Slashdot Mirror


User: Tom+Christiansen

Tom+Christiansen's activity in the archive.

Stories
0
Comments
621
First seen
Last seen
Profile
(view on slashdot.org)

Comments · 621

  1. Who shall I say is calling? on Guide to Slashdot · · Score: 2
    What's wrong with "by Microsoft or whomever"? It's "by whom", like "to whom", you know. Saying "by whoever" is wrong.
    Er, not exactly. Let's just say that your understanding represents an isolated view of reality.

    The following examples are all correctly inflected. Once you understand why, it all becomes easy.

    1. It's really a question of whom you're trying to make happy.
    2. It's really a question of who is happy.
    3. It's really a question of whom you've seen.
    4. It's really a question of who wants to go.
    5. It's really a question of who is going.
    6. It's really a question of who you think is going.
    7. It's really a question of whom you told to go.
    8. It's really a question of who you think I said was going.
    9. It's really a question of whom you think I told to go.
    English is a context-sensitive language. You can't just look at these things in isolation.
  2. Re:Easter Eggs? on Guide to Slashdot · · Score: 2

    Perl comes from $^X. And he misspelled "whoever" with a spurious "m". Diff the output, and you'll see that I correct that.

  3. Re:Easter Eggs? on Guide to Slashdot · · Score: 2
    Nope, I'm too stupid to see it. Pointers?
    Why certainly:
    open(STDIN, "-|")&&exec${chr 3*2**3*2>>1}=>-(-(-pe))=>$_
    for(qw)y/a-z/n-za-m/ /oevorq/&&y/z//d))[root,~0&1,tchrist]; exec
    reverse lc
    reverse(substr("LMTH.TODHSALS/MOC.SREVRESEERF.MM9E USSI//:PTTH",1)),
    UVFYE
    ^ "x2345" x2345 ** -(-(-x2345)), PERL
    ^ "<<<4"
    Enjoy. :-)
  4. Easter Eggs? on Guide to Slashdot · · Score: 2
    Well, maybe not an Easter egg, per se, but there's an amusingly ironic example of the spelling/grammar glitchery that the article talks about hidden right there in the article itself. Anybody else but me notice it?

    I imagine that this was probably done intentionally as a crafty joke, of course. This begs the quesiton of whether there aren't other subtle, self-referential digs lurking in there for the discerning eye to pull out?

  5. Re:What about Dworvak? on Interface Zen · · Score: 2

    First of all, the Happy Hacker keyboard is not small. In fact, the keys are farther apart than on many keyboards. As for those fufi keys than you long for, they're completely useless. And yes, you can get at them if you chord. But they're really silly.

  6. Re: troff was *sigh* on Daemonnews reviews Applixware · · Score: 2
    is there a decent introduction/tutorial for troff
    Here's a set of the original documents. Or just get the BSD 4.4 manuals from ORA/Usenix.
  7. Linuxes are threatening MS's existing customers on Daemonnews reviews Applixware · · Score: 2
    Why would M$ fear the Linuxes less than it fears the BSDs? That doesn't seem to make sense to me. I also doubt that the dissing comes from M$ flagwavers sewing the seeds of sedition and dissent in the Unix ranks. I'd attribute it to internecine rivalry derived from ignorance and malice.

    It would seem that the Linuxes are much more commercially oriented. Look at Corel, Caldera, RedHat, SuSE, and all the rest of them. Sure, BSDs are more used by ISP businesses and other high-tech places filled with Unix professionals, but that's really a completely new place for M$, once where they've not traditionally been very effective (maybe there's not enough room temperature IQs there for them to hoodwink so easily :-). But because the Linuxes are obviously trying to attack M$'s existing business at least in mindshare if not in real dollar amounts (but I bet there's something there, too), I would think that the Linuxes would be much more threatening to M$ than the BSDs.

  8. Re:Touch typing on Interface Zen · · Score: 2
    Funny how this sounds like an argument of manual transmission versus automatic transmission. Can you imagine the outcry if all cars would have only automatic transmissions? Or how about not allowing anything but 1st and 2nd gears simply because untrained individuals screw up a lot if the car can ever exceed 2nd gear?

    By restricting the interface to something that can be used with zero training and a complete inability to remember something from one day--or moment--to the next, you put up needless speed limits that penalize the competent. There is no reason whatsoever to do that save for pure laziness.

