Is this common with today's interviews?



  • @dhromed said:

    @ASheridan said:

    a lot of people get confused, and don't understand that there is a difference.
     

    What are the use cases where the difference becomes a problem?

     

     

    I gave a very clear use case in a post not too far above this one.

     



  • @blakeyrat said:

    Read the article I linked.If you read his essay, you'll see that with minor changes (making all non-printing characters illegal in filenames; changing the IFS separator list; decreeing that all filenames must be stored as UTF-8 all the time), all of the currently broken shell scripts suddenly start working as if by magic, with no change to the shell language.

    I guess - after reading it (and I don't disagree with what he proposes, but unfortunately reality means it sadly ain't gonna happen), my point was simply that sometimes the shellscripts are broken because of specific characters in filenames, but some are broken because of the short-sightedness of the developer to cope with those particular characters. Yes, I know it's impossible to cope with ALL of them, but the fact that some apps are tolerant of "strange" characters and others aren't indicates different thought processes during the design stage for the input module.

    @blakeyrat said:

    There's a certain class of people who spend all their work time pretending to be 5 times more busy than they actually are, to avoid getting more work assigned to them, or in a misguided effort to make themselves appear more critical to the business. He could just be one of those people.

    Obligatory link to a Dilbert cartoon featuring workshy-Wally...
    Yeah, I'm convinced that the "I'm too busy" is also an oft-used fob-off for "I just can't be bothered to see those people." It happens.

    @blakeyrat said:

    And your griping about the performance issues are pretty out-of-date, when did you try it, Windows 2000?

    Prior to that - I worked at a place where we put in NT4 + Small Business Server boxen for financial apps. We flicked compression on to give that "you've paid for a 10G disk but really you're getting up to 20G of storage" sales bollocks. On some machines where compression wasn't enabled, we noticed backups were much faster - further investigation indicated that hardware compression on the tape drives relieved the CPU.

    I'm guessing with the oompf that more modern kit provides, it's a moot point now. I've not really used Windows much since the 2000 days.

    @dhromed said:

    I disagree. You may've been brought up with the idea that zip files are files, but conceptually, they're closer to folders than files.

    I've been bought up with the idea that zip archives appear as a single file (and can be attached to an email, have their contents amended, etc) but conceptually they're a single object that contains many objects within. I see RAR, GZIP, TGZ, TAR in the same fashion. I suppose a similar concept is an excel file being a workbook containing many excel spreadsheets, rather than just singular.

     



  • @Cassidy said:

    (and I don't disagree with what he proposes, but unfortunately reality means it sadly ain't gonna happen)

    That's what I meant when I was talking about the broken culture. The culture is so broken that an obvious deficiency, with security, usability, automation, "heisenbug" implications will never get fixed. If you're ok with that, then fine. I'm not, personally.

    @Cassidy said:

    my point was simply that sometimes the shellscripts are broken because of specific characters in filenames, but some are broken because of the short-sightedness of the developer to cope with those particular characters. Yes, I know it's impossible to cope with ALL of them,

    It's literally impossible to write a shell script that can traverse a list a files correctly. It's not a case of, "oh the developer of this particular script did the wrong thing", it's a case of, "it's impossible to do the right thing using the standard tools". (IIRC, he did come up with a working solution that relied on non-POSIX tools but even that only worked with certain commands. Staying within the realm of POSIX, though, it's impossible.)

    @Cassidy said:

    I'm guessing with the oompf that more modern kit provides, it's a moot point now. I've not really used Windows much since the 2000 days.

    Then maybe not post your experience from < 2000 as if it's still relevant today? You've triggered one of my major pet peeves! "NTFS compression is too slow!" "When's the last you tried it?" "Over a decade ago." CHRIST!



  • @ASheridan said:

    You're right of course, a zip is a file and not a directory, so why doesn't Windows Explorer suggest that? I mean, it even gives it an icon that looks like a directory, and opens it up as if it were a directory, and as Windows hides file extensions by default, a lot of people get confused, and don't understand that there is a difference.

    If Microsoft did manage to make it work across the whole OS, I guess it would work the same way as it does on KDE (generally my preferred window manager, just because of features such as this) which uses KIO slaves to handle things like zip files (and other archive formats) as if they were local files. Very convenient and one of the main features I wish would make its way into Windows at some point. The current .zip archive situation in Windows is a bit broken in my opinion.

     

    It keeps that distinction by virtue of only Explorer being able to do it.  And I'm sorry but if you're working or even using a computer in 2012, you know what a zip is.  My 86 year old grandmother knows what a zip is.

    Plus, HTML aside, if you tried to run an exe from inside a zip, explorer does pop up a dialog letting you know that the files are archived, and you should probably extract them before running.  And even extracts them to a temp directory so you don't mess up your desktop, and cleans them up afterward.  

