• P (unregistered) in reply to KattMan
    KattMan:
    Mike:
    I don't know guys, if you are all so unhappy with the stories why don't you just ask for your money back?

    This site to me is an enjoyable 5 min visit, what is it to you?

    Actually I did ask for my money back, and got every penny I previously sent them. Great guys here!

    Is that all? I got 10x my money back!

  • ejtaw (unregistered) in reply to no laughing matter
    no laughing matter:
    Rather good WTF ruined by story-embellishment!

    Fire that Erik Gern, he won't stop destroying WTFs by his story-writing!

    is what happens when they look for writers by having a creative writing contest....

  • D B (unregistered)

    I didn't finish the story. It is so embellished that I will not finish it. If the whole "WTF" is the communication between two departments then it should never be fictionalized.

    This might be the last time I click my dWTF bookmark...

  • zxbfz (unregistered) in reply to RobFreundlich
    RobFreundlich:
    Mike:
    I don't know guys, if you are all so unhappy with the stories why don't you just ask for your money back?

    This site to me is an enjoyable 5 min visit, what is it to you?

    I second this.

    Besides, I liked the story. Maybe it's because I've worked in Draconia, where interdepartmental politics really were like the Cold War. Or maybe it's just because embellishments can take a rather basic WTF and turn it into a fun story!

    And to preempt all of the cries of "but the WTF sucked they need better WTF's that don't need embellishing": they work with what they've got. If you want better WTF's, SUBMIT THEM YOURSELF!

    I have. Multiple times. And they either get butchered or ignored, so fuck off

  • Juni (unregistered) in reply to FreeMarketFan
    FreeMarketFan:
    If you don't have a good article to post - just don't post anything. Please.

    This was just trash. Pure trash

    CAPTCHA: populus - the populus demanded that this crap never see the front page again

    oh, then we'll have a repeat of the "isn't this the DAILY wtf?" crisis of 2012/2013

  • The Truth, The Hole Truth, And Nothing Butt (unregistered)

    Should there be something in the story which makes a reader say WTF? Or should just the story itself be enough?

  • Norman Diamond (unregistered) in reply to faoileag
    faoileag:
    Krunt:
    faoileag:
    I called IPC via windows registry "innovative". It seems I was mistaken in that evaluation. Even Microsoft thinks information exchange via the windows registry is a valid method of IPC...
    No, it is in fact a terrible idea.
    Where in my post does it say that I find it to be a good idea?
    Krunt:
    And the article you linked is for the .NET Compact Framework. Version 1. From 2006.
    So MS in 2006 did think of it as a valid IPC mechanism for pocket devices. This is 2014 and I said I was mistaken in assuming the method was "innovative". Because Microsoft acknowledged the method way back in 2006. What's your problem?
    The pedigree is longer than that. A design decision seems to be random, whether to communicate between device drivers and user space via DeviceIoControl[*] or the registry.

    [* Name invented so no one will ever guess that the API was copied from the other side of the iron curtain.]

  • Norman Diamond (unregistered) in reply to Tom
    Tom:
    When I submitted changes on both sides to use interprocess pipe, they revoke the change and threatened disciplinary action
    MSMQ, sockets, pipes, COM components, an SQL database, or even drop files would be better than passing application data through the registry.
    I think you accidentally cleared up that mystery. Even though there obviously are Linux programs that manipulate Windows registries, it didn't make any sense in the context of this article.

    I bet their IPC used .ini files. I bet the article's mention of the Windows registry was a failed attempt to anonymize .ini files.

  • Norman Diamond (unregistered)

    By the way, who got those floors backwards?

    .Net comes from a famous Great American Capitalist Company.

    Linux got its start in Suocialist Finland, right next to the old USSR. Ever notice what colour that hat is? How about the name of gcc's preprocessor?

  • A Critic (unregistered)

    Erik Gern makes Mark Bowytz look like William Faulkner.

  • foo AKA fooo (unregistered) in reply to Norman Diamond
    Norman Diamond:
    By the way, who got those floors backwards?

