During the MAX 2009 conference, Duane Nickull was interviewed by DZone on the subject of RIA architectures. As an architect, I appreciated Duane’s comments about the responsibility of architects, versus developers, where RIAs are concerned (e.g. focusing on and valuing interaction design and user experience, distilling key business requirements by working closely with those the RIA will serve, being mindful of the framing process-oriented context, etc.).

In particular, don’t frustrate users resonates with me (e.g. it’s a non-technical answer to “what is an RIA?” (or “reeyah”). Duane’s Revenue Canada example (or not “getting” this) is a good one. It just so happens that I encountered my own today…

First, two exhibits off Twitter:

• A – I try to help shine a light on a community voice
• B – (less than six hours later) subject vendor asks for community input

Don’t get me wrong, more vendors need to be reaching out and engaging with their communities in deep and meaningful ways. So, I’m not suggesting that creating a forum for community discussion is bad. However…

What if, instead, the forum was seeded (pre-launch) by a reasonable distillation of those who’ve already voiced their concerns, like Pie? The Web is there to be culled–”listened to” if you will–you just need to mine it.

As Duane and his co-authors talk about in Web 2.0 Architectures, more and more of us are living declaratively. Certainly this is true when it comes to providing candid feedback and standing behind things we believe in (e.g. vendors we want to succeed…and those we don’t).

So, an alternative forum post could have listed Pie’s identity management feedback alongside the feedback of others–fully annotated with community profiles, source links, etc. Of course, those supporting the forum could proactively reach out to folks like Pie to confirm that discovered feedback is appropriate for syndication and could enable contributors to easily follow the conversation moving forward–ideally in the medium of their choice (e.g. email, feed, etc.).

Update 12/29/2009: Of course, Pie, being the strong advocate he is, reached out again and updated the forum post himself.

If you’re at the Adobe MAX conference this week, then you already know: Adobe LiveCycle Enterprise Suite 2 has launched.

And judging by the attendance at the pre-conference session on LiveCycle, there’s significant interest in building user-centric applications in the enterprise–exactly what LiveCycle ES2 is designed to unleash!

As you will see on the main LiveCycle site, ES2 is all about:

• Delivering intuitive user experiences
• Streamlining business processes

• Effectively managing communications

Given my work on smart client architectures, I have to say that I’m particularly excited about the potential in LiveCycle Mosaic.

Mosaic provides a compelling framework that brings together aspects of business and collaboration to drive richer context pivoting. The task (goal) at hand is more richly represented at all times, allowing the user to pivot more effectively and efficiently and leading to better outcomes more rapidly.

Mosaic is not merely about aggregation as is portal technology. Rather, Mosaic is about intuitive, contextual composition that puts the focus back on the user’s task rather than all the supporting systems underneath. Users can access their mosaics either in their browser or on the desktop (via AIR support). Catalogs of mosaic application assets like tiles can be shared to encourage reuse and to simplify future composition.

LiveCycle Mosaic should be a boon for user-centric, content-enabled applications development.

By the way, if you’re not at MAX (like me), you can still participate online. For example, view the top three sessions from each day at MAX on demand. (Today’s keynote stream was five-by-five at my desk!) Check out the MAX Companion, too, while you’re there. October 11 will see all the MAX content posted online, too!

Cheers!

Update 11/25/2009: Please read the “what’s new in ES2” document. I also recommend that newly revised LiveCycle Developer Center (aka DevNet site).

Today Silverlight 3 officially launched; so, I decided to cut-over to v3 from v2. Here’s what worked for me:

1. Ensure that all browsers are closed before proceeding.
2. Open the Control Panel to begin uninstalling older Silverlight 2-related software.
3. Select “Microsoft Silverlight Tools for Visual Studio 2008 SP1 – ENU” and uninstall the older IDE integration.
4. Select “Microsoft Silverlight 2 SDK” and uninstall the older SDK, which, in my case, was installed via the IDE integration installer.
5. Select “Microsoft Silverlight” and uninstall the older runtime. In my case, I was running Silverlight 2 GDR 1 (2.0.40115.0). You can determine your current version, if need be, here.
6. Since I didn’t have an older version of Expression Blend on my machine, I could proceed. If you have an older version (release of beta), you should remove it, first, then continue on here. Before you install please know that once you do, your machine now becomes a Silverlight 3 development machine. You cannot down-level target to Silverlight 2.
7. Install Silverlight 3 RTW (3.0.40624.0).
8. Install Microsoft Silverlight 3 Tools for Visual Studio 2008 SP1, which includes the Silverlight 3 SDK.
9. Install the Release Candidate of Microsoft Expression Blend 3 + SketchFlow. (I was hoping for the final release of Blend today, but I’ll settle for this (60-day trial) RC (version 3.0.1921.0).)
10. Upgrade your Silverlight 2 solutions to Silverlight 3 solutions via the upgrade wizard. Etc. Etc.