    I'm all for user friendliness, but at some point the users do need to use their brains.  Even the best software in the world can't do what you guys are asking it to do.

     


  • ♿ (Parody)

    @Master Chief said:

    Even the best software in the world can't do what you guys are asking it to do.

    Why do you believe this? Is there some justification for this? Are you saying that the example already provided (kioslaves) doesn't exist? Do you also deny the ability to mount an ISO file and use it as though it were part of the file system? What's the conceptual difference between a ZIP archive and an ISO image? What about the disk images used by VMs? Or encryption software like TrueCrypt? Are you posting from 1982?



  • Ok so I'm on KDE and using "kioslaves" whatever the fuck those are. Advocating for the devil.

    I open my email client. I try to attach a .zip file to an email. I get an error saying, "you can't attach a folder to an email, you can only attach a file."

    Right? Or is there some magic juice in "kioslaves" that makes .zip files behave like folders when they should but also behave like files when they should? And how does it know when to do A and when to do B? Or is the solution just that the KDE people had to go and manually fix every single email program on Linux to do attachments right?



  • @blakeyrat said:

    If you're ok with that, then fine. I'm not, personally.

    I didn't say I was okay with it; I think I'm more resigned to the fact that it's a shitty situation which ain't gonna change, meself, so I won't waste effort seething about it.

    @blakeyrat said:

    Then maybe not post your experience from < 2000 as if it's still relevant today? You've triggered one of my major pet peeves! "NTFS compression is too slow!" "When's the last you tried it?" "Over a decade ago." CHRIST!

    Erm... I thought I mentioned it was historical, and that the subject had been done to death back then.  Perhaps I didn't make that too clear, sorry. Is NTFS compression used/commonplace under modern Windows, then?

    Either way, the original point of my experience is that Windows handled compression/uncompression transparently - although, as you mentioned, this was a feature of the underlying filesystem and not of the file mamagement tool - so it should be possible for Microsoft to extend that idea to ZIP archives.



  •  @blakeyrat said:

    Ok so I'm on KDE and using "kioslaves" whatever the fuck those are. Advocating for the devil.

    I open my email client. I try to attach a .zip file to an email. I get an error saying, "you can't attach a folder to an email, you can only attach a file."

    Right? Or is there some magic juice in "kioslaves" that makes .zip files behave like folders when they should but also behave like files when they should? And how does it know when to do A and when to do B? Or is the solution just that the KDE people had to go and manually fix every single email program on Linux to do attachments right?

    Actually, the behavior is dependant on what you're using to interact with them. If I'm attaching a file to an email, it will attach the whole zip file, and not try to open it, which I think is the preferred behavior here. If i'm in my file browser (say Konqueror, or Dolphin) the zip archive opens as if it were a directory, and from there I can use the context menu to attach an individual file from the archive to send via email, or drag and drop into my email client, or I could just choose to edit the individual file, and the kioslaves let the app treat it as a regular file. All in all, I'd say the behaviour definitely makes things easier, it's pretty logical and fairly seamless.

    I think even you might be impressed if you used them!



  • @PedanticCurmudgeon said:

    @blakeyrat said:
    @PedanticCurmudgeon said:
    @blakeyrat said:
    Or Linux's allowing virtually any character in a filename. Or Linux's pig-headed insistence that "everything is a file". Or... pretty much all of Linux is an example of this, frankly.
    Have you read the Unix Hater's Handbook? People have actually been complaining about these things for decades.

    Yes, but the real WTF is that the culture is so broken that none of the issues ever get fixed.

    The key word here is culture. Cultures generally have customs that may seem horribly broken to outsiders. Have you considered the possibility that there's a good reason the issues never get fixed: they're not unmanageable problems for people within the culture?
    But it is a problem, even for people well versed in the *nix "culture".  They know it's a problem, they admit it's a problem. 

    Ultimately, the problem isn't really "culture".  The problem is that long, long ago somebody made some really, really stupid design decisions.  Seriously, what sane person thinks it's a good idea to allow a filename to contain a backspace, carriage return or other such nonsense?  And after a few years go by it gets harder and harder to fix the problem.  Not because they don't know that there's a problem and not because they don't want to fix the problem, but because they're afraid of breaking something .  The end result is that now, people have to jump through all sorts of ridiculous hoops to make sure that things work properly.

    The article that Blakey linked to points out that there are things you can do (making all non-printing characters illegal in filenames, etc), but, none of that should be necessary, because the operating system never should have allowed it in the first place.


  • ♿ (Parody)

    @blakeyrat said:

    Ok so I'm on KDE and using "kioslaves" whatever the fuck those are. Advocating for the devil.

    I open my email client. I try to attach a .zip file to an email. I get an error saying, "you can't attach a folder to an email, you can only attach a file."

