Where's the outrage?



  • @cartman82 said:

    You can do version control without a version control tool?

    What are Team Foundation Server and Perforce?


  • Discourse touched me in a no-no place

    @HardwareGeek said:

    Try positioning an image at the the top of a column. Notice that it is almost, but visibly not quite aligned with the text at the top of the adjacent column. Move it 1 mm to align with the text. This causes the text that it's invisibly attached to to move slightly, and suddenly the image jumps to the previous page. Undo. Move the image 1/2 mm. Ok, it didn't jump, but your picky client-from-hell (my ex-wife) says it's still not aligned. Move it a little more. It jumps. Undo. Move it half that distance. Ok.

    At that point, wouldn't it be easier to adjust the top-of-paragraph spacing a tad on the other column?


  • Discourse touched me in a no-no place

    @accalia said:

    and if their definition of "good software" differs from yours?

    Well, if the usability sucks, they're wrong.


  • Discourse touched me in a no-no place

    @accalia said:

    Rule one is "Do not act incautiously towards unarmed wizened old men"

    I thought it was "people are stupid".



  • @aliceif said:

    What are Team Foundation Server and Perforce?

    The former is an attempt by Microsoft to make something similar to the suite of Atlassian products, and when I say attempt I mean totally fuck it up and make something that has all the features but is utterly fucking confusing to use (I have used TFS for five years and their Web UI is utterly shit).

    The latter is the result applying occult thinking to a version control system and somewhere there in there is something that keeps track of what version of what is one each machine.

    Both products have version control in there somewhere, but they have another 100 things tacked on the side.

    Git is just version control system.


  • I survived the hour long Uno hand

    @accalia said:

    their definition of "good software"

    There's another fucking problem: people treat quality like it's some soft science. Does the application give you warm fuzzies? There is basic shit you can fucking measure, but people blow it off because it doesn't make them feel good to realize they have dozens of security holes and their appliation runs slow as shit.



  • Any software developer who thinks they can decide if the software they create is good/bad/ugly should not be allowed to develop.


  • Garbage Person

    Okay. So CVS. SVN. Any of the dozens of VCS that came before Linus decided to make Git to solve not the generalized problems of version control but the specific ones for his specific project's specific politics and procedures.



  • Everyone seem to ignore how a free alternative pushes the quality of paid software up, and the price down.

    And bsd code even allow for a conpany to make money improving it's UI, without having to write it from the ground.

    OSS helped Apple and Google to create marketable alternatives to Microsoft monopoly.



  • Which happen to work well teams that want to collaborate globally. TBH Git really didn't do anything new Sun had something similar in the 90s (or so they claimed).

    I think Git works well enough for most people (myself included) and it is pretty much the industry standard in most places now. Bitching and moaning about it because you don't like it isn't going to change that.

    Also most of the alternatives to Git are worse or don't have any substantial benefit.



  • I'm happy that last year I convinced the people in charge to move a project that was developed in ftp for many years to our existing svn server.



  • @Weng said:

    Okay. So CVS. SVN. Any of the dozens of VCS that came before Linus decided to make Git to solve not the generalized problems of version control but the specific ones for his specific project's specific politics and procedures.

    Being distributed was a requirement. The one they used just changed the license on them, so switching to another like Perforce (if Perforce even had distributed capability back then, I don't remember) would be a quite silly move. The only real alternative at the time was Monotone, I think, but I believe it didn't cut it performance-wise.


  • Garbage Person

    Every time I've ever needed DVCS, Mercurial has sucked less in every way.

    Just because something is a standard doesn't mean it's unassailable.

    The entire fucking point of this thread is that the entire fucking tech industry standardizes around suboptimal stuff more or less at random, ossifies, and stops progressing.

    See: JavaScript. Git. Web apps. Markdown. BbScript. SQL. HTTP. Fucking IPv4. SCSI. USB. QWERTY. 3.5mm audio jacks.


  • Garbage Person

    Agile development methodologies. The word Scrum. Open plan offices. MicroUSB charging cables. Touchscreens. Mice. Sexist wages. Image Memes. YouTube. Email.


