Change

Yesterday I got this week’s issue of Information Week. IBM had taken out four full pages ads in the magazine. That’s a large portion as this is a thin weekly magazine. The theme of these ads was “Conversations for a Smarter Planet”. Here is the kicker. I had no idea what they were talking about.

Two of the four ad pages were giant images. The other two pages had the following headlines: (1) A mandate for change is a mandate for smart, and (2) How can we make our smart systems smarter? These pages then had a lot of text on them. The closing line was “Join us and see what others are thinking at ibm.com/change”.

I figured these ads were just teasers, and that the meat was at the web site. I decided to bite. There seemed to be five bullets at the bottom of the page. Somehow I already got to the second bullet which was about intelligence. The byline was “intelligence to build solutions for business.”

By clicking some buttons I proceeded to the next bullet. The text talked about smarter governance. There were also links to topics like SOA and RFID among others. I saw something about being green and recycling as well.

At this point I had viewed some ads in the magazine, and was trying to navigate the web site. But I still had little clue as to what IBM was selling. There were two further bullets at the bottom of the page which were inaccessible. They were collaboration and infrastructure. No matter how many links I clicked, I could not get these buttons enabled.

Yeah I know IBM is making some good money these days. However I don’t think their latest advertising efforts are working. At least they make not sense to me.

VOB Versus View

I normally deal with only two views for my development work. One view is my development view which holds C++ application source code. The other view is my database view which holds scripts and stored procedures. These two locations for code I work with are also located in separate VOBs.

Whenever I need access to the data from one of the views, I mount the VOB for that specific view. This works fine when I just work with one of these views. However strange things happen when I mount both VOBs. Each view can see the folders in the expected VOB, as well as the other VOB.

This would be cool if I could get access to everything I need through one view. However it does not work like that. If I have the database VOB mounted, but use the development view, I can see the database folders. However I cannot do any work with files in those folders. I have to use my database view to access the files correctly.

What is this? This can’t be the normal way Rational Clearcase works. It just does not make sense. I get that I may need to use more than one view to access more than one type of files. However why is everything visible in the view, but not usable? That seems like something is wrong.

I would not put it past the people who set up my views and VOBs. Perhaps they configured stuff incorrectly. The problem is that I don’t know a lot about the details of how views and VOBs work. However something is wrong in the first place if the tool allows you to set up a bum configuration in the first place.

In the past I never thought I would say this. But I am starting to miss PVCS Version Manager. We had a clunky old version of that software made for Windows 3.1. However there were a lot of things that worked right and easily in that system.

Clearcase View Problems

I continue to have problems with views I do not access frequently. Personally I blame the Clearcase admins who manage our customer's repositories. They got some type of scripts which disable then delete inactive views. I would not mind if my views got deleted and I could recreate them. However once the script does its job, the view remains but I cannot use it.

Normally you need to open a trouble ticket with the help desk to get issues like these resolved. But I have found that is more trouble than it is worth. I took matters into my own hands. I had work to do. First I viewed the properties of the View (no pun). Then I clicked the Advanced tab. I copied down the Global Address of the view.

With the Global Address location in hand, I navigated the network to the location. Each of my views had a folder on the network. Usually I would not like to mess with file structures that Clearcase uses. But I needed to get past this problem. I renamed the directory for my old corrupted view. Then I was able to recreate the view. Problem solved. Something still sucks with this process though.

Clearcase Defaults

Every time I start up Clearcase, the rightmost pane displays a bunch of text starting with the following paragraphs:

Dynamic View

Your project's source files are stored in one or more ClearCase data repositories called VOBs (versioned object bases). This view allows you to access those source files.

Each time you check out, modify, and check a file back in, ClearCase creates a new version that records the changes. Version control is just one of the features that enable your team to manage changes and coordinate access to sources.

The text goes on and on about dynamic views. All of these may be interesting information. However I don’t need to know that any more. I want to see the contents of the select folder in Clearcase.

So each time I launch Clearcase, I have to realize this, find the Windows splitter bar, and move it down to expose the pane with the folder contents. This is very poor design. Why is Clearcase wasting my time? Can’t I set it up to hide this stupid view?