    I don't see how that would happen, unless you have an email client that's really aggressive with that sort of thing.

    @blakeyrat said:

    Right? Or is there some magic juice in "kioslaves" that makes .zip files behave like folders when they should but also behave like files when they should? And how does it know when to do A and when to do B? Or is the solution just that the KDE people had to go and manually fix every single email program on Linux to do attachments right?

    I just tried it with some local, static html + image files. I zipped up the directory. Kioslaves basically look like any other protocol, like file or http or whatever. Interestingly, this works for me in konqueror, but not dolphin (well, surely not interesting to blakeyrat, but...). Dolphin behaves like Windows Explorer, so I can easily navigate into the zip file, but it seems to extract to a temp folder rather than actually using the zip kioslave. But in konqueror, my path changes from the normal file path to starting with "zip:" plus the rest of the path. I was able to go into the html pages, see the images and navigate around in them as though they were sitting on the file system by themselves.

    It seems that the browser needs to also use the kioslave, because when, from konqueror, I told it to open the file in chrome, I got the old temp file, broken navigation behavior. And I can't simply enter the zip: protocol into chrome and get it to work. Chrome and opera both do a search, firefox tells me that it doesn't know what to do, since the zip: protocol isn't associated with anything.

    So it's not as seamless as mounting something like an ISO image as a filesystem, but I suppose it's more ad hoc. There are other types of kioslaves, so that you can (e.g.) open files over ftp or sftp or make a normal audio CD look like the tracks are mp3s (or wavs, or whatever).


  • BINNED

    @blakeyrat said:

    Ok so I'm on KDE and using "kioslaves" whatever the fuck those are.

    So you actually use Linux. Interesting. I remember discussions a few years back, in the days before Ubuntu, about increasing user adoption of Linux. The point was made that an enlarged user base meant that most of the new users would be from outside of the culture and would have little or no interest in adapting to it. This was held to be a bad thing. I guess that prediction has proven to be correct.

    Anyway, the point I was trying to make earlier is that the core set of Linux (and Unix) users actually likes things the way they are. If the culture seems broken to you, that may just mean that you're an outsider.


  • ♿ (Parody)

    @El_Heffe said:

    Ultimately, the problem isn't really "culture".  The problem is that long, long ago somebody made some really, really stupid design decisions.  Seriously, what sane person thinks it's a good idea to allow a filename to contain a backspace, carriage return or other such nonsense?  And after a few years go by it gets harder and harder to fix the problem.  Not because they don't know that there's a problem and not because they don't want to fix the problem, but because they're afraid of breaking something .  The end result is that now, people have to jump through all sorts of ridiculous hoops to make sure that things work properly.

    I imagine that originally, the thought was that it wasn't worth the resources (back when things were really tight) to enforce stuff like that. Because, hey, who would do something stupid like that? Like almost every other aspect of computers or computer related behavior, enough people don't need a good reason to do something stupid. And since most sane people continue to stay away from the behavior, it's not a high priority to do something about, since it would have the negative consequences of breaking certain systems that rely on the behavior.

    In the end, it's probably a rarer problem than the janitors unplugging the computer at night, but it's a convenient thing to whip out as a demonstration of why something that some people dislike is The Worst System In The World.



  • @ASheridan said:

    If I'm attaching a file to an email, it will attach the whole zip file, and not try to open it, which I think is the preferred behavior here.

    Yes, but how does the email client know it's a single file, if KDE is telling it it's a folder? Or, more generally, how does KDE know whether to tell arbitrary application X that it's a file and arbitrary application Y than it's a folder?

    @ASheridan said:

    it's pretty logical and fairly seamless.

    If it's seamless it seems to me that's only because the KDE folks went and patched a whole bunch of third-party apps to support it seamlessly. Microsoft doesn't have that option in Windows.



  • @PedanticCurmudgeon said:

    Anyway, the point I was trying to make earlier is that the core set of Linux (and Unix) users actually likes things the way they are. If the culture seems broken to you, that may just mean that you're an outsider.

    And if Linux fans were going around saying, "oh it's just a hobby OS, it's not professionally engineered and there's really nobody interested in fixing its major flaws", that would be fine. The problem is, the Linux fans are going around saying, "Linux is better-designed than Windows and you should use it for all your servers and all your desktops and basically it's great, it's so superior to the competition." That's where the problem comes from.

    The Linux culture treats it like it's a hobby OS, not being used for serious work. But at the same time, they promote it as being a robust, solid piece of software that should be running tons of servers. You can't have it both ways!