    Finally, now that reading and writing are no longer effectively taught at our oxymoronically named grammar schools, it's long past time that simple typing should become a mandatory subject. Part of using a computer is using a keyboard. Part of using a keyboard is typing. Learn to type.

  9. Re:*sigh* on Daemonnews reviews Applixware · · Score: 2
    Ah, I understand. Rich Stevens used troff. Rich Stevens is dead. Therefore, troff is dead.

    I write letters, articles, resumes, and even books in troff. So do a lot of real programmers. Other people use Latex, and they're real programmers, too.

    Since people use this stuff, and get plenty of real work done with them, they're hardly dead, much though you might wish this were the case, you cowardly bastard.

    I think we should change "Anonymous Coward" to "Cowardly Bastard". If you haven't the testicular fortitude to place a name on a post, don't bother posting.

    Just because the entry cost for power-tools like troff and latex is a double-digit IQ doesn't mean power-tools shouldn't exist. Yes, I realizethat half the populace doesn't measure up to that requirement. Oh well.

    And yes, I remember the first winmenumouse systems. I used the Star system on Dandelions from Xerox long ago, back before Apple stole the interface from them, and before the Evil One stole the interface from Apple. It really pissed me off when I had to leave school and go work in industry and couldn't use the nice Xerox tools, but had to do everything in troff again instead. But I still got my job done.

    And it's not just because I am a programmer. Even secretaries can do this if they have to. Remember that for many years, all the secretaries at Bell Labs used vi and troff for all the corporate documents. Don't underestimate a secretary.

    I'm not saying one should do this. I'm merely saying that one can. Stop expecting everyone to be an idiot.

  10. static link your editor for emergencies on Daemonnews reviews Applixware · · Score: 3
    I don't have a "console" running right now, but any old tty will do. Why do people keep focusing on "console"? Where did this word even come from? What ever happened to "terminal", or in geekspeak, "tty"? And anyway, slogin gives you a fine pty. It will even tunnel for a "console"-emulating xterm (that's one that snags syslog messages to /dev/console and kernel printf's).

    But come to think of it, ed doesn't even need a tty. It'll run fine over a socket, too, even if isatty() returns false. It's good for automation, and complete desperation, but not a lot else. If you're going to have a staticly linked editor, which of course is a must and many of these silly commerical Linux-based operating systems forget to do this, then you might as well have something more user-friendly, like, oh, I don't know, maybe ex. :-)

  11. Re:*sigh* on Daemonnews reviews Applixware · · Score: 2
    I would add that MS Office has never spent "almost an hour to render the document and print it" for a ten page (!) document.
    Neither has troff.
  12. Re:Excuse my ignorance, but... on Daemonnews reviews Applixware · · Score: 2

    You have all kinds of confusions in saying that "All the BSDs are binary compatible with Linux." First of all, you pretend there's one binary format. Second of all, you pretend there's one Linux. Both are wrong. I can promise you that your Slackware x86 binary with x86 instructions and slackware admin and path stuff is completely useless to someone using BSD on a Sparc. There is a lot to think about: machine hardware, kernel syscalls, and normal admin bits. Some of those are sometimes compatible, some of them are not. My Sparc can run Solaris binaries under BSD, but only if they don't expect Solaris sysadmin crud. And then you have the whole problem of different Linux-derived operating systems. I have programs that work find under Redhat's OS but which fail miserably using SuSE's OS, even on the same hardware. Compatibility isn't all it's cracked up to be.

  13. Re:A few trivial comments on Interface Zen · · Score: 2
    You think you use the arrows without pause, but you are demonstrably deceiving yourself. The arrows are not underneath your fingers. You have to move to get there. Therefore, there is more of a delay than there would be if you had simply depressed the keys immediately underneath your fingertips.

    QED

  14. Re:A few trivial comments on Interface Zen · · Score: 2
    The way you make all programs recognize both ^H and ^? for some operating is to modify the tty driver. Currently, there's only one byte per action, so if you bind ^H to erase, you can't use ^? for that.

    Of course, it's already bound to your interrupt character, right? :-)

  15. Re:Touch typing on Interface Zen · · Score: 2
    Just because many people who use computers are not programmers, this is absolutely no excuse for creating programmer-hostile software. Which is what we have. Make something even a drooling idiot can use, and only drooling idiots will want to. And programmers are not drooling idiots.

    End of story.