Since the web is already flooded with plenty of additional commentary and sample applications, I’ll stop here and keep it practical for now. However, I’m excited about this new release and have some ideas to pursue in Silverlight 3. Stay tuned…

P.S. You may also want to download the HTML Help documentation for Silverlight 3.

I’ve been patiently waiting to blog about this for some time. Now that Microsoft has finally gone public with this during this week’s TechEd…

Acropolis‘s vision is to revolutionize the way that organizations define, develop, configure, deploy and manage client solutions. Clearly, this is a broad vision, and it will be realized over a set of releases. Having worked with David Hill, Microsoft’s architect on this endeavor, in the past on smart client architectures, I’m confident that this vision will realized.

Organization-wise with Microsoft, Acropolis fits into the same developer-oriented division as the Windows Forms, ASP.NET and Silverlight folks. For example WPF and “Cider” are sister teams to “Team Acropolis.” This is worth mentioning simply because it shows that Microsoft is organized to deliver on its unification vision where application development is concerned.

So, going to the next level of detail, what is Acropolis and why should you care is you’re building clients (applications) in the Microsoft development environment?

Here is what excites me about each piece to the puzzle shown above:

•  Component model layer: support for pluggable metadata consumers (strategies or providers) in a pattern-based (pipeline and dependency injection) foundation…sounds really cool! Seriously, the way metadata flows here is powerful. Stay tuned…
•  Component application layer: infrastructure support for data synchronization, concurrency, hosting, etc. allows me to focus on my specific problem domain rather than on plumbing
•  UI-specific application layer: the line-of-business focus is timely. More below…
•  Composite application framework: you already build business-oriented components; now you can drop them into a framework that simplifies their development and management, while improving their user experience (parts (basic unit of composability), forms (part-based, goal-oriented orchestration), services (pluggable providers and strategies) and connection points (patterns of part interaction))
•  Tools: visual composition and declaration empowers more imaginations to take what I offer in directions unimagined and productive, than just those belonging to coders

WPF is evolving to address line-of-business scenarios (i.e. 3.5 and beyond). Acropolis is being developed, among other things, to complement and fill-out the next release of WPF where LoB development productivity and usability is concerned. (That being said, Acropolis is not beholden to WPF, but it does target it primarily.)

From the above stack diagram, you can see that Acropolis (and certainly WPF, for that matter) are about more than just LoB apps. However, enterprise and “breadth” ISVs are the focus of version one.

So, what is the relationship between Acropolis and the just-released Smart Client Software Factory (SCSF) that is built upon Enterprise Library 3.1 (EntLib)? What Acropolis in relation to the Composite Application Block (CAB)?

SCSF/EntLib targets both .NET 2.0 and .NET 3.0. Currently Acropolis targets .NET 3.5. There is no relationship between Acropolis and the SCSF, apart from the fact that the patterns and practices team will be evolving their offerings over time to add value onto the current platform. With this release of CTP1, though, the Acropolis does expect the CAB to become deprecated and the SCSF to evolve, targeting Acropolis.

Visual Studio 2008 (“Orcas”) Beta 1 is required for its WPF designer (“Cider”) functionality. While Acropolis design-time experience requires .NET 3.5, it should be possible to target down-level runtime environments (e.g. .NET 3.0, Windows Forms, etc.). These details are still being worked out. As noted below in closing, feedback is actively being sought on Acropolis by its team; so, if specific down-level support is required, be sure to articulate it.

Here is what those in Microsoft’s DevDiv are saying about Acropolis:

• Brad Abrams
• David Hill
• Kathy Kam

Looks like it’s also worth paying attention to Don Burnett’s Acropolis commentary.

The Acropolis team is asking for your early, formative feedback, and I encourage you to provide it to them.

Update 6/7/2007: Glenn Block answers more common questions about Acropolis relative to p&p offerings.

It’s good to finally see “Smart Clients: Essential Requirements and Candidate Services” published on MSDN.