  • Garbage Person

    DevOps. Continuous Delivery. Regimented and scheduled sprints. Passwords. Password management. Biometrics. Smart cards. Single sign on. Oauth. Security questions. Wish it was two-factor authentication. Two-factor authentication. Captchas.



  • That's the microsoft way. Each org doesn't know what the other is doing. And they change their mind and rebrand everything every year.


  • Garbage Person

    Twitter. How the fuck did we standardize around Twitter? Everyone I know who heard of Twitter in 2006 and 2007 agrees: It is the fucking stupidest thing ever. And yet here we are.

    SMS. MMS. The SF Bay area. India. Eastern Europe. Web conferencing. Telepresence. Telework. VoIP. Bluetooth. PDF. JPEG. MPEG. H264. LPR. Spreadsheets.


  • Garbage Person

    Why do I even work in this godforsaken shitshow of an industry?



  • @Weng said:

    The entire fucking point of this thread is that the entire fucking tech industry standardizes around suboptimal stuff more or less at random, ossifies, and stops progressing.

    No that is your opinion and you are very much in the minority. It is certainly not a fact.

    Good enough in most instances is absolutely fine. I have a "good enough" car to drive to work in, I have a "good enough" telephone and I have a "good enough" television to watch tv and play games. Why it should be any different with regards to development tools?

    Expecting everything to be perfect and be exactly how you like it demonstrates a lack of pragmatism, rather than something wrong with the industry itself.

    EDIT: I actually like the fact that most of the tooling is opensource because I can get it for nothing, and stuff does get fixed and it works and that is all I really give a fuck about.


  • Trolleybus Mechanic

    @Weng said:

    Why do I even work in this godforsaken shitshow of an industry?

    The pay is good
    The working conditions are good
    The management structure is good
    The benefits are good
    The commute is good
    The tasks you work on are enjoyable and good
    The product is good
    The career growth potential is good

    I don't fucking know, actually


  • Trolleybus Mechanic

    Appropos of nothing, but if someone had just done this:

    curl http://somefuckingwysiwyg.example.com/latest > /CommunityServer/javascript/wysiwyg.js

    (paracoded)

    Then we would never have had to deal with Discourse. Ever.


  • Garbage Person

    Perfection is not achievable. Progress is. We gleefully discard lessons learned towards one problem in order to latch onto solutions for unrelated problems.

    And then we arbitrarily decide that our "standard" solutions will be the ones that incorporate the least forward progress. We reinvent everything, all the time, generally less well.

    This is all fucking broken. And it's not about development tooling, either, though that's one of the worst offenders. It's literally everything.



  • @Weng said:

    And then we arbitrarily decide that our "standard" solutions will be the ones that incorporate the least forward progress.

    Good enough with some improvement in a lot of areas has been demonstratively been seen as the best approach, it is called the "The Value of Marginal gains".

    While it is rooted in sport (and it has been shown to work), all the principles can be applied in other industries including ours. You fix or improve one tiny thing in each area and they all tend to add up to make everyone lives easier and is much more attainable than trying to perfect everything.

    A lot of the tooling I have used hasn't changed that much in how it works, but there are lots of small improvements and they have made a massive difference when everything is considered as a whole.


  • BINNED

    @cartman82 said:

    Why shouldn't it be released? Because you don't like it?

    Yes. No software should be released without the Blakeyrat Seal of Approvalâ„¢.



  • @Weng said:

    3.5mm audio jacks.

    Awful awful awful.



  • @flabdablet said:

    NINE YEARS since that fucking thing first appea

    Why are you using the decade-old version and not the one released this year? To ask a stupid question to which the answer is undoubtedly, "I'm a luddite who hates change".

    The ribbon might not be perfect, but it's a shitload better than what came before. And it's also FYI gotten a shitload better since 2007.

    @fbmac said:

    @blakeyrat is complaining about the quality of Microsoft's oss, as if releasing the code reduced it's quality by magic.

    They're made it open source. Simultaneously, the quality has dropped. You explain it.

    In any case, I also complained about the quality drop in Visual Studio 2015, which is not open source, so you're also lying by omission. Or possibly didn't bother reading what I typed.

