jump to navigation

Stuff I’m Working On, Mash-UP SOA for the Intention Economy March 17, 2006

Posted by poseidongroove in Apple, Chaos, Collaboration, e-WorkPlace, Edge Thinking, ESB, eWorkPlace, Graveyard Portal, iTunes, Mashup, Mashup SOA, Microsoft, SOA, Social Software, Web Services, Weblogs.
1 comment so far

I feel vindicated that Ray Ozzie was on the Gillmor Gang articulating some of what I’ve been saying again, on the Gillmor Gang (WIKICalc Gang) that there is value in Web 2.0 beyond media content, Podcasts and shared calendars. A full expose on this will come later, been reading Doc Searls article here on Intention Economy Great read. It’s pretty much describing some of the ideas I’ve been working on and articulating, it’s called Collaborative Services Description Language – I’ll describe it as an extension to Web Services Description Language and BPEL. However, CSDL goes beyond mere description of network services and simply wiring plain old business processes in a sequence for execution. If you’re wondering why not just use Business Process Execution Language, (BPEL) It’s because we need to get away from the straight-jacket of Business Processes described only in the context of IT Users alone and UDDI directories and allow for web mash-ups by real users to happily co-exist with business applications that have a social context of use. So maybe we extend BPEL !!! As long as your grandma can interact with it, that’s okay!

Think of a software agent that’s using underlying messaging workflows and rules engine on an ESB to connect multiple websites together to fulfil a composite transaction. So a Super Paypal might be the one that mediates the payment between intermediate sites that are part of this multi-step transaction.

So CSDL from an architecture perspective describes discrete events and services that any end user can easily orchestrate from a SOA or Mash-Up perspective, the important thing here is you allow chaotic services to happily co-exist with Business Processes. Loosen up that nasty straight jacket in Enterprise Applications which is really engineering user disconnect from enterprise applications. The most important thing is that some recognisable icon needs to exist to allow users interact with services in a fool proof manner from their browser. This icon, like the RSS feed icon that you, the end user can configure a personal agent. Okay, we’ve got enough acronyms to worry about, the important things is let’s provide a standard way of doing it.

For example, find me the cheapest multi stop flight to Hawaii from Paris, Book a Sea View room at my stop over hotel in Florida, Create a Watch List Agent that notifies me of any location on my specified route where I can watch Bon Jovi live or similar bands from liveplasma This mash-up spans a typical user browse session. In fact, it could be saved as a “Personal Holiday Agent” that other’s can clone or consume for a Service Bookmark Site!!! Imagine, this can complement the Lonely Planet Guides; you can just clone the experience. This is possible using Enterprise Service Buses with their messaging capability and routing capabilities but we need to build interaction services on top of them to make them useable. So where the money exists on the Web is not Attention Economy, it’s the Intention Economy, we’ll happily part with our hard earned cash if anyone can give you an experience or service that can take the burden off you !

  • The browser UI Interaction can be completely written in
    AJAX whatever given the interactive nature of these services,
  • Provide for a standard interface that describes these first class objects and their properties Forums, WIKIs, Blogs, Documents, Media Content, a Tag-able Content from an abstract perspective, therefore, If company a ships over the wire a WIKI objects and all it’s properties from IBM to MSOFT, Oracle or BEA’s collaboration platform, the stuff is created without any problem,
  • Give semantic meaning to user interaction/transactions in a social network context, as well as making it meaningful to computers and competing social networking or collaborative platforms,
  • Provide a unified social networking services or so called mash-up SOA,
  • Allow different platforms to exchange collaborative data/services in a structured manner,
  • Combine SOA Services and Social Network Services(Web 2.0) transparently from an end user perspective,
  • The primary aim of CSDL is to provide a means to unify the semantic web and provide a Create, Read, Update and Delete invocation mechanism in XML
  • Enable a standard metaphor to organise bazaar applications,

So whilst everyone is focussing on Apple and iTunes, calendars, Digg, Newsvine, and Flikr. They’re really missing the whole point. Web 2.0 is the first time we really made an effort to empower end users, yet everyone is wetting themselves on shared calendars. Its companies and services that provide a compelling service and infrastructure around helping us spend our money in a way we feel good those are the real winners in this new world.

I thought, I’d have time to walk through the XML Schema tonight, However, I’m struggling to stay awake now. Will post later. Goodnight.

