One of the complaints about the mysterious problem codes is that we have no glossary to explain what they mean. All that we really need, though, is the ability to associate the individual problems with each other, and distinguish the different problems from one another. We don't care about the universe of all possible problems, just the problem that we're looking at right now.
If we're looking at someone who's reporting a bX-wj8w1r, maybe we want to see everybody else with the same problem. So we search on bX-wj8w1r. We can see all of the cases found, examine the stated symptoms en mass, and see what the actual problem is. Since we know that nobody describes their symptoms the same way, seeing all of the reports of the same problem, together, is much easier if we use the code.
One of the problems with problem analysis is that dissimilar problems get lumped together, but similar problems are never reported together. It's impossible to analyse "login problems", because there are so many different problems, yet so many ways that any specific login problem can be described.
Look at my article Help! My Blog Is Gone! for an example of how many symptoms might be reported for one problem, yet how many individual problems might cause one symptom.
If we assign a "bx-" code to each specific problem, we can aggregate the problems and find out, at any time, how many people are being affected by a malfunctioning login script. Maybe we'll find that the people reporting a new problem just updated their personal firewall, which is now blocking a login script from running.
Another use for the codes is the ability to see what new problem is being reported. If we simply search for "bx-", and look at the most recent reports, we can see that.
The value of the "bx-" code is that it's a unique string. You'll probably not find "bx-" part of anything but these problems. If I searched for "rat" on the web, I'd find hits from "borat", "congratulations", and "rationalization", among others. If I search for "bx-", it's pretty likely that all hits will be people discussing a Blogger problem.
That will make it possible to script a search, and compile a dynamic glossary at any time. Just search for "bx-", sort by date descending, and see what you get.
>> Top
If we're looking at someone who's reporting a bX-wj8w1r, maybe we want to see everybody else with the same problem. So we search on bX-wj8w1r. We can see all of the cases found, examine the stated symptoms en mass, and see what the actual problem is. Since we know that nobody describes their symptoms the same way, seeing all of the reports of the same problem, together, is much easier if we use the code.
One of the problems with problem analysis is that dissimilar problems get lumped together, but similar problems are never reported together. It's impossible to analyse "login problems", because there are so many different problems, yet so many ways that any specific login problem can be described.
Look at my article Help! My Blog Is Gone! for an example of how many symptoms might be reported for one problem, yet how many individual problems might cause one symptom.
If we assign a "bx-" code to each specific problem, we can aggregate the problems and find out, at any time, how many people are being affected by a malfunctioning login script. Maybe we'll find that the people reporting a new problem just updated their personal firewall, which is now blocking a login script from running.
Another use for the codes is the ability to see what new problem is being reported. If we simply search for "bx-", and look at the most recent reports, we can see that.
The value of the "bx-" code is that it's a unique string. You'll probably not find "bx-" part of anything but these problems. If I searched for "rat" on the web, I'd find hits from "borat", "congratulations", and "rationalization", among others. If I search for "bx-", it's pretty likely that all hits will be people discussing a Blogger problem.
That will make it possible to script a search, and compile a dynamic glossary at any time. Just search for "bx-", sort by date descending, and see what you get.
>> Top
Comments