Like a Hand Reaching from the Grave...

I came to Marseille to start some new work, perhaps study some things I'd not had time to do before, and otherwise just have "fun" doing research (i.e. minimal numbers of meetings!!). For the most part, it has been successful. But the last four days...

For better or worse, in research, instead of accomplishing tasks, we take on responsibilities. The difference is that you never quite shed the responsibility. Even if you move onto something else you'll still get called to pick up what you've done. Or, as is often likely, after you've finished a task it becomes your responsibility to maintain it. In short, you are really never done with a project.

Something similar to that has been taking 100% of my time over the last four days. One of my current responsibilities is co-leading the data acquisition group (DAQ Group) at DZERO. The intense, 24/7 work occurred on this project over 4 years ago now (hence the hand from the grave title). It has been so successful it isn't uncommon that a week goes by with very little to do. But this system is key to the experiment's ability to take data. If it breaks, it means whomever is responsible has to drop everything and spend all their time getting it working again.

This happened to a bit of the system that I authored. I think there are sections of this code that are over 10 years old now. The experiment upgraded another component that inexplicably caused my bit problems (don't all bugs start out "inexplicably"?? :-)).

The last four days I've been digging back into that code -- places I'd forgotten exist. Some of it has been fun -- like adding new features to improve its interaction with other components. And some has been really quite boring, like upgrading the code I use to test the component.

But I'm close now. The experiment and accelerator will be down for a full day today (today == Chicago time), and I should have time to put my modifications online.

This code hasn't gone through a major change in over 3 years now. I wonder how many little bugs I've introduced with these improvements, and how much time over the next month or two I'll be spending getting them worked out so it can go another 3 years before it has trouble again?? ;-)

Comments

[Benn Tannenbaum]

and you wonder how many people have used the current "features" of your code and will be negatively affected by your "fixes" upstream...

[Gordon T Watts]

Ha! Right. They certianly have learned behavior patterns based on those "features" and some of those patterns are going get them into trouble after the improvements go in. ;-)