  • BINNED

    @El_Heffe said:

    @PedanticCurmudgeon said:

    @blakeyrat said:
    Yes, but the real WTF is that the culture is so broken that none of the issues ever get fixed.
    The key word here is culture. Cultures generally have customs that may seem horribly broken to outsiders. Have you considered the possibility that there's a good reason the issues never get fixed: they're not unmanageable problems for people within the culture?
    But it is a problem, even for people well versed in the *nix "culture".  They know it's a problem, they admit it's a problem. 

    Ultimately, the problem isn't really "culture".  The problem is that long, long ago somebody made some really, really stupid design decisions.  Seriously, what sane person thinks it's a good idea to allow a filename to contain a backspace, carriage return or other such nonsense?  And after a few years go by it gets harder and harder to fix the problem.  Not because they don't know that there's a problem and not because they don't want to fix the problem, but because they're afraid of breaking something .  The end result is that now, people have to jump through all sorts of ridiculous hoops to make sure that things work properly.

    The article that Blakey linked to points out that there are things you can do (making all non-printing characters illegal in filenames, etc), but, none of that should be necessary, because the operating system never should have allowed it in the first place.

    Let me get this straight. You're saying that the only reason they haven't prohibited non-printable characters in file names is because they're afraid of breaking something? What exactly do they think would break?


  • ♿ (Parody)

    @blakeyrat said:

    @ASheridan said:
    If I'm attaching a file to an email, it will attach the whole zip file, and not try to open it, which I think is the preferred behavior here.

    Yes, but how does the email client know it's a single file, if KDE is telling it it's a folder? Or, more generally, how does KDE know whether to tell arbitrary application X that it's a file and arbitrary application Y than it's a folder?

    You have to actively use the kioslave. The app might do that when you click on the file (like file managers do), but generally, the result from the open file dialog would simply be the path to the file, and presumably your email client would treat it like any other file you clicked on. Just like you wouldn't expect it to open your media player if you selected an mp3.

    @blakeyrat said:

    @ASheridan said:
    it's pretty logical and fairly seamless.

    If it's seamless it seems to me that's only because the KDE folks went and patched a whole bunch of third-party apps to support it seamlessly. Microsoft doesn't have that option in Windows.

    I don't know much about the underlying implementation and how it's used at that level, but the key seems to be to have the protocol registered with the kioslave, and for the app to hook into however that happens. Browsers seem to have this general capability. Either way, it doesn't seem terribly unreasonable to need to be a KDE based app to use a KDE based feature.

    I couldn't say if you could register kioslave protocol for non-kde apps, but the analogy seems that on Windows, you could register the protocol with the Win32 API, so that anyone using the Win32 API could take advantage. Or maybe shell extensions are a better analogy. It may be that there is no good way to implement this sort of feature in windows. But that's a far cry from, "This is impossible!" that has been expressed in this thread.

    But I believe that one of Windows 8's new features is built in ISO mounting. I'll ask again: Why that's any different conceptually from mounting a zip archive? It really just sounds like something that was initially implemented at the wrong level (i.e., Explorer) to realize its full potential, and either now it's too hard to go back and fix, or the changes aren't worth the marginal improvement.



  • @blakeyrat said:

    Yes, but how does the email client know it's a single file, if KDE is telling it it's a folder? Or, more generally, how does KDE know whether to tell arbitrary application X that it's a file and arbitrary application Y than it's a folder?

    Generally, an individual app will treat it as a file, because you're browsing using the standard file:// protocol. Within some apps (like Konqueror) it will automatically use kioslaves as you're browsing files because it assumes (correctly in my opinion) that you'd rather treat them as a directory as opposed to a file in that context. A file dialogue opened from within an application though expects, and is given, a file. However, if you open a zip as a directory within Konqueror, and then right-click on a file within that and edit it (a text file for example) then the system passes the file across to the editing application using a kioslave, which the app sees as a local file, so when you save back it's saving into the archive directly.

    @blakeyrat said:

    If it's seamless it seems to me that's only because the KDE folks went and patched a whole bunch of third-party apps to support it seamlessly. Microsoft doesn't have that option in Windows.
     

    The KDE folks didn't patch any third party apps to make this work, the apps just make use of the kioslaves because that's what the OS makes available. Microsoft does have that option in Windows, I just can't fathom why they've not taken advantage of it yet, as it's a huge time-saving bit of functionality of the OS.

     


  • ♿ (Parody)

    @PedanticCurmudgeon said:

    Let me get this straight. You're saying that the only reason they haven't prohibited non-printable characters in file names is because they're afraid of breaking something? What exactly do they think would break?

    I believe that there are systems that use the filesystem as a database, and regularly make use of non-printable characters. You obviously don't read Raymond Chen's blog, because you have no imagination about the stupid things that get perpetuated in the name of backward compatibility.


  • BINNED

    @blakeyrat said:

    @PedanticCurmudgeon said:
    Anyway, the point I was trying to make earlier is that the core set of Linux (and Unix) users actually likes things the way they are. If the culture seems broken to you, that may just mean that you're an outsider.