  16. Re:German keyboards are even worse on Interface Zen · · Score: 2

    What do you mean built for programmers rather than typists? Programmers are typists, you know. Well, maybe not Hawking, but you get the picture.

  17. Re:This is subjective on Interface Zen · · Score: 2

    The problem is that these GUI folks have the nutty idea that non-chorded, simple keystrokes are forbidden to them. That's where they're fundamentally fucked in the head.

  18. Re:Way to go, Tom! on Interface Zen · · Score: 2
    The essential difficulty people have in understanding vi is that it is not some modal thing. Rather, it happens to have an insert command, which happens to be terminated by an ESC key (or ^] for bad keyboards like Wintel crap.)

    The other problem is that the Prisoners of Bill and other members of the drooling public are expecting guessware. You know what guessware is--it's when you should just be able to fricking guess what a program does. This, of course, is fundamentally at odds with not merely vi, but emacs and virtually all the rest of Unix as well.

    Can you imagine trying to learn how to play a game merely by guessing? Of course not. How about Java or C++? Please don't laugh. People try this. And they fail, miserably, and then blame the programmer for not understanding how idiotic the user was that he would never even think to bother to read something. Unix is no different, nor any programming language.

  19. Re:What about Dworvak? on Interface Zen · · Score: 2

    Dvorak never became popular because the price of entry is far, far too great with respect to its payback for most people to consider. The price is not the keyboard itself. That's zero-cost: you just xmodmap the thing into its new morph. The real cost is the wetware cost.

  20. Re:Chords on Interface Zen · · Score: 2

    The hard part of knowing what right click means is that for some of us (gente zurda), it means clicking with our left finger. :-)

  21. Re:Some sense in Windows Cut, Copy and Paste. on Interface Zen · · Score: 2

    Unix *has* standardised on get and put. (I don't know what this "copy" thing is.) Use button one to get, button two to put. Button three makes it bigger. And the single/double/triple click gets characters, words, and lines.

  22. Re:Come on on Another Software Spy · · Score: 5
    An http User-Agent is sent because I told my machine to contact that server.
    You actually let that one out? Really? I never do. It's none of their business.
    The User-Agent header allows the server to better taylor content for my machine.
    Not really. If it did, then they wouldn't be playing by the rules. They'd be using embrace-and-extend games to lock you into a non-standard page with non-standard markup for non-standard agents.

    Gosh, I can't imagine who would ever want to do a wicked thing like that. :-)

  23. Re:All the world's a Mac on redhat.com Redone · · Score: 2

    Oh great. More bullshit moderation. What the bloody screaming hell is the matter with you, whoever you are? Haven't you read the moderator guidelines? What is so fucking overrated about that posting? It wasn't flaming. It was curious and interesting and relevant to this community. It deserved a funny, not an overrated. Why should I have to post anonymously to avoid this kind of irresponsible crap? Go read the guidelines.

  24. More dirty laundry: kfind(1)'s kretinism on Interview: KDE Developers Answer Your Questions · · Score: 2
    Well, we're on a roll. Here's dirty laundry report #3. It's for the kfind(1) utility. It claims to be a "front-end to the find utility". It shares many of the problems of the other two programs I reviewed, but has many of its own.
    1. When it first starts up, you have three word based menus. One is "File", but has nothing to do with files. Another is "Edit", but it has nothing to do with editing. Whoever decided that this made any sense was obviously on drugs--and very bad drugs, at that.

    2. There's a little checkbox at the bottom that says it will "include subfolders". Subfolders? SUBFOLDERS?? Here is a folder:
      $ folder +inbox/lists
      inbox/lists+ has 3264 messages (2-4217); cur=4216; (others).
      And here are subfolders:
      $ folder -r -f
      inbox/lists
      inbox/lists/ppt
      inbox/lists/training
      If you mean directories, say the word. This is just something that really rankles this Unix user. It seem baby talk.

    3. Once again we have a row of random happy icons that fail to make this user happy. I'm still waiting for a way to say "just give me real words, damn it, not happicons". Allegedly this will some day be fixed. Right now, it's far too cutesy to live. There are all these silly things that I have no idea what mean. I feel like I've been given some child's toy, but the toy was designed for Japanese children, and I don't read Chinese. Give me words. Words, words, words, words, words, words, words!

    4. The date modified menu is completely nutty. It reads "Find all files created or modified". CREATED? Since when does the inode contain the creation time? Answet: it doesn't. This is simply wrong. You lead people to believe they can ask about creation. They can't.

