Skip to main content

Debugging Stories: Problems Logging in

I came across an interesting Debugging Story about login credentials that only worked when the user was seated. Whenever they attempted to login while standing, their credentials failed. Until one wise debugger asked the right question. Read the story here, in this book excerpt from Programming Perls by Jon Louis Bentley.

It reminded me of one of my own stories: the new user whose initial password refused to work.

I was part of a small team making a web portal for a medical insurance company. It would let medical practitioners submit predetermination-of-benefits requests before the proposed procedure was performed. We created a first roll-out to three partner offices, with a minimal set of useful features. We called it our Beta release.

Two of the offices began using the system right away. The third could not log in.


We reset and re-sent the credentials for office #3. Nothing changed. The user got a "login failed" message on every attempt.

"Make sure you type the user name and password exactly as they appear," we reminded the user. But still she could not login.

We examined the server-side logs, which were unfortunately pretty sparse on this initial roll-out. But we had enough to see that it was not a connectivity problem, the attempted but failed login appeared in our logs.

We triple-checked our authorization system, to make sure that the user account did indeed exist, with all the same permissions as the other Beta-testers.

With frustration mounting in one-third of our Beta-testers, we decided to send one of our team across town to the medical office, to trouble-shoot the issue on-site. He watched the user enter the credentials a time or two. While he could visually verify that the id was correct, the masked password input field meant that he could not confirm how the password was being spelled.

So he asked the user to please enter the password one more time, but to only use her index fingers on the keyboard. When the user did so, our guy saw right away that the user was inadvertently capitalizing the first letter of the password. Since the password was a case-sensitive mix of lower-case letters and numbers, of course it was failing.

Even after pointing this out to the user, our guy had to audibly shout "fingers up! fingers up!" when she attempted to login again, while using all her fingers. Her unconscious tic was so ingrained that it took multiple attempts before the user could reliably and correctly type the password.

Popular posts from this blog

How to do Git Rebase in Eclipse

This is an abbreviated version of a fuller post about Git Rebase in Eclipse. See the longer one here : One side-effect of merging Git branches is that it leaves a Merge commit. This can create a history view something like: The clutter of parallel lines shows the life spans of those local branches, and extra commits (nine in the above screen-shot, marked by the green arrows icon). Check out this extreme-case history:  http://agentdero.cachefly.net/unethicalblogger.com/images/branch_madness.jpeg Merge Commits show all the gory details of how the code base evolved. For some teams, that’s what they want or need, all the time. Others may find it unnecessarily long and cluttered. They prefer the history to tell the bigger story, and not dwell on tiny details like every trivial Merge-commit. Git Rebase offers us 2 benefits over Git Merge: First, Rebase allows us to clean up a set of local commits before pushing them to the shared, central repository. For ...

Git Reset in Eclipse

Using Git and the Eclipse IDE, you have a series of commits in your branch history, but need to back up to an earlier version. The Git Reset feature is a powerful tool with just a whiff of danger, and is accessible with just a couple clicks in Eclipse. In Eclipse, switch to the History view. In my example it shows a series of 3 changes, 3 separate committed versions of the Person file. After commit 6d5ef3e, the HEAD (shown), Index, and Working Directory all have the same version, Person 3.0.

Trigger Windows Scheduled Task from Remote Computer via Jenkins

One thing I love about working in Information Technology is the opportunity - the NEED - to constantly learn new things. If a week goes by in which I have not looked up something on StackOverflow or other message boards, I start lobbying my team for more challenges. This week, I learned the power of running " SCHTASKS.exe " from a command-line script for a remote server in a Microsoft Windows environment. If you don't know Schtasks, you can read up on it here: https://msdn.microsoft.com/en-us/library/windows/desktop/bb736357(v=vs.85).aspx In a nutshell, it is the command-line interface for the Windows Task Scheduler, and allows you (or a system administrator) to create, change, run, query, terminate, and delete scheduled tasks on a work-station, either the local one or a remote one. Not all of the features are available in older versions. In my scenario below, this was relevant as the local computer will be a Windows 8 machine, and the remote server is, shall we ...

Updating Oracle javapath symlinks on Windows

A Java-based application on my Windows 10 machine recently started prompting me to upgrade my version of Java. Since I wanted to control it myself, I declined the app's offer to upgrade for me, and downloaded and installed the latest Java 8 from Oracle. In my case, Java 1.8.0_171, 64-bit version. The upgrade went fine. But when I launched the app, it again said I needed to upgrade. Why was it still looking at the old location? I made the change using Settings, to change the JAVA_HOME environment variable to point to the location of the new upgrade. But no change, the app still insisted that I needed to upgrade. A little research into the app's execution path showed that it was using c:\ProgramData\Oracle\Java\javapath to find Java. When I looked in that folder, I found symbolic links to my old Java installation. Normally, this hidden bit of information gets updated automatically in the upgrade or installation process. I have read of cases where, when downg...

Abort a Git Merge or Cherry-Pick

Recently a colleague of mine used the Git Cherry-pick feature to bring one of their commits from one branch of our repository to another. They hit a somewhat complex merge conflict and, in trying to resolve all of the conflicts in the file, they got confused about what needed to be done. They came to see me with the question: can they cancel their cherry-pick and partial merge, and start over? The answer is Yes: it is possible to abort a merge or a cherry-pick Git operation. Most of the time it is not needed; with a little work and human intelligence, the merge conflicts can be resolved without too much trouble. But sometimes, in cases like my colleague faced, a more complex merge winds up confusing the developer, and they want to go back and start over. If you use git from the command-line, it’s as simple as: git cherry-pick --abort or git reset --merge On my team, about a quarter of us use Git from the command line, but most - like my colleague in this story -...