    And if Linux fans were going around saying, "oh it's just a hobby OS, it's not professionally engineered and there's really nobody interested in fixing its major flaws", that would be fine. The problem is, the Linux fans are going around saying, "Linux is better-designed than Windows and you should use it for all your servers and all your desktops and basically it's great, it's so superior to the competition." That's where the problem comes from.

    And this is precisely where the Linux culture is broken. You see, the fanboys you mention aren't really a part of the core culture and largely don't understand the history; if they did, they certainly wouldn't claim that Linux is for everybody. And this is in fact part of the problems foreseen by others that I mentioned in my earlier post about the subject.

    @blakeyrat said:

    The Linux culture treats it like it's a hobby OS, not being used for serious work. But at the same time, they promote it as being a robust, solid piece of software that should be running tons of servers. You can't have it both ways!

    Now, about this just a hobby OS thing: I'm sure you'll agree that the ability to put backspaces in a filename isn't nearly as much of a problem in practice as bugs that cause crashes and security holes (step one: don't put backspaces in file names; step two: don't deal with those who do)? I'd grant you that argument if you were claiming that there were root exploits that the community refused to fix.



  • @El_Heffe said:

    Ultimately, the problem isn't really "culture".  The problem is that long, long ago somebody made some really, really stupid design decisions.
     

    @boomzilla said:

    I imagine that originally, the thought was that it wasn't worth the resources (back when things were really tight) to enforce stuff like that.

    It's one of those things where knowledge is not finite - decisions made with today's knowledge may prove to open huge cans of worms in tomorrow's environment when more knowledge becomes available, and hindsight is always 20/20.  Unfortunately, not changing or amending that decision early on means the results of that (arguably daft) decision becomes standard behaviour, and correction attempts will break that which is broken yet working.  The "empty string is NULL" Oracle bug/behaviour is a good example.

    @blakeyrat said:

    The problem is, the Linux fans are going around saying, "Linux is better-designed than Windows and you should use it for all your servers and all your desktops and basically it's great, it's so superior to the competition."

    I think you'll find that's the behaviour of Linux bigots. Yeah, they're fans, but their blinkered behaviour is no different to Microsoft bigots with their "but can it play games?" sneers, and ageing mainframe programmers ranting about the stability and power of their steam-powered cast-iron beast compared to these piddly little personal computers that need regular restarts when the processing gets too much for their frail design.  All are as bad as each other - unfortunately the vitriolic bellowing from this vocal minority draw attention away from those that accept their lot and just simply get on with it.

    Drop them all in a pit, along with a couple of chair legs. It'll be fun.



  • The kioslaves thing makes sense to me now. You guys could have just said it's an "opt-in" thing so only programs that understand what's happening see .zips as folders instead of files.

    @boomzilla said:

    @PedanticCurmudgeon said:
    Let me get this straight. You're saying that the only reason they haven't prohibited non-printable characters in file names is because they're afraid of breaking something? What exactly do they think would break?

    I believe that there are systems that use the filesystem as a database, and regularly make use of non-printable characters. You obviously don't read Raymond Chen's blog, because you have no imagination about the stupid things that get perpetuated in the name of backward compatibility.

    This is mystifying to me, because Linux hates backwards compatibility in all other contexts. (Not as much as Apple does, but still.) Seriously, you don't have a stable driver ABI, but this is what you choose to focus your backwards compatibility efforts on? That makes no sense to me.

    @PedanticCurmudgeon said:

    Now, about this just a hobby OS thing: I'm sure you'll agree that the ability to put backspaces in a filename isn't nearly as much of a problem in practice as bugs that cause crashes and security holes (step one: don't put backspaces in file names; step two: don't deal with those who do)? I'd grant you that argument if you were claiming that there were root exploits that the community refused to fix.

    Again: the article says that this filename issue has major security implications, and major potential to create heisenbugs. Those are problems in practice.

    If every Linux user on Earth says, "ok I don't use non-printable characters, or encodings other than UTF-8, or start filenames with a dash", then why the fuck doesn't the OS just enforce those rules in the first fucking place?!

    Moreover, why wasn't this one of the fixes introduced when Ubuntu started (for example) to make Linux "for human beings"? They were (supposedly) starting from scratch; why didn't they fix this? What "human being" wants to see a file consisting of a carriage return and three bell characters? I mean the whole point of founding Ubuntu in the first place was to ditch the old anti-user culture and create a new, friendly, culture... right?

    It boggles my mind. But then again, I'm not a Linux user, and I obviously don't understand the Linux culture.