Advertisements

Top-Down vs. Bottom-Up Approaches To SOA. No! There is a Third Way. February 24, 2006

Posted by poseidongroove in Bazaar, Chaos, Collaboration, e-WorkPlace, Edge Thinking, ESB, eWorkPlace, Graveyard Portal, Mashup, Mashup SOA, SOA, Social Software, Web Services.
7 comments

I was passed a link to a Gartner Report report (Sorry you need to have an account to view the report) recently by my CEO, Dan French. The Gartner article claims there is a process war being waged by a “Top-Down Brigade” who are process centric, that see SOA as a framework that underpins the notion of representing business processes in composite applications, that access functionality decomposed into a set of services. The other group, “Bottom-Up” camp are essentially the integration brigade. They see SOA as an improved way for creating interfaces between application. I think I’ll describe this category as adaptive vendors or companies that started off flogging EDI, EAI, Web Services technologies. You might also include Application Server Vendors trying to move up the food chain with so called Enterprise Service Buses.

The article also goes on to compare the two camps as the battle between the purist versus the pragmatist, the value generator versus the cost saver and the Business Evangelist vs IT implementer.

Even though the two camps have good points they still don’t get what SOA really means. Rearranging the Powerpoint Diagram is not enough. I have to take anything Gartner says with a pinch of salt. I don’t respect any company or bunch of analyst for that matter that constantly rig their reports based on which vendor pays the most tip. If they don’t rig or be easily swayed by particular vendors, how the hell does a bunch of unrelated APIs by any vendor suddenly become Enterprise Service Bus or Enterprise Content Management Hence, I call it the “Magic Quadrant Syndrome” I’m glad I’m not the only one that is suspicious of Gartner, James Governor has also bemoaned the problem caused by groups like Gartner and their “hommie” relationship with vendors. If you want any evidence of rigging, checkout this post and the threaded comments from a few years ago. Funny enough, I think the most lucrative job in IT might be Vendor Qualification and RFP/RFI Consulting. I think I’ve decided to go into this line of work!

The challenge that is facing us today is missing in this whole arguement. Where is the third way? I’ll repeat this again. Until SOA grafts on collaborative social networking services it’s going to stay in IT. Hard core processes locked into SAP, Siebel, PeopleSoft are meaningless. The word process drives me nuts. Even when they’re broken we persist with them. No wonder creativity is dying in organisations today.

SOA needs to bring that mash up revolution outside the firewall inside the enterprise and into our homes !!!. The Process Brigade need a social context. e.g. make mash ups become services in SOA, The fact that an activity is undefineable doesn’t mean it can’t be called a process. If you’re looking for good examples, you only need to look at Government Department IT Projects. Total waste of our taxes. There is no real process in government departments. All they really manage is exceptions. How can an exception become a process. You’re more likely to need to converse with a human being somehow to resolve most exceptions. The Bottom Up Brigade need to recognise that Integration for it’s sake is useless unless the integration brigade delivers the services into pervasive end user technologies forget it. A possible reference platform for this mash up is already out there checkout ServiceMix Talk about being Open Source and built from ground up to be Service Oriented. A word of caution, just like App Servers were plumbing so are platforms like ServiceMix. However, the potential exists to build something pervasive on such an open platform. My ideal reference services architecture looks like this. I think this is a starting point.

nextgen-eWorkPlace.GIF

Next Generation Services Reference Architecture

Here’s a checklist of the Third Way:-

  • Focus on Context of Pervasive Use not UDDI, RSS didn’t become useful overnight. It wasn’t until the word didn’t matter that it had a context of use.
  • End User Presentation Services will drive adoption of SOA technologies. How? By infiltrating everything they do at play, at school and at work.
  • Manage Chaos. Social Software and Mash-up needs horizontal and transparent means of control especially behind the corporate firewall. Remember, there can be Order in Chaos. SOA can be just that. If you focus on what people can’t do you’ve already lost the battle.
  • Build Trust not suspicion, Security Groups in Enterprise IT are out of control. They’re stifling innovation.
  • The word SOA Governance in SOA is being abused. So is the word compliance!
  • Aggregation Services are not as useful as they seem. If they were useful enterprise portals will not be Graveyard Portals today behind your company firewall.
  • Focus on multiple channels of interaction, e.g. email, fax, sms, etc can be service oriented as well. If I can orchestrate in Blackberry or iTunes you’re on to a winner.
  • Mash Up SOA needs an appropriate orchestration language similar to RSS, Something users can click on, browsers can orchestrate, what ever, I call it Collaborative Services Descriptive Language. (CSDL) a possible combination might be a more useful version of BPEL, I’m not comfortable with all the implementation examples I’ve seen so far because it’s not expressed in a typical grandma Josie or Uncle Harry’s context of use.
  • Uncle Harry, for example would like to create a multi step mash up service that monitors the cheapest and best locations based on specified criteria for his next cruise holiday. Think about car hire, guide, villa, golf, flights etc. These types of service need to be available for end users. Not some lonely sandal brigade UDDI Flunkey kicking tyres at an Order Process Orchestration!!!.
  • If you can’t affect the price of fish forget it. Metaphorically speaking.
  • I passionately believe in the vision above hence, I’m diligently going to pursue this vision with or without any help until I see it come to pass.
  • We need a Mash Up SOA Icon like the Universal RSS Icon, The RSS icon now easily recognisable, SOA needs to become like that!!!