    .Net comes from a famous Great American Capitalist Company.

    Linux got its start in Suocialist Finland, right next to the old USSR. Ever notice what colour that hat is? How about the name of gcc's preprocessor?

    Except Linux wasn't mentioned at all in the story.

  • (cs) in reply to Fritz, a.k.a. Fritzo
    Fritz:
    Oh wow you're completely clueless. People 'whining' is the most entertaining stuff on this site (which reminds me: bring back Mandatory Whining Day) and you want to take it away?

    FTFY

    (reading comments...)

    Oh, I just realised this must be one of those days.

  • Noumenon (unregistered) in reply to The Truth, The Hole Truth, And Nothing Butt
    Should there be something in the story which makes a reader say WTF? Or should just the story itself be enough?

    Ha ha ha, I'm glad I read to the third page of the comments. This should be the featured comment.

  • PB (unregistered) in reply to faoileag

    According to the MS document referenced in an earlier comment either Windows notifications or a P2P message queue (Presumably part of the Windows API) would be more suitable for small datasets. The registry is designed for storing hierarchically structured data but it's best seen as a single-writer model. If the data is fully public (this is questionable), has a simple hierarchical structure and doesn't change often then this may be a good choice. What little can be gleaned from the article suggests otherwise and more importantly, the application writing the data is not the application defining the keys.

  • faoileag (unregistered) in reply to Norman Diamond
    Norman Diamond:
    Linux got its start in Suocialist Finland...
    Nice one, that one... :-)
  • anonymous (unregistered) in reply to P
    P:
    KattMan:
    Mike:
    I don't know guys, if you are all so unhappy with the stories why don't you just ask for your money back?

    This site to me is an enjoyable 5 min visit, what is it to you?

    Actually I did ask for my money back, and got every penny I previously sent them. Great guys here!

    Is that all? I got 10x my money back!

    Is that all? Every time I've bought a page from TDWTF, my money was mysteriously returned to me at the top!
  • Oxin (unregistered) in reply to Mike

    ditto. not sure why every wants to complain about the story being a story. i doubt most of us would even visit if there was only a single sentence each day.

  • Valued Service (unregistered) in reply to The Truth, The Hole Truth, And Nothing Butt
    The Truth:
    Should there be something in the story which makes a reader say WTF? Or should just the story itself be enough?

    I come for the comments.

    But I have to read the story for the context.

  • (cs) in reply to faoileag
    faoileag:
    Norman Diamond:
    Linux got its start in Suocialist Finland...
    Nice one, that one... :-)

    Best comment on this article.

  • steelcobra (unregistered) in reply to Met
    Met:
    If I hadn’t," Tariq said, "and either one of us screwed up a routine, a race condition could corrupt the Windows Registry. We’d have to have our support tech reinstall windows every time the registry corrupts on a user’s machine."

    This article is B.S.

    What method were they using to update the registry? How could a race condition cause you to have to reinstall Windows? If an app can corrupt the registry in a way that requires a reinstall of windows and not the app, virus writers need to start using this exploit so Microsoft will fix it.

    To put it bluntly, the registry should be used as little as possible, preferable only to point permanent keys that never change in the software. Any changes made to the registry affect how the OS runs. If you need to store config settings like what the user's phone number is? Put it in an .ini or .cfg file. Treating the Registry like an open database for data transfer is pretty much kindergarten grade Bad Industry-specific Software practice.

  • last straw (unregistered)

    I've been reading this site for years. I kept reading when you changed the name to "worse than failure", I kept reading when you changed it back, I kept reading even when you had that godawful "mandatory fun day" webcomic. I kept reading through years of steadily more floridly-overwritten and less believable posts.

    This one finally pushed me over the edge. I'm done.

    (I know, I know: I won't let the door hit my ass on the way out. Normally I'd just unsubscribe quietly, but I used to really enjoy this site, I miss what it used to be, and if I were running a site like this and was losing readers I'd want to know why.)