  • ♿ (Parody)

    @blakeyrat said:

    This is mystifying to me, because Linux hates backwards compatibility in all other contexts. (Not as much as Apple does, but still.) Seriously, you don't have a stable driver ABI, but this is what you choose to focus your backwards compatibility efforts on? That makes no sense to me.

    I would have said that backwards compatibility just isn't nearly as high a priority as with, say, Windows. Of course, if you cooperate with the system, and the source for your drivers are in the kernel source tree, the kernel guys generally do whatever is required to update your code to keep it in sync with their changes. I'm not involved or pay attention to how well that works in practice, but it seems pretty decent. As with most systems, things are difficult if you work against, rather than with the system.

    @blakeyrat said:

    If every Linux user on Earth says, "ok I don't use non-printable characters, or encodings other than UTF-8, or start filenames with a dash", then why the fuck doesn't the OS just enforce those rules in the first fucking place?!

    That's a good question. I would guess that it's just not enough of an actual problem to rise high enough in the priority queue to actually fix.

    Of course, then you get into the issue of who's responsibility it is. The operating system's kernel is different from a file system. I wouldn't expect to see the Linux kernel changed in response to weirdness in NTFS, for example. And there's no monolithic file system that everyone uses. The way forward would probably be to create a new filesystem that doesn't allow these characters. I'm not enough of a filesystem dork to know if this has been done or even contemplated seriously.



  • @blakeyrat said:

    And if Linux fans were going around saying, "oh it's just a hobby OS, it's not professionally engineered and there's really nobody interested in fixing its major flaws", that would be fine. The problem is, the Linux fans are going around saying, "Linux is better-designed than Windows and you should use it for all your servers and all your desktops and basically it's great, it's so superior to the competition." That's where the problem comes from.

    Well, I don't really know any Linux fans that I've ever heard say either of those things (at least not any that I take seriously), but I'm sure you wouldn't let your personal biases get in the way of you speaking for them.

    But seriously, it's a fantastic false dilemma you seem to be implying: Linux can only either be a hobby OS [i]or[/i] the greatest-fucking-thing-evar-omg!!1!  Choose!

  • BINNED

    @blakeyrat said:

    Moreover, why wasn't this one of the fixes introduced when Ubuntu started (for example) to make Linux "for human beings"? They were (supposedly) starting from scratch; why didn't they fix this? What "human being" wants to see a file consisting of a carriage return and three bell characters? I mean the whole point of founding Ubuntu in the first place was to ditch the old anti-user culture and create a new, friendly, culture... right?

    It boggles my mind. But then again, I'm not a Linux user, and I obviously don't understand the Linux culture.

    Actually, I agree with you in the case of Ubuntu given the target market. They could have just patched the kernel and refused to support any apps that didn't comply. By the way, they didn't start from scratch; they started from Debian, which until recently had the reputation of being one of the most difficult distros to install (ironic, isn't it?).


  • @PedanticCurmudgeon said:

    ... which until recently had the reputation of being one of the most difficult distros to install (ironic, isn't it?).
     

    It is? I understood that Debian has a reputation for being one of the more stable (and until yum came along, had much better package management systems than RPM-based distro) but I don't recall finding it particularly difficult to install. It seemed to be as idiot-proof as installing Fedora, CentOS, SuSE and Solaris.

    I know Canonical wanted a stable and robust distro that they could prettify for desktop use, so settled on Debian and added in the end-user bells and whistles (Mint takes that a stage further, I believe).

    I will say that the way in which Debian has "tainted" the way configuration files threw me. It took me a bit to fathom out Apache (compared to RPM-based distros) and the framework in which Debian/Ubuntu arranges Pure-FTPD is completely arse-backwards. If ever I wanted to kick someone in the junk via TCP, it's that author.


  • BINNED

    @Cassidy said:

    @PedanticCurmudgeon said:

    ... which until recently had the reputation of being one of the most difficult distros to install (ironic, isn't it?).
     

    It is? I understood that Debian has a reputation for being one of the more stable (and until yum came along, had much better package management systems than RPM-based distro) but I don't recall finding it particularly difficult to install. It seemed to be as idiot-proof as installing Fedora, CentOS, SuSE and Solaris.

    Debian isn't difficult to install any more, but had that reputation for a long time, and certainly did when the Ubuntu project started.



  • oh, okay - bit before my time, then. I was largely brought up on DeadRat and its RPM cousins before branching into DPKG territory.

     



  • @Master Chief said:

    @ASheridan said:

    You're right of course, a zip is a file and not a directory, so why doesn't Windows Explorer suggest that? I mean, it even gives it an icon that looks like a directory, and opens it up as if it were a directory, and as Windows hides file extensions by default, a lot of people get confused, and don't understand that there is a difference.

     

    It keeps that distinction by virtue of only Explorer being able to do it.  And I'm sorry but if you're working or even using a computer in 2012, you know what a zip is.  My 86 year old grandmother knows what a zip is.

