• Bradley (unregistered)

    I don't understand what this story is trying to say punchline is. My best guess is that this is one of those Chinese restaurants named "translation server error" things?

  • Bob (unregistered)

    "After a long moment, he closed the editor and the browser, deleted the code from his hard drive, and closed the ticket "won't fix."

    Hopefully adding a detailed reason for not fixing it so the next poor sod didn't do the same thing.

    But yeah, how do you manage bug reports on an application which as deliberate bugs in it :/. Raise them during development I guess and mark them as won't fix (or just mark all bugs for that version of the application won't fix).

  • (nodebb)

    @Bob Why would you deny the next poor sod, all this fun? Not to mention the fact that this is how the next poor sod will learn how to do his job.

  • (nodebb)

    Being new to a team but an experienced developer has to be one of the worst situations ever: You'll constantly be challenged when you say anything about how to improve the code, despite years of experience, nobody will want to listen anything you've learned at other companies and in general you'll be treated like a clueless noob who doesn't know anything.

  • Ex-EMIS Dev (unregistered)

    @DocMonster - if the company puts you through the "we don't do it that way here" then you need to reply "well this is the way I leave. Bye!"

  • (nodebb) in reply to DocMonster

    Well, then you just pack your bags and dash through the doors at good ol' speed o' light.

  • (nodebb)

    Alternatively, mount a charm offensive. Go in with your eyes and ears open, and build relationships with the people you'll be working with. People don't care what you know until they know that you care.

  • Dave (unregistered) in reply to DocMonster

    Or, you'll be treated like a new team member who lacks the social or verbal skills to go with his or her technical expertise. If someone doesn't understand what you're telling them, you can write them off as an idiot, or get better at explaining what you mean. If they won't listen because your approach puts them off, ditto.

  • TrollingMagician (unregistered) in reply to DocMonster

    Not to mention, that if you've changed industries (say, from mortgage finance to healthcare), the customers still expect you to already be intimately familiar with the details of their operations. After all - you're experienced.

  • Altaree (unregistered)

    The thing he learnt was to always search for the error string in the code.

  • OldCoder (unregistered)

    @Bob: It isn't a deliberate bug.

    It isn't even a real bug.

    It is a customer requirement!

    Coded to spec. Doesn't need to close WONTFIX. Should be closed ASREQUESTED.

  • Wyrm (unregistered)

    That is nothing really unusual. Migration projects with the requirements that "existing bugs must be kept in place" is nothing good, but you can't argue with the customer. I can understand why some known bugs can be kept around (mostly things like calculation errors or typos used in automated string look-up), however this one seems just plain ridiculous.

  • Dave (unregistered) in reply to Wyrm

    I have to say, that attitude is responsible for a large proportion of the WTFs we see here. Client management is a skillset worth having. I retired from anything IT related and moved into a completely different field, and earn good money solely from being ok at that.

  • Vic (unregistered) in reply to Wyrm

    Many decades ago I worked on a project where a company was migrating from IBM 1401 to IBM 360 computers. The design was typical for the era - all the data was on tapes, so you'd have something like (sort the transaction tape into the order of the parts tape) (validate the transactions against the parts tape and generate a validated transaction tape) (sort the transactions tape into the order of the vendors tape) (validate the transactions against the vendors tape and generate a validated transaction tape) (sort the transactions into the order of the master transactions tape) (apply the transactions and generate a new master tape and reports).

    We were required to keep the same architecture, which was ok. One advantage was that we could do program level parallel testing, by using copies of the production tapes input to a step, and comparing the outputs.

    One of my programs failed the parallel testing because the final total on the reports didn't match - I had a high-order digit which didn't appear on the production report. I suspected that there was an overflow in the field in the 1410 version, but went through the multi-hundred page report and added all the page totals, which confirmed my suspicions. I explained this to our client's project manager, and he insisted that I drop the high digit and match the report; he suggested that the department might have a reason to not want the high-order digit. I balked, and asked if I could at least speak to someone in the user department.

    The user was spending half their day at a Monroe calculator recalculating the report total, and since it was wrong, they were also recalculating all the page totals. They had asked the IT department years ago to fix it, but were told they were busy and would get back to the user. Some months later they asked about the status, and were brushed off, being told that IT would get to it when they had time and the user should stop asking, So they never did.

    I was allowed to use the correct totals.

  • Vic (unregistered) in reply to Vic

    I also once asked to be allowed to change the format of one of the reports. It was about 500 pages long. Each component used in manufacturing their products was listed, one per line. There were three columns, components in sufficient supply, components in short supply, and components in critical supply.

    Because most components were in sufficient supply, the last two columns were sparsely populated. This meant carefully working through 500 pages to find components which needed to be reordered or manufactured ASAP to address the shortages.

    I suggested that we could take advantage of the memory available on the 360 to "bubble up" the parts in the second and third columns to make it easier to identify those items.

    Again, the PM was against it, but did let me talk to the user. I found an elderly woman who handled the report. Her job was to read the report, and type a page or so each listing the parts in short and critical supply. They didn't care at all about that first column.

    I realized that if I implemented what I suggested she'd be out of a job, and agreed with the PM that we shouldn't change the program.

  • siciac (unregistered) in reply to DocMonster

    You'll constantly be challenged when you say anything about how to improve the code, despite years of experience, nobody will want to listen anything you've learned at other companies and in general you'll be treated like a clueless noob who doesn't know anything.

    No, you're treated like someone with no balls. Take the initiative to improve the code, push it out, and when they cry about it, show them how it's better.

  • Zenith (unregistered) in reply to siciac

    At the risk of summoning Bert, that's a great way to end up unemployed. Happened to me twice and both times it was about inheritance.

    The first place inherited all modules from a base class. Their deployment/versioning system hid unfinished code behind IsDebug() calls that could be registry, configuration, static function, instance function, static property, instance property, or global variable (but never preprocessor directives for some reason). When I saw that, I suggested putting a common IsDebug() definition on the base class. Not long after, "it just isn't working out."

    So then I was somewhere else that used master pages but didn't understand why. This was the place with the global-changing Indian causing tons of surprise rework every week. Well I saw this pattern where I was doing the same stuff on a two dozen pages and moved the relevant code into the relevant master pages. Also fixed a tag soup problem that was causing scrolling issues (they had fits about pages scrolling for any reason). I was the only person working on the only pages that used this master page, mind you. Next morning, the Indian flipped out screaming about a breaking change. There was no such break but he had everybody in such a panic that I was forced to waste three days backing out the code reuse. Wasn't long after that, out the door (I was already interviewing for other jobs but they got ahead of me).

    The crazy part? Both places asked questions about inheritance during the interview. Almost every place does. But nobody knows how to actually use it, outside of the boilerplate generated by the IDE's new form/page wizards. It's like the worst thing you can do some places is actually know everything on their laundry list of wants.

  • Lorens (unregistered)

    This is the first WTF in a long time that literally made me laugh out loud.

  • (nodebb)

    "Padma" is a woman's name.

  • gracioushex (unregistered) in reply to Ex-EMIS Dev

    We had a new guy in that boat recently, very experienced at the tech we use but new to the company. We explicity asked him to find every case of "we don't do that here" and challenge us on it, helped us find several opportunities to improve our environment.

  • Paul Neumann (unregistered) in reply to gracioushex

    And shortly after, he was made redundant?

Leave a comment on “A Hard SQL Error”

Log In or post as a guest

Replying to comment #497147:

« Return to Article