Epistemic Arrogance – Opportunities and Threats in Software Development February 10, 2006

Posted by poseidongroove in Absolutism, Chaos, Collaboration, e-WorkPlace, Edge Thinking, ESB, eWorkPlace, Graveyard Portal, Magic Quadrant Syndrome, Mashup, Mashup SOA, Social Software, Web Services.
add a comment

It’s not often you listen to a podcast two or more times except, it’s a music mix. I found myself replaying an excellent podcast by Nassim Nicholas Taleb titled What Do We Know – The Scandal of Prediction from the IT Conversations Feed. I listened to the podcast several times today as it brings to the fore something I often remind myself. We’re prone to making assumptions and are more conservative than we’re prepared to admit in the software business.

I like to ask myself frequently what do I not know? What I don’t know now might be relevant in affecting a significant change in my design pattern or an accepted architectural orthodoxy. Therefore, I cannot make a prediction that what I know now is the only and best answer for the problem. I guess, that’s why I’m not hung up on any specific software platform. It’s also meant, I’ve worked on several platforms in my career. We can’t afford to be arrogant about what we think we know.

Once, we’ve come up with a design pattern or architecture, we think job done and there’s not much else to add. The edge thinkers are more or less excluded once the mainstream conventional thinkers have had their say. It’s this predictability that makes the Magic Quadrant by Gartner and such likes a bore. Once Fortune 500 companies have looked at it and decided the supplier based on analyst predictions end of story, innovation isn’t coming from anywhere else.

The freshest thing about the social networking phenomenon is the ease with which ideas can be shared, learned and debated. I’ve never known a time when application development was such an innovative activity. As an Architect that specialises in Collaborative Technologies and Web Applications Development it’s fascinating to see that finally, we can eliminate the “limits of scale” of organisational intelligence as described by Tom Gruber the inventor of Intraspect now owned by Vignette.

It’s with this in mind that the edge part of my brain didn’t like Web 2.0 and SOA as mutually exclusive technologies or what ever you wish to call them. The collaborative part of my brain and the Enterprise Architecture part tell me that SOA is meaningless unless it has a social context and is a means to this social end. That end is as a consumable service beyond UDDI Directory to end user at home and at work in anything from cars to Media Players.

Social Networks are viral. SOA can become viral and useful in an organisational and Web 2.0 context. I’ve been working on an Idea I can only describe here as the Collaborative Services Description Language (CSDL). I’ll develop this in another post. There is a fantastic opportunity for IT to finally become a “necessary good”.

SOA must become pervasive. Our conventional thinking right now means we’re building Enterprise Service Buses expecting UDDI to be relevant beyond the IT department. I think not. Right now, were building or thinking more of hard core business processes. Very little attention is paid to exceptions to the process. These classes of activities that fall through the design crack are actually the ones we need to be conscious of the most. They are the ones that end users will most benefit from. It’s also where the social context exists. They are very coarse grained e.g. Dispute Resolution Team Room between suppliers, an Ideas Generation, Document Approval Process, Brand Management, Drugs Discovery, etc.

These class of processes are not best served by hard edged SAP, eCRM type applications such as salesforce.com, it’s only possible to create such applications using a class of platforms that have agile written all over them.

Would it not be nice if we had a common language that allows Business Processes and Collaborative Work Processes to be orchestrable and re-useable regardless of CRM, ERP Platforms. This means, a create Team Room or Document Approval Request can be propagated from SAP to the e-WorkPlace application bundling the meta data associated with the request, an event message to the SIP compliant platform, a bunch of documents and video files associated with for example, an Invoice in Dispute into an RSS Feed. SOA patterns sometimes describes it as Workflow Automation. I think this is a cop out. Oh It’s too amophorous lets leave it out !!

Folks lets stop the bigoted arrogance of our conventional thinking, we got work to do.