    Plus, HTML aside, if you tried to run an exe from inside a zip, explorer does pop up a dialog letting you know that the files are archived, and you should probably extract them before running.  And even extracts them to a temp directory so you don't mess up your desktop, and cleans them up afterward.  

    I'm all for user friendliness, but at some point the users do need to use their brains.  Even the best software in the world can't do what you guys are asking it to do.

     

    Well you, I, and your 86 year old grandmother know what a .ZIP file is because we've been using them for years.  The problems arise, as ASheridan points out, because newer versions of Windows suggest that .ZIPs are folders by...

    • describing them as "Compressed (zipped) Folders" when you create them in Windows Explorer;
    • giving them a folder-like icon; and
    • allowing users to browse them like folders in Windows Explorer

    Given these three things it should not be unexpected for users to treat them like folders, and it is this behaviour that can cause problems.

    E.g. If I perform the following actions what would I would expect to happen?

    1. Launch Windows Explorer.  Explorer launches.
    2. Double click the "Documents" folder.  The "Documents" folder opens and the contents are listed.
    3. Double click the "Work" folder.  The "Work" folder opens and the contents are listed.
    4. Double click the "Scripts" compressed folder.  The "Scripts" compressed folder opens and the contents are listed.
    5. Double click "Test1.docx".    The "Test1.docx" opens in Microsoft Word.
    6. Make changes to the document then click "File" > "Save".


     



  • @blakeyrat said:

    What "human being" wants to see a file consisting of a carriage return and three bell characters?
     

    :D ... MUST... CREATE ... LOLPIC

     

    Edit

    Actually, just use UTF-8: ↲♪♪♪



  • @blakeyrat said:

    I'm used to Classic MacOS where you could use any printable character except colon.

    Apologies for the necro, but Apple weren't as restrictive as you imagined. I remember Norton Disk Doctor complaining about finding files on my Mac with NULLs in the name. The cause was a piece of software written in C with what I presume was a broken C string to Pascal string converter, that would leave several stray null terminators at the end of filenames. The only reason Norton even cared, was that these files were not openable in Mac OS X. There was no concern otherwise, as Mac OS 9 didn’t care: the files opened fine. The lack of a command line rendered this entire issue irrelevant – filenames were always passed using fully compliant APIs.

    I also deliberately put non-printing characters at the start of names in order to affect sort order in the Apple menu – I even wrote a program that would append one to the start of the name of any program dragged onto its window.

    As I understand it, the colon, being the path separator, was the only restricted character. The only real problem I had was with the odd program writing long names out via Carbon APIs (since HFS+ supported names far longer than any of the classic APIs could deal with), and the Finder was using the standard (volume ID, directory ID, name) tuple to access files, rather than (volume ID, file ID), so the names it was getting were truncated, leading to files that couldn't be opened, renamed or deleted.



  • @dhromed said:

    @blakeyrat said:

    What "human being" wants to see a file consisting of a carriage return and three bell characters?
     

    :D ... MUST... CREATE ... LOLPIC

    Apple’s bitmapped Geneva font had a different animal as its final character for each size. Note that although the screenshot depicts Disk Doctor, that’s not the same Macintosh on which I was getting Mac OS X filename compatibilty warnings. Norton were not prescient.

    Note: image does not constitute a LOLPIC.



  • @Daniel Beardsmore said:

    @blakeyrat said:
    I'm used to Classic MacOS where you could use any printable character except colon.

    Apologies for the necro, but Apple weren't as restrictive as you imagined.

    I read your entire rambling reply, and you never actually got to the point where you explained why Mac Classic filenames "weren't as restrictive as I imagined."

    I'd call you a pedantic dickweed but... I'm not even sure what the hell you were going for there. It's like you had this anecdote that you thought was just so amazingly great that you couldn't resist posting. "But wait," you said, "this anecdote is completely unrelated to anything else in the thread!" You puzzled over this, before a light bulb went off above your head, "of course! I'll just write an introductory sentence that makes it look as if I'm replying to Blakeyrat! Nobody will notice."



  • My bad. I forgot that NULL was printable on Macs. It's been too long since I last used Macs properly.



  • @Daniel Beardsmore said:

    My bad. I forgot that NULL was printable on Macs. It's been too long since I last used Macs properly.

    What the... what the holy fuck are you talking about? Do you have that alien parasite from Star Trek II in your brain?

    I give up on this thread.



  • @blakeyrat said:

    What the... what the holy fuck are you talking about? Do you have that alien parasite from Star Trek II in your brain?

    naaa just plain old lice. The rock lover says that "find -exec" and "find -print0 | xargs -0" cope really good with whatever weird filenames you can come up.



