Apr 28 2008

100,000 Airplanes

Published by Jeff under Quality Ready Assessment

A few months ago we started our Quality Ready Assessment (QRA). This is a straightforward web tool to help companies and individuals measure their overall software quality level by asking them a short series of questions around overall application lifecycle. Companies who take this get a customize assessment, and internally we look at anonymize trends in the data.

We’ve gotten a tremendous response so far, and we continue to slice and dice the data to better understand how people who work in and around application development get things done. As a starting point, I’ve looked at the data for SCM usage and found some interesting trends.

The first one that jumped out at me was that 59% of the organizations have an SCM tool that does not support notifications and triggers. This means people are wasting time, plain and simple. This is a classic pull versus push model, where the “puller” is a human needing to know when changes have occurs. Anyone who’s been in the car of a family vacation knows how unsatisfying that kind of “Are we there yet?” conversation is from both sides. Far simpler is to get a call when something has changed.  A simple example of how I personally use this is to have Surround send me an email whenever one of the design or requirement documents change. Since the email includes the name of the file, the comments the person made when checking in the file and even a direct link to the file in Surround I always know when changes happen to the documents I care about.

Another fascinating one is that 55% of organizations have no integration between CM and their issue management solution. Being able to check in code and link that code to the issue that it addresses makes everyone more productive. Developers can mark issues as fixed as part of their normal workflow. QA engineers can see what parts of the application has changed to make their work more targeted. The whole organization is able to learn over time, rather than relying on memory (“What did you do to fix that bug last year?”,”Something with file handling. I think it’s in Foo.vb. Or maybe Bar.vb. Or something like that”.)

Something that won’t surprise anyone is that a recent SD Times survey found the top two SCM tools were SourceSafe and CVS, both of which lack this kind of functionality. While cheap is nice, the time it takes everyone on your team to manually work around these short-comings dwarfs the cost of any tool.
Grant has some related data that show, again to no ones surprise, that companies that use tools with these kinds of shortcomings suffer in the quality and schedule department as well.

The title of this post is the same as an episode of the TV show West Wing, and refers to Franklin Roosevelt’s bold plans for America’s effort around the WW II. Every good company should make big plans for their product. But asking people to achieve them using SCM tools without capabilities like notification and two way integration with issue tracking is like making 100,000 planes using hand tools. You could do it, but it sure isn’t very efficient.

One response so far

Apr 24 2008

Pilot

Published by Jeff under Uncategorized

My name is Jeff Amfahr, and I am the product manager for Surround SCM here at Seapine. After a groundswell of demand (thanks Dad!), I’ve decided to start writing about some of the issues exciting ideas and opportunities around SCM in general and Surround SCM in particular. In addition, I believe it’s only a matter of time before I’m slandered by Grant so I wanted to acquire my own mouthpiece. As Rupert Murdoch already snatched up the Wall Street Journal, this seemed like my best bet.

Over time, I’ll be keeping you apprised of what’s happing with the Surround product, soliciting feedback on directions and features we’re looking at and talking about some of the challenges exciting opportunities of working in this space.

Surround is a very interesting product, since we use it internally. This is a real mixed blessing. On the one hand, when things aren’t working well we feel it ourselves. I get feature requests all the time from team members. Developers “get” almost all of the use cases.

On the other hand, it’s easy to be blinded by our internal development process and use cases. “No one does it that way” and “It’s really important that XYZ be fast as people do that all the time” are easy to slip into the thought process.

I have to try and balance making features that I and other team members want/use to be perfect versus optimizing features that we don’t use, but some large part of our customer base might. And the challenge is how to find out what people are using (and why), since the feedback loop is rarely perfect. Hence, this blog as another avenue to let people in on the fun.

Hopefully you find the information here useful, as I give some additional insight into how Surround is evolving and you give me insight on how it should evolve. If nothing else, if you regularly read this as well as Grant’s postings, you run the risk might have the chance to be deposed at the inevitable lawsuit and/or criminal case.

One response so far

« Prev