Expert Texture Home Contact me About Subscribe Digipede Connect on LinkedIn rwandering on Twitter rwandering on FriendFeed

rwandering.net

The blogged wandering of Robert W. Anderson

No contract-first in Workflow Services

I’ve been looking a bit at .NET 4 Workflow Services as a part of AppFabric.  I’m surprised to find no way to build these services from existing message contracts.  I can understand imposing limitations on doing so – even that there might be no reasonable tool support – but this is a real stumbling block.

Certainly contract-first is not the only way to build services, but it is if you are implementing a published standard.  Maybe there is some way to tinker around with the output of the tools sets to make a Workflow Service compatible with some existing WSDL, but would it be worth it?

One answer might be to put a pure WCF facade over the Workflow Services.  Kind of a headache, but maybe workable.

This makes me think that .NET 4 Workflow Services are really targeted to be internal to the enterprise or at least where contract definition is flexible and controlled by one entity.

This follows a standard Microsoft pattern: help the enterprise dev in V1 and then expand from there.  This strategy makes sense, I just want it all in VS 2010, not in V.next.

Tags: , , , ,

More on the death of WinFx (Part 3)

Andrew Hilton recently commented on WinFx is Dead Part 2.  I am promoting it to a post so I can better comment on the comments.

Andrew says:

I’m of the opinion that it is not yet the official windows API because it simply wasn’t ready, and using .NET 2 with WinForms would have been too restrictive and no doubt would not have integrated with the subsystems of Vista (such as graphics). The .NET framework 3.0 was not released until almost the final release of Vista. The risk of developing the windows front end under the managed API and still ship on time would have been too great.

I think you are correct regarding the .NET 3.0 components of WinFx: there was an early plan that Vista and Office would take advantage of these technologies, but of course, everything was late.   Using .NET the .NET 2 WinForms really didn’t make sense either.

From what I’ve heard early versions of Vista and Office 2007 did use the managed API (or at least an early version of WPF). My feeling is that the dev teams simply couldn’t cope with a large framework like that shifting beneath their feet causing all manner of chaos. Hence the move by both Vista and Office 2007 back to COM, which also gave leeway to the WinFX teams to change stuff radically if they needed to. My assumption is that given the large technical leap of WPF, WCF etc there would have been many back steps as they were trying to go forward.

Yeah, but I’ll bet the delays of .NET 2 had a bigger impact on the likelihood of managed APIs in Vista than, let’s say, WPF or WCF.

I’ll be very suprised if the consumer version of Vista does not include some managed code WPF apps. All the indications are that WPF will do very well (for example see the recent video of the WPF version of Yahoo messenger), . . .

Yes, I think this is all true, but hold on.

My post wasn’t about .NET 3.0 (or what they were calling WinFx), but of the original idea for WinFx: a full (or fairly full) Windows API.

WPF and WCF have little to do with that, certainly WF has nothing to do with it, and Info Cards, not really either.  Of course, this last one is a part of Vista though as kind of an add on. 

All of this stuff is very cool, but WinFx was supposed to push managed APIs deeper down into the OS, not to be extra layers on top of it.  The internalization of these APIs were to supply a more robust and secure OS. 

And it is that WinFx that appears to be dead.

Tags: , , , , , , , ,

A Vista plea to Microsoft

So, Vista is delayed. One must assume that Microsoft has made the best choice with the information at its disposal — certainly better info than the second-guessers who have pounced on Robert Scoble.

I am left with a couple of questions, though:

  1. What does this mean about Vista Server? That was already delayed beyond Vista Client. Can we hope for that in 2007? 2008?
  2. What about WinFx? I can understand why Workflow Foundation may need to track the Office 2007 release, but do Communications Foundation (WCF) and Presentation Foundation (WPF) have to wait for Vista? Do these parts really need 6+ months more work? And what about Atlas?
  3. And, of course, what about IE7? Is IE7 really that far from ready?

On my 2nd and 3rd questions, it is hard to imagine that Microsoft will release final parts early — it would take some of the wind out of the Vista sails — but as a software developer, I hope they do.

It is very difficult for ISVs to incorporate these new technologies into products while the dates keep slipping.
My point here isn’t to jump on the dogpile, but to make a plea to Microsoft:

Please, while retaining focus on quality, release WinFx and IE7 as soon as possible. Please do not wait for Vista’s release to make these components ready in final form.

Can anyone from Microsoft comment?

Tags: , , , , ,

VSLive! David Chappell

Not the sonic-software Dave Chappell, but the one of Chappell & Associates on Building Modern Software: Services, Workflow, Integration.

Exuberant speaker. Ex-funk pianist.

Walked through service-orientation as the current default enterprise architecture; and how the new Microsoft technologies underly this architecture: WCF, WF, and BizTalk.

This was a good overview of these issues (but, I wish I could get to the Digipede VPN from here . . . ).

Tags: , , , ,

More on SSDL

In response to the renewed interest in SSDL prompted by the IEEE article, Jim Webber recently posted about how SSDL compares to the existing W3C/OASIS stack. I blogged about the article in my previous post, I like SSDL.

The post is worth checking out if you are interested in SSDL; however, I’m disappointed by the last comment:

Of course the practicalities of the situation with huge existing investments in the W3C/OASIS stack means that SSDL will most likely remain nothing more than a toolkit for academics and geeks.

Pragmatic and perhaps true, but I hope that doesn’t turn out to be the case.

I’m still hoping to see this in a future version of WCF — I’m counting on Savas to make this happen ;)

Tags: , , ,

Ode to cool names

After my last post on some of the cool Microsoft technologies at PDC05, I started thinking about acronyms and product names, and product positioning. Specifically, I started to miss some of the old Microsoft code names.

Indigo is now WCF.
Avalon is now WPF.
Longhorn is now Vista.

It feels so generic, so corporate, and so stodgy. While never having talked to anyone in Microsoft branding / naming, it seems like this is intentional. Microsoft is making a terrific push into the breadth of the enterprise. No longer is Office the solution to every problem and C++/VB the only tool. Microsoft is serious about serving this market beyond the OS/Office and, I think, the products names reflect this.

Of course, Microsoft doesn’t have an easy job on their hands – I don’t mean to imply that they do.

But, I miss the old names. At the very least, I wish Longhorn had kept its name. Or had taken the name of Indigo. Now that is a cool name.

Tags: , , , ,

Tech @ PDC05

Microsoft is presenting on and / or announcing a lot of new technologies at PDC05 that are interesting for us:

Windows Communications Framework (was Indigo) The more integrated the communications services are with Windows and the more transports-independent they are, the better. I look forward to giving our customers more options on how our components communicate (regarding protocols, transport, and security) to best fit their specific operating environment. Of course, a lot has already been written on WCF. I look forward to hearing more about future directions.

Windows Workflow Framework and Windows Workflow Services (was Windows Orchestration Engine): while we plan the enhancements in the workflow capabilities of the Digipede Network, integrating with existing systems is always a requirement. A no-brainer for us was/is to build a BizTalk adapter. WWS looks interesting (as it looks to replace the existing BizTalk Orchestration services) – it will allow us to integrate with BizTalk 2006 as well as Office 12. This means that Microsoft has just reduced the number of integration points we need while increasing the flexibility of the entire solution. Very cool.

Windows Server Compute Cluster Edition (CCE) This product is going to make a big splash in the growing 64-bit clustering market. The work that that team is doing to improve the platform for distributed computing is all good. As Dan said in a recent post, this is a critical move. I look forward to Kyril’s talk and to catching up with him on their plans.

And more. I’ll blog on that later . . .

Tags: , , ,