through web services or RSS (e.g., see Jon Udell demo that includes integrating with SalesForce.com). David Mattison writes:

The problem, though, is that the wiki, by itself, has limited programmability. You can enter text and make links, but that’s not quite the same as building an actual application. While both SocialText and JotSpot appear to offer additional components that they’ve built, the next stage may be for them to start promoting easy integration with additional outside apps and components developed by others. If you start thinking of the wiki as less an open whiteboard for text, and more an open workbench for integrating web services-based applications however you’d like, things start to get much more interesting.

So if JotSpot and other "application wikis" like SocialTextand XWikican come up with rich enough scripting languages, the wiki becomes the natural platform to power the Big Red Button Platform. The real work comes in creating the XML models that capture the semantics of access control, workflow, and some of the other aspects that real apps require. But much of that stuff is presumably already in the Wiki! So with some thinking and work, I can picture these aspects as little JotSpot applications/services that get stitched together with some XML configuration and viola, we've got XML schema-driven applications -- and a Big Red Button!

Middlespace: Context for Discussing Collaborative Software

Ross Mayfield presents a few case-studies to help explain what he calls Middlespace

Bottom-up phenomena has accelerated in recent years because of social software. A relatively simple decentralized pattern of enabling more connections and groups to form has complex results. These results (for example: open source, the long tail, heterarchical organization, emergent democracy, wikipedia and participatory media) hold great promise. Bottom-up production is driven by social incentives, comes at a lower cost, realizes economies of speed and enhances quality through diverse and greater participation. Despite these benefits, Bottom-up phenomena is perceived as a significant risk because the dynamic of control is uncertain. But every risk has its rewards and can be managed if known.

Where the bottom-up and top-down meet -- middlespace -- is the realm of policy, metrics, incentives, cooperation and sharing control. The practice and politics of this realm are best explored through new case studies. [full post]

Acknowledging and defining this 'middlespace' provides context and frames the discussions that must take place when collaborative software spreads beyond early adopters into an organization's mainstream.

Build vs. use Open Source? Koders.com can Help.

Google for open source projects has appeared: Koders.com.  Looks like they scoure open source repositories and let you search for code by keyword, language, or license. 

But even more interesting is their "Build vs. Integrate Open Source" tool, which allows you to ballpark the cost to develop a given component yourself using some of your own cost metrics.  For example, here's their assessment of the cost to duplicate the popular Hibernate project's functionality. Good stuff.

ehick Migrating to TypePad.com

I've decided to migrate my blog to TypePad.com. From now on, I'll be hanging out at http://scottmcmullan.typepad.com/blog

For syndication purposes, my Feedburner RSS/Atom XML feed is still the same: http://feeds.feedburner.com/scottmcmullan.

Open Source Inside?

I had the pleasure of meeting Palamida founder Ray Waldin recently. Palamida is addressing a growing problem for ISV's: "what sort of open source might we be using, and are we adhering to the licenses?"

With a smorgasbord of ready-to-integrate code and a the proliferation of licensing terms, it's getting harder to ensure a complex, long-lived software product doesn't mistakenly run afoul of any licenses. Both Palamida and their competitor Black Duck Software are bringing a welcome service to market -- best of luck Ray!

The Dells of Open Source, Part 2

A little over a week ago I blogged SourceLabs, a newly minted Open Source aggregator aiming to sell support and services around productized open source components that have affinity. I expected to see a flood of these (it's a good idea after all), and last week we were indeed treated to another: SpikeSource. Turns out Kim Polese is CEO and the company was incubated and funded Ray Lane and Kleiner Perkins. Nice start!

Digging deeper, an interview with founder Murugan Pal (former Asera CTO) at TheServerSide quotes him as saying,

Almost all the other companies in this space talk about the same thing--but what are the objectives metrics that they are measuring? Is it integration? Standards compliance? Regression? Negative tests, boundary tests. How can you integrate these tests in multiple dimension? How about IP rights classification? Security threat assessment?

But thinking about this for a few seconds longer, I see some cracks in this foundation. For starters protecting the investment Murugan mentions above. Once SpikeSource's LAMPJ stack becomes known as versions a, b, c, ... j of open source components 1, 2, 3, ... 10, and this info is widely disseminated, why pay the subscription? Certainly some will, but how many, and how much can you charge? And then there's the low or no cost competitors, who "knock off" your package with their own free distribution the day after each new release. I can picture tiny me-too companies shadowing SpikeSource's "products" and getting a free ride off their rigorous processes.

I'm obviously glossing over a whole host other value-add they have in mind. However in general, the basic value proposition of selling access and support to a, "vetted and test set of open source software as a product" seems better suited to software higher either higher up the stack, like Gluecode's portal server, or to a "sell the services" model, like JBoss or Orbeon.

Excite co-founders launch JotSpot -- getting INTERESTING!

I've been reading Joe Kraus' blog Bnoopy on entrepreneurship since he started it last month. His posts recounting lessons learned as Excite's co-founder have been entertaining gems for entrepreneurs. Inovie Software, the San Diego-based business collaboration company I founded in 1997 and later sold (now at UGS), was a minnow compared to Excite. Even so, we faced most of the challenges Joe describes.

So given my background and interest in business collaboration and entrepreneurship, I was very excited to read that Joe's new company, JotSpot, is joining SocialText to bring the power of wikis to a broader audience. Nice!! Ever since I first used a wiki I've been hooked. The first thing I do now when getting involved in a new project is ensure a wiki is in place -- I don't want to work without one! A Wiki provides a type of ad-hoc, free-form collaboration we never achieved with TeamCenter. If I knew then what I know now, I... Nevermind! ;)

Dan Farber writes:

JotSpot is based on wiki technology, which lowers the barriers to creating Web-based collaborative applications. JotSpot extends the concept. "Wikis run out of steam in that they don't allow you to add structure or build apps," Kraus said. "On the surface, JotSpot looks like Wiki and you can use as a wiki, but it allows you to start with unstructured data and add structure incrementally."

That's the stuff right there. I'm really looking forward to seeing where SocialText, JotSpot, and some in open source (e.g., XWiki) take this "Application Wiki" concept.

ps. Jon Udell has blogged a Flash Demo for anyone interested in seeing JotSpot in action.

Web 2.0

I'm enjoying reading some of the coverage of the Web 2.0 conference going on this week. Coverage, pontifications, thought provoking soundbites, and new company and product launches are flying out fast and furious! Many ideas great to silly for sure -- I should have gone :(

One particular new company/product announcement (blogged by Jeremy Zawodny) that caught my eye was Rojo, a new web-based "feed" aggregator. I'm overwhelmed by my current 82.6 (and growing daily) sources of news, blogs, commentary, etc., and Rojo sounds like a promising solution. Not exactly sure yet how it compares to other similar products like Bloglines, but i'm currently evaluating both and will post some comments soon.