This article was something I helped craft when the Microsoft Architect Advisory Board was active. As the article states, MAAB has been retired and many, if not all, MAAB members are now Solution Architect MVPs.

One important name not listed in the acknowledgements is Jon Tobey. Jon took the collaborative work of a group of architects and clarified it into something more publicly consumable. As you might imagine, this was not an easy job–my colleague Greg Poirier characterized it as follows: “herding us like a bunch of alley cats to the finish line.” Thanks, Jon.

And thanks to all my fellow Smart Client Architecture Working Group teammates. Cheers!

Back in February of this year, I posted on the maturation of smart client development artifacts from Microsoft. What I called “SC-BAT” then is now the recently released Smart Client Software Factory on MSDN (i.e. SC-SF since 5/1 apparently). I look forward to upgrading my GAX and GAT to the Jun-06 CTP’s and upgrading to the new SC-SF, too–per getting started.

Aside: It’s interesting for me to observe the morphing of “Baseline Architecture Toolkit” into “Software Factory” given my time spent on Software Factories as a reviewer of the book by that name.

Update: I see that Sam Gentile thinks highly of the new release, too.

If you’ve been a reader of mine for awhile, you know that I’ve blogged about smart clients in the past (defined). It’s been great to see a fair bit of progress from Microsoft concerning more robust infrastructure for the development of smart client-oriented applications. Insufficient common plumbing was a central concern of the Smart Client Architectures working group within the, now all-but-closed, Microsoft Architect Advisory Board. Each one of us in the WG saw numerous one-off efforts that would benefit significantly from a more capable, robust platform from which to build smart clients (e.g. focus more on the business/problem domain instead). The following efforts are yielding these anticipated benefits: Enterprise Library, Composite Application Block, Guidance Automation Toolkit and Smart Client Baseline Architecture Toolkit.

In case you’re not familiar with these Microsoft (and community) provisions, I encourage you to visit the following links:

• Enterprise Library (EntLib) – January 2006 release for .NET 2.0: GDN home, contributor blogs, the goods (bits), overview and core architecture presentations, and (3) ARCast recordings. I also recommend this good explanation of the Object Builder framework within EntLib.
• Composite UI Application Block (CAB): home, community, the goods (C# bits), hands-on labs (C#), and introduction presentation. CAB specifies a composition-oriented architecture and framework for the presentation layer. It is a smart client framework for building composite enterprise applications (versus small standalone applications), enabling solution developers to focus on their business domain instead of application plumbing. CAB has a dependency on EntLib.
• Guidance Automation Toolkit (GAT): home, a short history, and an article. If you’re familiar with Enterprise Templates from VS.NET03, GAT is like ET on steriods. GAT requires Visual Studio 2005 (i.e. its DSL support). DSL is short for domain-specific language. DSLs are an essential part of the Software Factories paradigm.
• Smart Client Baseline Architecture Toolkit (SC-BAT): community, history, drop 3 details, and (2) ARCast recordings. The Smart Client Baseline Architecture Toolkit provides comprehensive, end-to-end guidance for building common Smart Clients. The reference implementation of SC-BAT has dependencies on CAB (which depends on EntLib) and on GAT.

First another annoyance and then a plug for a forthcoming development tool…

Another annoying observation…when I delete a folder (e.g. a sub-folder under my Exchange-based account Inbox folder) why does Outlook regularly prevent me from emptying my Deleted Items folder on the same mail account (i.e. complain that synchronization between Outlook and Exchange must first be completed before deletion can occurr)?

This is nonsense; I’m deleting a folder. I don’t care about Exchange’s problems–I don’t want to see this folder and its content any more (sync or no sync). The dialog Outlook 2003 is presenting to me smacks of poorly crafted UI such as “you asked to do Foo…are you sure that you really want to do that?”

It strikes me that a better user experiences would be to allow my delete/empty operation to complete successfully at least from a presentation viewpoint if not also from a mailbox perspective. If mailbox processing is still required, Outlook can privately tell Exchange of this requirement. Unless someone can explain to me why transaction support for this processing is required, Outlook doesn’t have to worry about Exchange failing to carry out mailbox cleanup in regards to this operation (i.e. no rollback semantics).