I am going to do some research. There is no reason why I should have to fight with the tool every time I launch it. Shame on the IBM Rational user interface team responsible for Clearcase.

Clearcase Labels

Our project uses Clearcase labels to control versioning in software releases. Development puts the label on files that get changed and need to be sent out to customers. Then configuration management (CM) picks up the files with the labels that development specifies. CM proceeds to use the new files to produce a build. Our project is in maintenance mode. So for the most part, we just change a few files here and there to fix bugs. However this year we added a bunch of new features to the applications. We also upgraded our tools which resulted in a lot of files changing. Development needed to label a lot of files in Clearcase. It was not intuitively obvious how to do this quickly.

Here is the process I normally use to apply labels to changed files in Clearcase. I use Clearcase Type Explorer to create a new label. Then I manually find the changed files using Clearcase Explorer. I get the properties of the file, and click on the Labels tab. At that point I added the label I previously created with Type Explorer. This works fine for a small amount of files. However sometimes we have a lot of files to add.

Previously we had a Java developer on staff who took over our build scripts. In fact he rewrote them using Ant. Part of the build involved putting a label on all the latest files. He figured out how to programmatically label all the files in our project using Ant. Unfortunately this developer has moved on to a new project. So when a current developer needed to label a large directory full of files, he followed in the Java guy’s footsteps and wrote an Ant script to do this.

All of this sounds very strange. I would think it is a common operation to label multiple files in Clearcase. I would hope that the solution is not to label them one by one. And I would also think that IBM does not want every customer to write some custom code to do this automatically. Why can’t I select multiple files in Clearcase Explorer, and apply a label to all of them?

Clearcase Access

IBM provides the Clearcase source control tool as part of the Rational Suite. This product is highly configurable, and perfect for revision control in the enterprise. However some Clearcase policy as demonstrated in the Software Maintenance blog are clearly counterproductive.

Developers make use of different views in the Clearcase environment to work different code sets or version. On bigger projects these views can grow quickly in number. Clearcase administrators need to develop policies to keep the view universe in check. However these policies should be created and enforced to the benefit of the developer user community.

Clearcase is, after all, a tool to be used by software developers. Its purpose is to maintain version control of files. It is meant as an aid to software development. Of course it benefits configuration management as well. However, it must serve its original purpose for it to be put to its best use.

Clearcase Woes

Every time I boot up my computer these days I get an albd server error. This is the first thing that is supposed to run when you start Clearcase. From past experience I know Clearcase Explorer does not work when I get the albd server error. So I manually choose Control Panel in Clearcase. Then I attempt to start Clearcase. For a long time the status is “start pending”. Finally Clearcase is started. So I launch Clearcase Explorer. It comes up so I am feeling better. However when I try to mount a VOB, I get no VOB choices.

This Clearcase has been a headache ever since we switched to using it for source code control. We only do it because our client has standardized on it. Yes I imagine Clearcase is powerful. However I am looking for a solution that works with minimal effort and upkeep from me. My job is not to muck around with Clearcase settings. I am supposed to code new features for our customer, with the occasional debugging of problems in the software we write.

Last time I had Clearcase problems, I went to our local Clearcase guy. He was able to get it working on my machine. However this is not his main job. The right way is to submit a trouble ticket with our client’s help desk. That will take forever. I need my source code control now. So I check out any web pages I can find on the albd server. It seems this process is required for VOBs and views. I could figure that out based on the things not working on my machine.

There is some promise from the IBM Knowledge Collection. They have a whole lot of ideas on how to debug a broke down albd server. The key advice is to check out the logs. So I go to Event Viewer on Windows. I see a whole bunch of Clearcase errors. They fall under to categories: (1) can’t find albd server on a specific host, and (2) can’t find albd. Nice details on that last error description huh? So I guess I need to find out what is wrong with the server that albd cannot find. Too bad that is not a server controlled by our team. I might just have to submit a trouble ticket after all. Bad show Clearcase.