We are closing in on CTP5 now and I thought I would give you a little preview of what we are working to deliver in this drop. Lots of UI work has been done around connections, options, and integration points. Greatly increase SQL 2005 support, better F1 help integration, lots more testing and bug fixing. This together with a lot of work overhauling the project system. Overall the product is coming very close to completion from a functional standpoint. As a whole if you have been working with our prior CTPs you should find this a nice step forward.
One of the interesting things during my year here at MSFT has been seeing the amount of time and effort that various teams put into building additional tools. A lot of times this is building power toy like extensions to existing environments or filling in gaps we haven't gotten to yet but there are also numerous attempts around a manner of development proces issues. In some ways it seems like a very frivolous way to go about business but we do get tremendous input from these projects and there is an ample audience that will try and comment about why your tool sucks. This helps drive, in a slow fashion, new functionality to market. Unfortunately it is slow and rather random because a lot of the tools that are created are too special case and they don't have the design, testing, and UX work that is needed to get them out the door.
A lot of the tooling and thought is around agile methods. The ability to go faster is a big concern at MSFT. It is an incredible challenge given the constraints in the business around support, legal issues, integration across products, volume of products and teams, and the complexity of platforms that we must support. I think our team has made good progress on this front but we have had a few special conditions that help: our baseline code--VS whidbey--is a finished product; we have few external dependencies; we have no customers to support, we have a solid focus on an incremental approach to market.
Key long term is not just can we "go faster" it is about can we be transparent, can we be agile in the adpative sense, and can we deliver software in pieces. This last piece, getting to more of a subscription and delivering value over the life of a product will be critical to can we go faster and that will continue to drive this team as well as the rest of team system.
So, looking at the comments I see mostly a tilt towards get it done. I certainly understand a desire for designers, if done correctly, they can be a great boon to productivity. It is the getting them right part that is tricky.
One of the comments talked about this being the best value of Erwin. There has been discussion before about us and modeling and one thing to recognize is that we will have a script import capability. So, you use Erwin, it generates a DDL script, import that to your project and we are ready to go. Not a great round-trip, always synchronized experience but if you have favorite tools you can share information back and forth. This is one of the great values of our product focusing on SQL scripts as the foundation of the truth.
Thanks for all the input so far.
As we have built the TSData story the focus has been around a SQL script experience because the current "designer tools" in VS: Tables, Queries, Databse, are all built for a connected experience, connect to a database modify it, see what happens. Our product is about changing this to a work disconnected, work against code, manage change with version control, have tools that help you understand the impact of your changes etc.
However, we realize that the designer tools do provide a good amount of productivity and are desired. The question at this point is does not having Table Designer really block your ability to use this tool? IE if we were to delay the product launch beyond the planned by the end of the year to add table designer would it be worth it? Or would you rather get the prdocut basically like it exists now with us completing the existing functionality and closing bugs and shipped ontime? Then start on V2 and build really solid data modeling, table design, etc. My natural inclination is to lean towards get the release done and out, stay the incremental agile approach not the just throw one more feature on the bus.
We are looking at what we will put into CTP5 and as we consider this a question arrises. This release may depend on having an instance of SQL Developer Edition installed on the local machine(currently we require SQL Express). However, SQL Dev edition by itself is a large download and presents setup callenges. What we would like to have you do is the following:
- Install VSPro or higher (should already be done if you are using our current CTPs
- Install SQL Dev Edition (if you have VS Pro or better you have the disk already to do this but have to run it separate from the VS setup).
- Run VSTE4DBPro CTP setup to install our bits
- Edit your VS configuration to point at the SQL Dev Edition instance, ie: Tools -> Options -> Database Tools -> Data Connections and change the SQL Server instance name to your SQL Dev
You can do this currently with CTP3 and CTP4. It becomes more important as we continue to fill out our parser/interpretter because it means we can use this instance to validate any enterprise constructs that you utilize. SQL Express doesn't support partition schemas and other Enterprise features but SQL Dev does. Thus if you have a project that utilizes any of these features when we call the sql instance to validate the sql code, the database (if SQL Express) would tell us that it is invalid syntax.