On the upside for Outlook developers, later this year Microsoft will ship Outlook support for the first time as part of Visual Studio Tools for Office 2005–what those of us in the early access program (and elsewhere) have come to call VSTO-O (e.g. Kjell-Sverre Jerijærvi). While it doesn’t cause any underlying change to the Outlook 2003 data model, it does reduce the amount of effort to produce compelling solutions that leverage Outlook 2003. If you’re looking for object model changes, you’ll have to wait for “Outlook 12.” Hopefully there will VSTO-O coverage (sessions + BOFs) in LA during PDC 2005.

One of my MAAB colleagues, David Hill, was heavily involved in the architecture of VSTO-O. I have a lot of respect for David’s insights into Smart Clients–we work together on the Smart Client Architectures working group within MAAB–hence, my involvement in and support of VSTO-O.

On a historical note, VSTO-O supersedes a previous effort called “Elixir” that was never released but used primarily as a vehicle for early architectural feedback.

[See also Architecture in a Priory.]

Over the past five months or so, the Smart Client Architectures working group of the Microsoft Architect Advisory Board (MAAB) (whew, that’s a mouth full) has met regularly to produce a reference model that is currently planned to outline a set of content that will map business needs to solutions enabled via smart client architecture. During this year’s (fourth) Microsoft Strategic Architect Forum (SAF) in Redmond, WA, we were able to present an overview of our progress to architects in attendance. It was encouraging to see the alignment between our work and our audience. A majority of interactive breakout sessions during SAF focused on concerns being addressed by our work (e.g. contextual integrity and transmission (aka references and relationships), simplified smart client production and potential expression as a software factory, etc.). Exciting times ahead!

It was nice to finally meet Rocky Lhotka in person. Shortly after our first working group meeting in May, Rocky joined the effort to advance smart client architectures. Anyway, we met up toward the tail end of the SAF–before the start of the MAAB meetings. One of the things I enjoy about blogging is the opportunity to gain another’s insight into a shared experience. For example, Rocky and I sat next to each other as Pat Helland presented another segment of his Metropolis series of talks (e.g. here is the presentation by Pat that shaped last year’s SAF). (Rocky blogs about it here.) What struck me at the time is the incredible value I receive when someone like Pat takes a hard look at history, correlates it to the present and suggest how to anticipate what’s to come–both opportunities and challenges. (It was also a treat to hear Pat’s lovely rendition of Mr. CIO Guy.)

Hopefully Microsoft will publish a MAAB page soon where you can gain addition insight into the makeup and focus of this group of architects. Ideally it will be a fairly collaborative area that allows you to interact with us (e.g. share your perspective; discover other feeds on architecture; etc.). Once Harry Pierson and team return from OOPSLA–wish I was there!!–look for these and other developments in the architecture space.

Earlier this week, I was in Redhill, UK at Nutfield Priory as a member of the Microsoft Architect Advisory Board (MAAB). Five focus groups spent two full, intense days at the Nutfield Priory collaborating on issues facing architects today. My group spent its time on Smart Client Architectures, and it was surprising how quickly the members of this group–the largest of the five–reached consensus on common vision of making it easier to build smart clients, which are admittedly hard to build today. Unlike some of the other areas focused upon by MAAB members this past week, which are five, even ten years out in their realization/implementation, smart clients are here today, driven by the growing demands of information workers, IT administrators, and software development staff.

David Hill joined our group as the representative architect from Microsoft. His suggestion that smart clients are rich clients done right resonated. Easily the most important feature to deliver in smart clients is deployment. Smart clients must be easy to deploy; if they aren’t the whole sell is a complete non-starter. However, deployment is not a one-time event; it’s a process. The process begins upon initial installation of software and continues throughout the life of the software and the role of the user.

Unfortunately there tends to be a default reaction when a new application is required in the enterprise–build a thin client (browser-based application). Not all user experiences are meant for the web, though. While this may seem like an obvious statement, it’s clear that today’s businesses knowingly and, worse yet, unknowingly sacrifice the productivity of their employees, partners and customers to the web. Do you need to work in your application offline? Without a network connection, the web disappears; so does the ability to do critical business that may be transacted on cached data. Is the responsiveness and richness of your application critical to decision making and support? What if the available network has only low bandwidth and/or high latency; will your users perceive your software, your service, your brand in the same manner as if there were high bandwidth and little latency? Will sales be lost, will collaboration cease, and will customer satisfaction drip away?