    @CatPlusPlus said:

    I "release" (what's "released" anyway, just available on the internet, something that was already arbitrarily marked as 1.0, something that made a press release about releasing?) all things I'm not getting paid for because there is absolutely no reason not to.

    I just gave you a reason not to.

    @CatPlusPlus said:

    That's software that is either a) an experiment/research thing or b) something I made for myself for very specific thing

    Ok; as long as you're adding the "this is not production quality" warning to the download page for it, per Cartman82's idea, no problem.

    @CatPlusPlus said:

    (e.g. Git falls into that category as well; that's why it's "released" — it was made for the kernel developers for developing the kernel because alternatives did not cut it).

    Bullshit; they were using an alternative that "cut it" for ages. They created/migrated to Git due to one of the Linux developers being an asshole and ruining their relationship with the vendor who was providing their last source control system. If anything, Git's a surrender, an admission that Linus can't control his asshole developers and stop them from being assholes.

    @CatPlusPlus said:

    Because see, open source is not a development model. It's a philosophy. And it's about sharing code, nothing more, nothing less.

    That's not what most open source fans talk about/tell me it is.

    @CatPlusPlus said:

    By all means call out big projects on not getting their shit together, but assuming any single open source project is striving to be that is a fundamental mistake.

    What about open source products that actively encourage me to use them? Which, again, Git falls under that umbrella. As does GIMP. What's the excuse there?

    @CatPlusPlus said:

    And telling people to not release their code is more harmful than the released code could ever be, even if the code is bad.

    I don't believe that's true and I don't believe it can be true. Perhaps in limited scenarios, like encryption.



  • @cartman82 said:

    You can do version control without a version control tool?

    You can do it without Git. Unless you're :moving_goal_post:, that's all that's needed to disqualify Git from your little rules.

    @cartman82 said:

    Because it was made, it works and a lot of people find it useful.

    So what does it do that couldn't be done before? Again: via your own rules.

    @cartman82 said:

    Why shouldn't it be released?

    Because it's awful, and it gives a bad name to all software everywhere.



  • @jaming said:

    Any software developer who thinks they can decide if the software they create is good/bad/ugly should not be allowed to develop.

    Amen to that.

    If there's one thing I've learned by posting to forums like this one, most developers have no idea what good software even looks like!



  • @fbmac said:

    Everyone seem to ignore how a free alternative pushes the quality of paid software up, and the price down.

    Prove it.

    I'm not "ignoring" it or acknowledging it until I see some evidence that it's provably true.

    @fbmac said:

    And bsd code even allow for a conpany to make money improving it's UI, without having to write it from the ground.

    ... huh? I'm not following you at all here.



  • @lucas said:

    I have a "good enough" car to drive to work in, I have a "good enough" telephone and I have a "good enough" television to watch tv and play games.

    Going from TFS 2013 to Git is like going from a 2015 BMW with all the extras to going to an Edsel.

    Sure, both of them "work" in that they both reach the destination in the same amount of time, but if you were given the choice, would you choose the Edsel?

    (And actually that argument kind of sucks, because TFS is awful software too. All SCM software is pretty awful. It's just good relatively because Git is so much worse.)



  • @blakeyrat said:

    You can do it without Git. Unless you're :moving_goal_post:, that's all that's needed to disqualify Git from your little rules.

    @blakeyrat said:

    So what does it do that couldn't be done before? Again: via your own rules.

    From your interpretation of my rules.

    Otherwise, there'd only ever exist one program from each category. And the moment a better app was made (better as determined by a tribal council presided by Blakeyrat), the old one would be retired, never to be used again.



  • @lucas said:

    While it is rooted in sport (and it has been shown to work), all the principles can be applied in other industries including ours. You fix or improve one tiny thing in each area and they all tend to add up to make everyone lives easier and is much more attainable than trying to perfect everything.

    But as we've discussed on this forum, the open source/Linux development model produces software that can't evolve over time. Because the software uses the CLI as an API, and any change in how the CLI operates would unavoidably break [unknown number] of other programs.