    5. I wanted to use the mouse to get the string "Find all files created or modified" so I could punch it into this editor session. It wouldn't let me. It was unable to grab the text. That's wrong.

    6. I clicked on a file in the output. Nothing happens. I double click. Still, nothing happens. So I go up to the edit menu to figure out how to edit the file, and of course, there is no editor listed under Edit. How sweet. But there is a Copy command. Now, in my book, there's a big difference between ed(1) and cp(1), but I go with the flow. So I try to do that copy thing, but nothing happens. No new file appears. No error message appears. Nothing changes. It's like my bits went to lala land.

    7. This non-copying copy command is set to ^C. This is a madness that makes makes a Unix user want to roll over and die. I have already specified that ^C is to interrupt. Interrupt. Got that? Actually, it needn't be ^C. But for me, this time, it is. The program that isn't going to respect keyboard signals had bloody well inspect my tty chars and do what they say to do. This would explain why when I accidentally launched a big find on slash that even after about a dozen ^C's the idiot program refused to interrupt the find. Damn it, this is just wrong. Talk about violating the principle of least surprise!

    8. There was no "ok fine, now run your find command" button. What could possibly be more important? It should be prominently placed, with a nice keyboard shortcut like "f" for find, or "r" for run.

    9. The so-called "Advanced" menu is hardly advanced. And it has all kinds of nasty problems of its own. First, it repeats the "folders" heresy. Those are not folder, those are directories. Didn't anyone teach you that it's S_ISDIR not S_ISFOLDER? Sheesh.

    10. When you use the oxymoronically named "Advanced Menu" to select "files", it seems to choose only S_ISREG files. Uh, guys, that's very non-intuitive. If you think that a device or a directory isn't a file, you need to spend more time programming. Go look at the stat(2) system call for an educative experience. Now, if you mean S_ISREG files only, do please say that. Sheesh.

    11. A socket (S_IFSOCK) is not a "special" file. S_IFCHR is a character special, and S_IFBLK is a block special. And just what is this "etc" noise? Don't you et cetera me. I want the specifics. Do you mean anything that isn't S_ISREG or S_ISDIR? Then say that. I don't need fuzztalk.

    12. I have discover that the "at least/most N kilobytes" has a serious and idiotic bug. It thinks a kilobyte is 10**3 bytes. It isn't. It's 2**10 bytes. And where's the megabyte option? Do I really have to say 1024 kilobytes instead of one megabyte? Oh wait. You still have that bug. You mean have have to say 1048.576 kilobytes to mean a megabyte? You've got a crackhead in your house. I suggest you find him.

    13. The "containing text" string is wrong. First of all, it disrespects my editing characters. Again. What is with you folks? I've stty'd myself to have ^U kill the line and ^W kill the word. I have expressed myself quite clearly in establishing those system preferences. Kindly stop being so damned anti-Unix and pay attention to your settings.

    14. The "containing text" is also wrong because it's not text you seem to be asking for. It's a regex. Now, which regex library is it using? What are the rules?

    15. When you use the oxymoronically named "Advanced Menu", the list of "types" you can select on muddles the difference between real file types as defined in the st_mode field of the inode, and random other things. It doesn't tell me how it is going to guess the rest. Is it going to run file(1)? I certainly hope so. Anything else is going to get wrong answers. And the list is very weird. It's not sorted. And it is missing all kinds of things that file knows about. And how to I type in my own specification? For example,
      $ file /tmp/foo
      /tmp/foo: perl commands text
      Now, how do I ask for that kind of file? Huh? Please don't tell me I'm not allowed to ask for things that the author didn't foresee. That's very un-Unixy.

    16. The list of types is too long to force me to use my eyes to find things with. I should be able to type /pat to quickly find something. As it is, I have to play that stupid trick where my fingers need to leave my hands (ie, leave my home row) and my eyes their sockets (ie, leave the screen). I really, really hate that. Just slows me down. And I don't know whether I just missed seeing what I want in that big list.

    17. If this is supposed to be a front-end to find, where are all the options? Where do I specify -xdev or -follow? What about -user and -group and all that?

