a trace of thought on...BizTalk Server, Team Foundation Server, Windows Mobile, etc. RSS 2.0
 Tuesday, January 06, 2009
I had a chance to record another podcast with Jeff Brand, Jason Bock, Rocky Lhotka, and Kirstin Juhl a couple weeks back - you can find it here.

Tuesday, January 06, 2009 10:19:16 AM (Central Standard Time, UTC-06:00)  #    Comments [0] -
Architecture
 Monday, December 15, 2008
Have you had a chance to play yet ?

Many folks have noted that a lot of the Visual Studio elements that have been present within BizTalk to support the development experience are no longer...quite so BizTalk specific!  BizTalk projects now build upon C# projects, and thus a lot of the differences that you used to see in navigating property pages, compilation settings and build mechanics are now gone.  This is a very good thing - it allows you to leverage skills you already have on your team.

MSBuild support is now first class.  Everyone who went through the trouble to install (and invoke) DevEnv.exe on their build server in order to build BizTalk projects will be glad to know that this is no longer required.  (Builds can be ever so much faster when you aren't relying DevEnv...)

Because of the close relationship with C# projects, you can now have C# artifacts directly in your BizTalk projects.  Many people have noted that when you "Add new item...", C# classes aren't offered as an option.  The product group has explained that this is because currently, the designers (such as the Orchestration designer) are unable to provide intellisense for types that are within the same assembly.  So, you are required to use "Add existing item...".

(This sort of reminds me of the interaction between pipelines and schemas - the former require fullly qualified assembly information at run time, which they won't get if you combine these two artifacts in the same assembly.)

If this limitation persists, will the feature get used?  I tend to think so.  There are often cases where the smart thing to do in an orchestration is to delegate to a component...but the work you do in that component is, at times, so specific to the orchestration that it makes sense for them to be co-located for deployment and organizational purposes.  What do you think ?

The support for unit testing is extremely welcome - check it out.  Debugging maps is great as well, but I'm now (often) partial to the external XSLT approach these days.

Monday, December 15, 2008 1:23:55 PM (Central Standard Time, UTC-06:00)  #    Comments [0] -
BizTalk Insights
 Thursday, November 06, 2008

A tool that has been an important part of the BizTalk Deployment Framework has been updated by Loren – see the codeplex project here.

XmlPreProcess is a general purpose tool in its own right for managing configuration files across multiple environments.  The tool has pulled in previously separated functionality (the excellent stuff done by Tom) so that it can consume spreadsheets (that describe environment variations) directly, rather than needing a separate process for that.  Very slick stuff !

 

Thursday, November 06, 2008 9:29:05 AM (Central Standard Time, UTC-06:00)  #    Comments [0] -
Deployment Framework | General
 Thursday, October 30, 2008
I had a chance to sit down with Jeff Brand while down in Omaha for my presentation at the HDC.  We recorded a podcast on all (most) things Scrum and TFS – you can check it out here.  It was great fun – thanks Jeff!
Thursday, October 30, 2008 12:17:25 PM (Central Standard Time, UTC-06:00)  #    Comments [0] -
Team System
 Thursday, October 16, 2008

I’m not able to attend the PDC this year, but its been interesting to watch the definition of Oslo take better shape after watching it for a long while.

I was reading Jon Flander’s post on the topic, where he reiterated the “Oslo as language, model repository, and visual editor aka Quadrant” theme.  The language, known as “M” – draws this quote from John: “visual models are useful for a certain portion of the developer population, but for the most part developers like to write code, which means text.”

Text is good.  I can store text in a version control system, and branch it when needed.  I can merge it with well-known tools later on, and compare the differences between versions easily.  I can have multiple developers work on it at once with a sane reconciliation process upon check-in. 

Even visual models that have a textual representation behind the scenes struggle with these basics…because the text is not first class – it is (often) just a convenient serialization format that is sufficiently opaque for the benefits of text to be lost.

So I’m all for “M”.  It will be fun to watch the vision unfold further. 

Thursday, October 16, 2008 7:25:04 PM (Central Standard Time, UTC-06:00)  #    Comments [2] -
General

I’m down in Omaha today at the Omaha edition of the Heartland Developer’s Conference 2008.  I’m presenting on Scrum with TFS 2008 – with just a little bit more content than at MDC 2008, since both the Conchango Scrum template and the Task Board UI were both updated recently (like…yesterday!) 

This is a great conference – they do a great job organizing this event, and there is a ton of great content.  I did my session this morning, and then had a chance to record a podcast on Scrum/TFS with Jeff Brand later in the day – more on that when it actually posts.

Slides I presented today are here.

Thursday, October 16, 2008 7:00:45 PM (Central Standard Time, UTC-06:00)  #    Comments [1] -
Team System
 Friday, September 12, 2008

I had a great time presenting at this year’s MDC – thanks to all of you who attended.  We covered Conchango’s newest Scrum template for TFS 2008, as well as their very cool new Task Board UI.  We covered eScrum, VSTS Lightweight Scrum, Scrum itself, process customization in TFS…come to think of it, it was a lot of content for an hour and a half…

My slides are available here.

Friday, September 12, 2008 10:31:32 AM (Central Standard Time, UTC-06:00)  #    Comments [1] -
Team System
 Wednesday, June 25, 2008

Suppose you have delivery notification set up on a logical orchestration port (along with a catch block for DeliveryFailureExceptions).

Gotchya 1: If that port is bound to a send port that executes the flat file assembler, and you send improper xml to the assembler, you’ll get a warning in the event log along the lines of:

The adapter "blah" raised an error message. Details "Cannot find definition for the input: foo”

But you won’t get a DeliveryFailureException (or any other, for that matter.)  Nor will the orchestration progress past that point.

Workaround: Consider putting an xml validator as the first step in the pipeline (the flat file assembler will come next.)  That will cause an exception to be raised that you can catch in your orchestration…

Gotchya 2: If the message heading for a send port is terminated within the admin console for some reason, you will never get a DeliveryFailureException, nor will your orchestration progress past that point.

Workaround: No idea on this one.  Leave comment if you do…

Wednesday, June 25, 2008 1:51:37 PM (Central Standard Time, UTC-06:00)  #    Comments [2] -
BizTalk Insights
 Tuesday, June 24, 2008

At this year’s MDC, I’ll be doing a talk on Team Foundation Server 2008 and Scrum – all about where state of the art stands with TFS 2008 and Scrum process integration.  We’ll cover Microsoft’s eScrum, Conchango’s Scrum for Team System, and perhaps a few others to boot.

September 9, 2008 – Marriott Southwest.

The whole conference looks great – check out the speakers and sessions.

Tuesday, June 24, 2008 1:18:27 PM (Central Standard Time, UTC-06:00)  #    Comments [0] -
Team System
Archive
<January 2009>
SunMonTueWedThuFriSat
28293031123
45678910
11121314151617
18192021222324
25262728293031
1234567
About the author:

Scott Colestock lives, writes, and works as an independent consultant in the Twin Cities (Minneapolis, Minnesota) area.

© Copyright 2009
Scott Colestock
Sign In
All Content © 2009, Scott Colestock
DasBlog theme 'Business' created by Christoph De Baene (delarou)