  • @blakeyrat said:

    @Daniel Beardsmore said:

    My bad. I forgot that NULL was printable on Macs. It's been too long since I last used Macs properly.

    What the... what the holy fuck are you talking about?

    More necroing, but in the spirit of helpfulness, here's a summary of the conversation just for blakeyrat:

    blakeyrat: ... any printable character except colon...
    Daniel Beardsmore: It was even less restrictive than that: I had a case involving NULLs in the filename.
    blakeyrat: wait, why is that less restrictive?
    Daniel Beardsmore: NULLs are not printable characters, but it was happy to use them.

    Unfortunately blakeyrat's sarcasm detector failed to trigger on the "I forgot that NULL was printable on Macs" line, causing confusion; I have therefore translated this line in the summary above. Other posters, please note this and do not attempt to use sarcasm in replying to blakeyrat until he can get his detector fixed.



  • One sentence explanation of why Daniel Beardsmore came across as a crazy-person and Scarlet Manuka is an idiot:

    Saying Mac Classic file name can have any printable character except colon doesn't imply that it can't have unprintable characters.

    Beardsmore was acting like because Mac Classic could store NULL in a filename my statement was wrong. Which is clearly Wrath Of Khan brain-slug crazy, and I replied as such. Why am I explaining myself to idiots who can't read? I really don't know.



  • If I said "You can have a beer or a coke", or "You can have any soft drink", asking for coffee wouldn't make sense. It's part of English communication that offering a set of choices suggests that anything not offered is consequently unavailable. It's not a rule; the swiftness and ease with which we can communicate, at least in English, comes at the price of ambiguity and resultant confusion. Precise communication is tedious. English sees abuse at so many levels; even expressions such as to "get dressed" don't make any inherent sense: you're not acquiring anything. Any reasonably intelligent English speaker will be well aware of how wretched a language it is.

    Tearing people new assholes never helps; all we can do is accept that English is painfully ambiguous, laugh, and smile at our follies. No matter how hard we try, we still end up writing sentences which seem clear and precise, yet still have multiple meanings depending on interpretation.



  • @blakeyrat said:

    One sentence explanation of why Daniel Beardsmore came across as a crazy-person and Scarlet Manuka is an idiot:

    Saying Mac Classic file name can have any printable character except colon doesn't imply that it can't have unprintable characters.

    I'd say it's pretty obvious that the statement does imply that based on how people interpreted it (or I suppose if you want to get pedantic they infered meaning from your statement that you did not mean to include).



  • @Daniel Beardsmore said:

    No matter how hard we try, we still end up writing sentences which seem clear and precise, yet still have multiple meanings depending on interpretation.

    Yeah, if you're an idiot who can't read. For the rest of us, the sentence was pretty fucking unambiguous. The sentence says nothing about unprintable characters. If I meant to make a fucking statement about fucking unprintable fucking characters, the fucking sentence would fucking have the fucking word "unprintable" in it somewhere. It doesn't.

    Look assholes, don't dredge up a fucking old thread just to start up the same fucking shit over again. I'm sick of the shit, ok? I'm sick of the shit in the threads going on today, I'm definitely sick of the shit going on in threads from months ago. What the fuck is wrong with you, there wasn't enough shit in your life? You need to dredge up more shit to argue about on the Internet? Fuck off and die.



  • @blakeyrat said:

    @Daniel Beardsmore said:
    No matter how hard we try, we still end up writing sentences which seem clear and precise, yet still have multiple meanings depending on interpretation.
    Yeah, if you're an idiot who can't read. For the rest of us, the sentence was pretty fucking unambiguous. The sentence says nothing about unprintable characters. If I meant to make a fucking statement about fucking unprintable fucking characters, the fucking sentence would fucking have the fucking word "unprintable" in it somewhere. It doesn't.

    Look assholes, don't dredge up a fucking old thread just to start up the same fucking shit over again. I'm sick of the shit, ok? I'm sick of the shit in the threads going on today, I'm definitely sick of the shit going on in threads from months ago. What the fuck is wrong with you, there wasn't enough shit in your life? You need to dredge up more shit to argue about on the Internet? Fuck off and die.

    Sorry blakey, he's right.  He even apologized for the confusion.  You, of course, went on your usual tirade.


  • @Sutherlands said:

    Sorry blakey, he's right. He even apologized for the confusion. You, of course, went on your usual tirade.

    Of course. That's what everybody wanted, isn't it? What other fucking reason was there to dredge this dead thread back up?

    LOLZ GUYZ!!! Blakeyrat is SO WRONG!!! Let's make fun of the Blakeyrat!! He's like a circus clown here for our amusement!



  • @blakeyrat said:

    He's like a circus clown here for our amusement!

    Wait, you're not?



  • Lame thread locked.


Log in to reply