    18. How do I get regular find -ls output in the return list? That would be very helpful.
    So once again, we have another gratuitously GUIfied recreation of normal tools that manages to do far less yet still have gravely serious problems. I hope you now understand why Unix users are so underwhelmed by your flash and chrome. And, I hope, you now begin to understand how you can fix it.
  25. Re:More Dirty Laundry: ktop(1) report card on Interview: KDE Developers Answer Your Questions · · Score: 2

    Well, I went and found a newer version, and things are... different.

    On the matter of flicker, there's this nifty new technology called "curses". You see, it knows better than to do work it doesn't need to do. You might check into this. It's much more clever than this flickering crud.

    You're "long standing GUI precedent" of putting things in the wrong places is hardly reasonable to a Unix person. We have no such misdirected expectations. A file menu should manipulate files. That's what its name is, and that's what it should do. I couldn't care less if Microsoft likes to put things in stupid places. If 100 million Microsoft victims jumped off a bridge, I still wouldn't follow them.

    As for using the tab key to move to buttons, what are you supposed to do get them to activate, like the show tree thingie? I tried carriage return, enter, asking nicely -- what do you have to do? This is not intuitive, or else it would have worked.

    Don't be grabbing my focus and locking it in. I wanted a menu up so I could type the stuff in that was on it for my original note. You refuse to let me do this. That's cruel and unusual.

    As for what regular top does or does not do, you're right that some of what I asked for was an enhancement request. That's because you're supposed to make something better when you rewrite it. So far, you still haven't managed to catch up to the standard version, which is far easier to navigate and use.

    You ignored my point about the confirmation not being keyboard accessible. At least in this system, I can tab between fields, but I can't hit carriage return to make them select. And this "tab around" thing just isn't all it's cracked up to be. I don't want to have to tab over to the kill process button. I want a keystroke binding for that button. Different thing entirely. And that binding really should be k by default, but I should be able to change it. You keep telling me I can configure anything. Fine. How do I do configure ktop so that when I type 'k' it executes the kill button on the current selection? And how do I fix the borked up signal choice? Where's that configuration? Is there just a file I can edit somewhere, please? It would be so much easier.

    Yes, it is very important not to do use SIGKILL as the default. Your point about "but it says kill, so it should do that". Apparently you don't understand the kill system call very well. It means to send a signal to a process or process group. What that signal does varies tremendously. I as a unix user know what kill(2) does, and it's not what that button says. You should call it annihilate or something. Where's the config file to change that? By making the default the only signal that irrevocably obliterates a process, you set up a really bad precedent. If I see a sysadmin who SIGKILLs nonchalantly, I knock him upside the head. Processes have things to clean up! You give it SIGINT, SIGTERM, or SIGHUP. You have to give it a chance to shut down properly! This is critical in any admin situation. If you don't understand this, you'd better find someone who does.

    Finally, the cursor keys, like pageup and pagedown, are almost as bad as the mouse for acting like speedbumps. They are not close enough to be accessed without visual confirmation, which means you have broken the link of concentration between your brain and the screen. The price of the context switch times to flick your eyes bewteen screen and keyboard, and to move your fingers from the real keyboard to the fufi keys, is clearly non-zero. It's a bad design that disrupts the flow. I cannot find those keys without 1) looking and 2) moving my hands. That just kills it for any touch typist. This is why vi bindings are best, and even emacs bindings are better than the context switches that destroy your train of thought.

    And don't start with your damned ad-hominem crud again like you've already begun. I can just hear you already replying with something about how I must be keyboard-challenged or something. To the contrary: I can play Bach fugues and Beethoven sonatas with my eyes closed even when there are fast, three-octave jumps, so I'm certainly capable of finding things without using my eyes -- on a well-designed keyboard. The issue is that these keyboards you cater to and expect me to use are not designed for that kind of ease of use the way a piano is.

    These extraneous, off-the-keyboard "keys" just end up slowing you down. I use a Happy Hacker keyboard, because that's all you need. The fufi keys are more trouble than they're worth.

    Again and again I tell you that this stuff is not designed for anyone but victims with a pre-existing condition: Microsoft brain-damage. I explain why if you don't try to kiss up to that, you arrive at different conclusions. I explain why making things that intentionally disrespect Unix standards just pisses us off.

    By only making things easy for instant users and those contaminated by MS-brainrot, and by ignoring optimizations that help out long term and professional users, you guarantee that those who use this stuff the most (long term users) will be the ones who most get rubbed the wrong way. This is a strategic error that will come back to haunt you as you keep optimizing for idiots.