    And it's not like they made this mistake 30 years ago and figured it out and since then it's been all peachy. No, that's the Microsoft way, and they've been suffering for it ever since. That's why they have little messages in the Registry that say, "dude, please, stop using this Registry key, it was only ever valid for like 4 weeks during one of the Windows 95 betas, STOP ALREADY!"

    The Linux way is: nobody's ever recognized or even attempted to solve that problem, ever.



  • @blakeyrat said:

    Going from TFS 2013 to Git is like going from a 2015 BMW with all the extras to going to an Edsel.

    TFS 2013 uses Git in some of it's tooling. I have been told it uses Git secretly behind the scenes even when it is setup as an oldschool TFS library. But why would you use that when using Git with TFS is quite productive, especially when using a sensible branching strategy.

    I think not only are you ignorant, but you are conflating what TFS and what Git do.

    TFS is massive SCM tool with Jira / Trello / CI stuff built in. Whereas Git is just a revision control system. They cannot be compared because they aren't the same fucking thing.

    @blakeyrat said:

    But as we've discussed on this forum, the open source/Linux development model produces software that can't evolve over time. Because the software uses the CLI as an API, and any change in how the CLI operates would unavoidably break [unknown number] of other programs.

    No this is what you have concluded with yourself as a reference, it isn't a fact.

    The part that I have emboldened is utter crap. Sure you can change how the CLI works especially if it is just tooling.
    e.g. you just depreciate flags as you progress or you can version it. There we are another insurmountable problem solved by some common sense.



  • @blakeyrat said:

    Bullshit; they were using an alternative that "cut it" for ages. They created/migrated to Git due to one of the Linux developers being an asshole and ruining their relationship with the vendor who was providing their last source control system. If anything, Git's a surrender, an admission that Linus can't control his asshole developers and stop them from being assholes.

    As far as I remember this was triggered by BitKeeper changing licensing terms first. Might be remembering wrong. Not really important. Creating a new solution because you evaluated existing ones and rejected them (on whatever basis) is not bad.

    @blakeyrat said:

    That's not what most open source fans talk about/tell me it is.

    Primary things created by both Open Source Initiative and Free Software Foundation are licenses and guidelines regarding sharing code. Philosophy is the driving force here. What do they talk about?

    @blakeyrat said:

    What about open source products that actively encourage me to use them? Which, again, Git falls under that umbrella. As does GIMP. What's the excuse there?

    Not defending them, really. For me Git in particular works well enough anyway, and the only other VCS I'd use for anything is Mercurial.

    @blakeyrat said:

    Because it's awful, and it gives a bad name to all software everywhere.

    Meh, commercial software really isn't any better. Most of my annoyances these days are in closed things (fucking Windows 10). Open-source at least will get benefit of the doubt and some patience from me on the account of being open-source. Shit I have to pay for gets no mercy.

    @blakeyrat said:

    Because the software uses the CLI as an API, and any change in how the CLI operates would unavoidably break [unknown number] of other programs.

    You're talking as if it's a rule and as if it's any different with any other kind of API. 😕 Also you do know that CLI-as-API originated in Unix which was not open source?



  • @Weng said:

    This is all fucking broken. And it's not about development tooling, either, though that's one of the worst offenders. It's literally everything.

    The computing industry's Big Lie is that a computer, like a washing machine, is a labor saving device. It isn't. It's a labor displacing device. Over time, the amount of labor required to operate within our increasingly IT-dominated culture stays about the same; given that the ability to sell one's labor is still our primary measure of social worth, this should not come as a surprise. All that IT does is raise the bar: we can do more and prettier things with computers than without, therefore more and prettier things must be done.

    I found one of my Dad's old stock market speculation references down in the back shed the other day. It's called The Intelligent Chartist and it was published when I was three months old. It appears to have been produced on an actual typewriter.

    In 2016, a new book like that would simply get ignored; prospective readers would take one look at the production standards and dismiss it as crank output. Not even business letters are made that way today, because they won't be taken seriously unless they appear to have been properly typeset. Is the information content in today's printed materials of higher quality? Unlikely. But they look nicer.