  • (cs) in reply to last straw
    last straw:
    I've been reading this site for years. I kept reading when you changed the name to "worse than failure", I kept reading when you changed it back, I kept reading even when you had that godawful "mandatory fun day" webcomic. I kept reading through years of steadily more floridly-overwritten and less believable posts.

    This one finally pushed me over the edge. I'm done.

    (I know, I know: I won't let the door hit my ass on the way out. Normally I'd just unsubscribe quietly, but I used to really enjoy this site, I miss what it used to be, and if I were running a site like this and was losing readers I'd want to know why.)

    Go to today's article, it's worth it.

  • chiefnewo (unregistered)

    Whoever is writing these, please stop. You're not funny and you can't write.

  • Krunt (unregistered) in reply to faoileag
    faoileag:
    Krunt:
    Smoke signals are a valid method of long-distance communication, but it doesn't make them a good idea when you consider the available alternatives.
    It also depends on what amount of data has to be transferred and whether the communication is bi-directional or not and what ressources you have.

    A smoke/light signal might even in 2014 be the best method to communicate "barbarians and the frontier!" if other methods are not available,

    Which is not what I said. I said if there are other methods available. And in Windows, there are other methods. Almost all of which are better and more suitable.

    The registry is a horrible method of IPC under any circumstances. It is designed for persisting settings, not realtime exchange of information. You're mistaking "innovative" for "retarded" in this context.

  • (cs) in reply to chubertdev
    chubertdev:
    Well, that's part of what I'm saying. You pay off your technical debt. But you do it by improving what you touch, not changing everything at once. If something desperately needs to be refactored, then it should be logged as a defect. When estimating time on a project, there is usually the "X hours" for going with the way it is versus the "Y hours" for refactoring and then making the change on the improved code. Anyone sane should hope for the latter being faster, since it does result in an improved codebase.

    That's bad and either lazy, myopic, or a cowardly approach to any sort of maintenance.

    See, in order to actually fix things and keep them working at peak performance you have to schedule routine examinations where you take a direct look at every portion of the system and evaluate how they interact. And if you find that yes, while each individual portion of the system is working ok, the overall system needs a full rebuild, then that's what you do.

    You can't just rely on fixing things as you find them, or fixing things as it comes up while working on other work. You have to actually go looking for the places where the system no longer works. Because often the broken bits aren't just in the things that users report, or the stuff that's easy to sum up and present to the business as an obvious defect. It's more often in things like having a framework that was never designed to be extensible, which now results in developers having to create workarounds and kludges, which results in an increased but not-quite-measurable percentage of defects and bugs as each workaround tinkers with the system in slightly ideosyncratic ways. Or in an increased likelihood that 10 years down the road some intern is going to need to make a change and fuck it up because the code isn't easy to change or understand. You can't take those things to the business on a case by case basis because by then it'll be too late, and right now the need isn't obvious.

    We, as developers, who strive to not just write code, but to design and write GOOD code, have an ethical responsibility (in the sense of professional ethics) not be complacent, or just sit back and do things badly because we are afraid of upsetting the paymasters. Why? Because ultimately it does both us and them a disservice. They get code with a higher likelihood of breaking, we get a reputation for writing broken code.

  • (cs) in reply to Snooder
    Snooder:
    chubertdev:
    Well, that's part of what I'm saying. You pay off your technical debt. But you do it by improving what you touch, not changing everything at once. If something desperately needs to be refactored, then it should be logged as a defect. When estimating time on a project, there is usually the "X hours" for going with the way it is versus the "Y hours" for refactoring and then making the change on the improved code. Anyone sane should hope for the latter being faster, since it does result in an improved codebase.

    That's bad and either lazy, myopic, or a cowardly approach to any sort of maintenance.

    See, in order to actually fix things and keep them working at peak performance you have to schedule routine examinations where you take a direct look at every portion of the system and evaluate how they interact. And if you find that yes, while each individual portion of the system is working ok, the overall system needs a full rebuild, then that's what you do.