I would love a more "automated" install experience but I want to keep the download size manageable so let me know how this would work for you.
CTP3 Project files can not be opened in on CTP4. If you are moving from CTP3 to CTP4 you will have to create new project files. If you have work that you want to capture a current state of a project you will have to deploy that state to a server and then RE it into a new project. When we made the change in project system to split the file and schema views this changed the project system file layout.
The CTP4 Release is now available for download from http://download.microsoft.com/download/6/8/1/681f2f35-365c-4b47-a1ac-044f9801efb0/TeamDataCTP4.exe
This release features a new project system with a separate project and schema view. You should see greatly enhanced performance when you reverse engineer a database. It also has a large number of bug fixes as we spent the last month focused on product stability. Download it and let us know what you think about the changes on our forum or by logging defects in Connect.
Channel 9 interviewed us a couple weeks ago right before TechEd and the
interview just published up on the site. You get to meet the management team, hear about what we are trying to accomplish in V1 and see a walkthrough of CTP3 (the TechEd bits). CTP4 is just around the corner so stay tuned.
Sachin Rekhi has added his blog to the growing list of team members blogging from Team Data. Sachin is focused on the areas that he currently designs for our team: unit testing, test data generation, refactoring and all the areas around database development.
Cameron Skinner and Gert Drapers on DataDude (#181) DataDude is the code name for Microsoft Visual Studio Team System for Database Professionals, announced shortly before Tech-Ed 2006. Carl and Richard interviewed Cameron Skinner and Gert Drapers from the DataDude team on June 12th at Tech-Ed, in this very informative live .NET Rocks! episode.
We are preparing the CTP4 release. This will relect a sprint in which we focused on fixing things rather than building new features. So, you won't see new capabilities as much as you will see a more stable product. You will see the new project UI and we will be very interested in your input on this. The release should be ready for download on 7/10.
We have just published two new downloads
http://download.microsoft.com/download/5/d/a/5daf1dba-13f4-450a-9005-8001bb638965/TE_Data_CTP3.exe
http://download.microsoft.com/download/5/d/a/5daf1dba-13f4-450a-9005-8001bb638965/VSTE_Data_CTP3_pdi.exe
The first is a self extracting zip file of our setup. You can use this instead of the current CD image file with no need to burn a CD or mount the image. You just get a directory with the setup that you can run. The second file is a new PDI file (this is already included in the zip file) that you can use if you already installed from the CD image. This will remove the problems people have had with repair/reinstall issues.
After we finished CTP3 we decided to do something different. Rather than jumping back on the feature creation stream we decided to keep working on stabilization and closing bugs. So the next CTP we drop won't have more features, just be better. We are making one big change though. In CTP3 the project system has an interesting combination of file and object views. Click on a Table and in the properties you will see that you are pointing to a specific file but other objects like Columns are part of the table, they are not separate files. This system creates a number of issues and we determined that the best course would be to head in a different direction. In the next CTP and going on to ship the database project will become much more like what you would see for a C# project. There is a File view where you can see, manage, and operate on the SQL scripts (and other files) in your project. There is also a Schema view which is like the C# object view. In this view you see the typical database hierarchy of objects Tables which contain Columns, Keys, Constraints, etc. This will mean changes to project files and your existing database project files won't be compatible with the new system. You will be able to drag files from an Explorer window into the project.
DataDude is the code name for Microsoft Visual Studio Team System for Database Professionals, announced shortly before TechEd in June, 2006. Carl and Richard interviewed Cameron Skinner and Gert Drapers from the DataDude team on June 12th at TechEd, in this somewhat shorter-than-usual but very informative live .NET Rocks! episode.
|
|
This is a photo taken by Dan Wygent - a fan from Huntsville, AL - of Richard Campbell, Cameron Skinner, Gert Drapers, and Carl Franklin at .NET Rocks! Live recorded Monday, June 12th, 2006 during TechEd 2006 in Boston. |
http://perseus.franklins.net/dotnetrocks_0181_teched2006.asx
Our own Jeff Welton does a little
podcast and finds that it is interesting "demoing" a tool in an audio podcast.