    This same kind of change has appeared in an endless variety of forms and guises, to the point where now have this ever-escalating spiral of expectations about the things our technological toys ought to be able to do, and we keep building ever-more-complex technological toys in order to do them. My best guess is that the complexity of our technological ecosystem increased beyond any possibility of human comprehension around the turn of the century, and it continues to grow, and we keep behaving as if Fred Brooks were wrong and there really is a silver bullet we can use to avoid being overwhelmed by it.

    Of course it's all fucking broken. We are completely, totally, irredeemably swamped by accumulated technical debt and we simply don't have the aggregate mental capacity to deal with it any more.

    On the upside, by the time any plausible candidate for Singularity is even close to appearing, the self-sustaining bug ecosystem we've been breeding for the last half century will be more than capable of eating its best design efforts for lunch and coming back for more. We will never, never be out-competed by superhuman artificial intelligence because we've done such an excellent job of submerging our entire culture in a foetid, fecund swamp of artificial stupidity.



  • @aliceif said:

    @Weng said:
    3.5mm audio jacks.

    Awful awful awful.

    3.5mm TRRS jacks: multiple incompatible kinds of awful awful awful awful.



  • Before the Internet, releasing broken software that required a patch to fix would be almost a death sentence for a product.

    Internet updates is what I blame for this drop in quality.


  • I survived the hour long Uno hand

    That's fair...

    But how much older crap was released broken anyway?



  • I don't remember finding any bug in any atari 2600 or msx game. Also ms-dos, windows 3.1, wordstar, turbo c, turbo pascal, none of this stuff. You wouldn't expect bugs in normal software usage in those days.



  • @CatPlusPlus said:

    What do they talk about?

    Well for example, if what you say is true, why would an open source fan ever brag about how many devices run Android?



  • @blakeyrat said:

    Why are you using the decade-old version and not the one released this year?

    Because for me there is entirely no difference. Just glancing at the thing makes my eyeballs skid uncontrollably off the screen. SO MUCH hate.


  • I survived the hour long Uno hand



  • @blakeyrat said:

    Well for example, if what you say is true, why would an open source fan ever brag about how many devices run Android?

    I don't know. Probably because it's built on Linux, and some people's understanding stops at that. But I think that'll probably be more Linux fans than open source fans per se.



  • @fbmac said:

    I don't remember finding any bug in any atari 2600 or msx game. Also ms-dos, windows 3.1, wordstar, turbo c, turbo pascal, none of this stuff. You wouldn't expect bugs in normal software usage in those days.

    Wow, bullshit.

    Seriously? Software was just as buggy then. Pac-Man has a rather famous bug that occurs when the player hits level 255 (and no, it wasn't intentional, and no, the fix wasn't too big to fit in the memory space available.) System 7.0.0, contemporary with WIndows 3.11, had a bug where if you uninstalled a font by dragging-and-dropping the font file out of the System Folder, the OS would corrupt itself and have to be reinstalled from the original disks.

    Atari 2600 carts are rather famous for being buggy. The Superman game for example. Oh, and of course E.T., a game frequently accused of being so buggy it actually caused the collapse of the entire home-gaming industry.



  • Maybe my memory is failing me. But ET was a failure, so it kind of helps my hypothesis. There was no patching for cartridge based games.

    Windows 3.1 was something you loaded to use word and excel at it's time.



  • @blakeyrat said:

    Pac-Man has a rather famous bug that occurs when the player hits level 255 (and no, it wasn't intentional, and no, the fix wasn't too big to fit in the memory space available.)

    Come on, did you ever make it to level 255? I didn't play the others.



  • My step sister could make it to level 255 on the C-64 cart version, which was pixel-identical to the arcade version. (Which is to say she was skilled enough to hit the level cap-- every level beyond that is identical. I don't think she ever just sat at the C-64 for like 7 hours to actually reach level 255 though.)



  • Believed to have been, together with friend Chris Ayra, first to reach the ultimate "split-screen" level 256 of Pac-Man in the summer of 1983.[1]


  • You forgot Windows 😉


  • Garbage Person

    And Linux. And FreeBSD. and OpenBSD. And OSX. And Ubanto and Debian and Slackware and..............

    Hey, look, kwalitee!


Log in to reply