To be clear, our work has just begun on this important facet of architecture. The roadmap will be incremental. Be looking for material to be published by this focus group, and by the MAAB in general. Keep an eye, for example, on the following resources:

• Future issues of JOURNAL
• Future PAG materials (blocks and docs)
• Future Microsoft Architecture Center Update newsletter(s)

I’m looking forward to contributing to smart client architectures as a member of the Microsoft Architect Advisory Board (MAAB). We have a meeting coming up in early May where this will be a focus. Hopefully we’ll have some material to publish and present more publicly thereafter. In the meantime, there is value in the following resources on the subject of smart clients:

• David Hill weighed in on the definition and makeup of smart clients (also here on his blog). Chris Anderson summarizes David’s original post.
• On 2/27/2004, the Smart Client Offline Application Block was released (announcement). This block is well-accompanied by David Hill’s article. On 1/27/2004, David gave an overview webcast on this subject, too, with Brenton Webster; he also points out a case study on a medical diagnosis smart client application.
• Naveen Yajaman posted his slide deck and demo code from the Smart Client Offline Application Block webcast he presented on 3/25/2004.
• The deployment section on windowsforms.net lists several resources worth mentioning as follows: Updater Application Block (pull-based, BITS-by-default solution)(6/27/2003), .NET Application Updater Component (reference; WebDAV, not BITS, solution)(5/5/2003) and Write Auto-Updating Apps with .NET and the Background Intelligent Transfer Service (BITS) API (Feb-2003).
• David Hill answers some questions (e.g. smart clients and no-touch deployment (NTD)) he received as feedback to his post, above, here and here. Jon Udell warns not to give up on NTD, which is here today, by waiting for ClickOnce. If you’re already actively developing with the current Visual Studio 2005 Whidbey Technology Preview, though, the April 2004 issue of MSDN Magazine features an article on ClickOnce by Brian Noyes that was recently posted online.
• GDN features a workspace for releases related to smart clients here.

7/12/2004 update: more smart client resources:

• Smart client deep dive
• Making a rich client smart: going offline
• Smart client architecture and design guide

(Part 2, Part 1)

Concerning architecture, patterns, etc.:

• Pattern libraries exist from IBM, Sun and Microsoft–and it’s especially important to consider each viewpoint where interop and heterogeneity are concerns in your enterprise and/or target market.
• Folks weighed in on services vs. objects, SOA, services vs. components, etc. ([1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]).
• David Hill weighed in on the definition and makeup of smart clients (also here on his blog). I’m looking forward to contributing to this particular area of architecture as a member of the Microsoft Architect Advisory Board (MAAB).
• If you are an architect, are you more of a Purist or more of a Realist? Michael Earls compares these two architect behaviors. Chris Anderson puts himself more on the Realist side; he also offers a great quote: one of my favorite things about Microsoft is that you are never the smartest person in the room. Later, Michael Platt blogs about architectural types.

Concerning web services:

• Christian Weyer reminds us not to blindly trust public information–this time in the context of WS-I Basic Profile conformance in ASP.NET Web Services (ASMX). Matt Powell comments. (Mindreef SOAPscope 3.0 has received a fair bit of positive press in the web services ranks, in general, lately.) Digging still deeper, Christian uncovers an IIS behavioral issue at the root of his previous post, and offers a workaround. I’m looking forward to full BP conformance support in Whidbey of later this year.
• W3C updated its Web Services Architecture document.

Concerning applications, tools & samples:

• Project Niobe is a prototype SDK that abstracts the VBA, COM and MAPI layers into a single managed object model. From here, you can develop applications that integrate with Microsoft Outlook using existing technologies, but develop against an object model that’s more in-line with the design of the classes and namespaces in the .NET Framework. – Simon Guest, author (more from Microsoft-Watch)
• Christ Pratley, OneNote GPM at Microsoft, talks about the genesis of OneNote, using OneNote for blog prep, why OneNote doesn’t have feature X, and OneNote’s development process.

Concerning blogs, RSS, etc.:

• The January 2004 issue of the Architect Journal featured an in-depth article on dasBlog, a weblogging engine based on .NET.
• .Text author Scott Watermasysk details how .Text handles Trackbacks.
• Robert Scoble talks with the FrontPage team, who demo’d a RSS-reading component for SharePoint. Looks like the team is interested in RSS and weblog usage scenarios. Not surprisingly, elsewhere in Microsoft an effort is underway to leverage RSS from Windows SharePoint Services (WSS). SharePoint Syndication hits GDN soon thereafter.
• After Dave Winer visits Microsoft, Robert Scoble comments on why Sharepoint really isn’t used for weblogging.