    You can't just rely on fixing things as you find them, or fixing things as it comes up while working on other work. You have to actually go looking for the places where the system no longer works. Because often the broken bits aren't just in the things that users report, or the stuff that's easy to sum up and present to the business as an obvious defect. It's more often in things like having a framework that was never designed to be extensible, which now results in developers having to create workarounds and kludges, which results in an increased but not-quite-measurable percentage of defects and bugs as each workaround tinkers with the system in slightly ideosyncratic ways. Or in an increased likelihood that 10 years down the road some intern is going to need to make a change and fuck it up because the code isn't easy to change or understand. You can't take those things to the business on a case by case basis because by then it'll be too late, and right now the need isn't obvious.

    We, as developers, who strive to not just write code, but to design and write GOOD code, have an ethical responsibility (in the sense of professional ethics) not be complacent, or just sit back and do things badly because we are afraid of upsetting the paymasters. Why? Because ultimately it does both us and them a disservice. They get code with a higher likelihood of breaking, we get a reputation for writing broken code.

    Do you work at a consulting shop, where you have loads of free time to fix thing? And every app that you've ever worked on is something that you created?

    Because, in the real world of big companies, it's only the squeaky wheel that gets the grease. You inherit a steaming pile of junk, you fix what's tragically broken, and make changes for what's demanded. The world isn't going to wait for you to gold plate your code.

  • SQB (unregistered) in reply to faoileag
    faoileag:
    Two departments in the same company not talking to each other although they sometimes have to cooperate? Yawn. Been there, done that. Haven't found a suitable t-shirt yet.
    "I haven't talked to the other department, and they didn't even get this lousy T-shirt"?

    (Captcha: caecus. Nothing funny about that. There's just some things you shouldn't joke about.)

  • 1234 (unregistered) in reply to last straw
    last straw:
    I've been reading this site for years. I kept reading when you changed the name to "worse than failure", I kept reading when you changed it back, I kept reading even when you had that godawful "mandatory fun day" webcomic. I kept reading through years of steadily more floridly-overwritten and less believable posts.

    This one finally pushed me over the edge. I'm done.

    (I know, I know: I won't let the door hit my ass on the way out. Normally I'd just unsubscribe quietly, but I used to really enjoy this site, I miss what it used to be, and if I were running a site like this and was losing readers I'd want to know why.)

    Yeah, I remember all of that too and I agree. I knew this story was in trouble with the over the top soviet-ish stuff in the first paragraph. But the big, "oh great another one" moment came when Simone turned into Wanda for a sentence.

    I used to love to come here, but the stories are just plain bad now. I mean I know that there will be some over dramatizing, but this one... cartoony stereotypes, a protagonist that no one cares about, no real WTF or if there was one it was hard to cut through all the garbage to see.

    But all the hyperbole aside, there's just no editing being done. The GHOST story had something named SPRIT that was SPIRIT later in the text. This one has the Wanda/Simone thing. Just bad form for story publishing.

  • (cs)

    The only good part of this article was having Galactus stare at me while I was reading it.

  • Anonymous (unregistered)

    Jesus Christ, what happened to TDWTF?! This is ridiculous. I don't even know why I keep coming back, expecting to learn something useful. That doesn't happen anymore. It's like some kind of amateur short story writing group now... These new authors are TRWTF.

  • Neil (unregistered) in reply to Tom
    Tom:
    20 minutes later, the post routine has written 2^32 records to the registry, filling up the entire hard drive and causing the server to grind to a screeching halt.
    Huh, I never realised that they removed the registry limit in an XP service pack.
  • Hannes (unregistered) in reply to Hanzo Is God
    Hanzo Is God:
    I rather read Hanzo stories than this.

    Sad thing is, I was thinking the same.

    No, actually I thought: "This is the worst TDWTF I've ever read".

    Story telling, and "anonymization" and "artistic freedom" are all fine and stuff. But this is far over the top and it really kills the WTF.

  • Mike (unregistered)
    Comment held for moderation.

Leave a comment on “Process Communication Hell”

Log In or post as a guest

Replying to comment #:

« Return to Article