RE: To the Debian Project, IMHO [long]
You Wrote: If you present someone with a challenge they will usually rise to it (as long as it is within their sphere of competence). So let's make life easier, not more difficult. Reply: This is an interesting way to lead into the following ... You wrote: Let's tell the truth Linux is like Unix. You need to know. We can help you, but your a big boy/girl now, and you have learn how to take care of yourself. Microsoft didn't and takes a bashing from the general users who complain that it doesn't live up to expectation, and it doesn't, you still have to know, despite what Microsoft or others may say. Reply: I agree Microsoft still requires knowledge. But the knowledge is somewhat transferable between programs. I also agree we need configuration GUIs in Linux. If you still enjoy the endless research to execute a simple command, by all means don't use the GUI. There are many times when I need a administrative task done more than I need to be doing it! I say endless for one reason. If it is so non-intuitive you need to go through 5 pages of help, you'll probably forget it in short order. Next time you'll be looking it up again. Don't get me wrong, people still need to learn some Unix. Point is, if you want the marketplace, you better go after the market people. The market people want windows offers. This is not by mistake. Microsoft has invested major $$$ in searching for what users want. Namely: 1. Simple peer to peer networking with resource sharing. This means you don't get in line for a cap and gown when you finally get it working. 2. Simple install. Most Linux variants are here. I think RedHat is the easiest to install. Debian is the easiest to maintain due to apt-get. My opinion. 3. Must integrate seamlessly into existing standard. Like it or not, the desktop standard in the majority market holder. At this point, that is clearly Microsoft. Linux needs to be a drop in replacement. 4. Simple maintenance. Linux is not rocket science. It just takes more ongoing administrative effort than windows NT Server. The major difference I see is the GUI. I don't run X, maybe that would change my opinion. I have run it before, but not on debian. Any comments here welcome. I might be missing a big part of debians administrative aids. Time and time again the best argument for Linux is stability. I have run NT workstation for over a year now. That's 24x7. I just don't turn it off. It has been rock solid. Under development conditions, I reboot maybe once per month. 90% of those reboots are because I have changed/installed something. I am not knocking Linux. I am just interested in the comments regarding level of development. Fact is, replacing Microsoft on the desktop requires one of two things. First you have to offer something they don't/can't/won't provide and you absolutely must have. Second, Linux offers enough that the difference between the two is negligible. It seems you don't like the progression of Linux from a standpoint it's going to put it into everyone's hands. Within the lazy man's reach if you will. If I were in marketing, I would yell YOU BET. To continue to grow, it must start capturing people who are less interested in how it works and more interested in what is can do for them. I.e.. a marketing person. Remember, they have computers to. Maybe the question is, should Linux be pushed beyond the server market into the desktop. That is the major hurdle. A system administrator may be able to justify the extra time. I know a died in the wool Unix administrator who chose NT 4.0 server because he had other tasks (programming) and needed low administration. Just my rambling opinions. Thanks for reading. paul -Original Message- From: Simon Martin [mailto:[EMAIL PROTECTED] Sent: Monday, September 20, 1999 7:15 AM To: Stephan Hachinger Cc: Debian User Subject: Re: To the Debian Project, IMHO Hi Stephan, No offence taken, with your mails, and don't get me wrong I quite like GUI interfaces for some jobs, command line for others. I just pick the tool I prefer for the job at hand. As far as I am concerned this thread is not about the GUI/OOUI/command line debate, but about the marketing focus, which can do a LOT of harm. If I go to a hardware store and by a drill to put up some shelves at home I have to know how to use a drill, what kind of wall I am going to be drilling, the type of rawl plugs I will use, etc. It is actually quite a complicated business, and the thing is most of us accept this as normal. Why make using a computer different. To be alble to write with pen and paper I had to learn about ink, hand position, types of paper (don't use a biro on bond because it smears easily, etc.), all things that have nothing to do with the actual activity of writing a letter, just the mechanics of writing a letter. Today I rarely use pen and ink, I use computer and printer, but does that mean that I don't have to learn the mechanics of my writing
Re: To the Debian Project, IMHO [long]
Those who want to build Debian Linux systems for simple users may do so. The tools to do so exist. Some systems could be built for such users, some to accomodate NT admins. and yet others for and by UNIX admins. BTW, I recently worked a contract for a corporation that uses nothing but NT servers and workstations. The machines were rebooted every two or three days, and complete images were installed to them once a week or more. Granted, though, the employees there were actually allowed to send and receive e-mail to their workstations via the Internet with no UNIX server to protect them! Art
Re: To the Debian Project, IMHO [not so long, but OT]
Art Lemasters wrote: [cut] BTW, I recently worked a contract for a corporation that uses nothing but NT servers and workstations. The machines were rebooted every two or three days, and complete images were installed to them once a week or more. Granted, though, the employees there were actually allowed to send and receive e-mail to their workstations via the Internet with no UNIX server to protect them! Semi-serious question: How does a UNIX server protect them against viruses (I assume that's what you mean)? Do they die in the arid environment of the server? ;-)
Re: To the Debian Project, IMHO [long]
Paul == Paul McHale [EMAIL PROTECTED] writes: Paul I also agree we need configuration GUIs in Linux. If you still Paul enjoy the endless research to execute a simple command, by all Paul means don't use the GUI. There are many times when I need a Paul administrative task done more than I need to be doing it! I say Paul endless for one reason. If it is so non-intuitive you need to Paul go through 5 pages of help, you'll probably forget it in short Paul order. Next time you'll be looking it up again. The word intuitive applied to computers is an Orwellism. You are NOT born with innate knowledge of menus and mouseclicks, this way or that. What you really mean when you say intuitive is familiar. You have used some kind of interface before and naturally find it easier to learn other interfaces if they are similar to the one you know. Now there may be many people who know the same interface you do. They may even be the majority of computer users. So what? That still doesn't make the interface superior. Paul Point is, if you want the marketplace, you better go after the Paul market people. The market people want windows offers. This is Paul not by mistake. Microsoft has invested major $$$ in searching Paul for what users want. Wrong, or at least just a small part of the truth. Microsoft has invested even more major $$$ in persuading people that they wanted what Microsoft had to offer. THAT is what 90% of marketing is concerned with today, and that is why I want no part of it. And the people we're talking about are not the end users. The end users didn't invent the (deplorable) I just want my job done and want to know nothing about my tools. It is the MANAGERS that like people working like robots. Paul Time and time again the best argument for Linux is stability. MY major argument for Linux (Unix in general) and against NT is CONTROL vs. the lack of it. If you don't know what I mean I'll explain privately. Paul To continue to grow, it must start capturing people who are less Paul interested in how it works and more interested in what is can do Paul for them. What IS this general obsession with growth? How dirty, smoggy, stinking must our cities get before we all understand that there are limits to technology? Why does everybody equate bigger == better? -- Ian Zimmerman Lightbinders, Inc. 2325 3rd Street #324, San Francisco, California 94107
Re: To the Debian Project, IMHO
Hi Stephan, No offence taken, with your mails, and don't get me wrong I quite like GUI interfaces for some jobs, command line for others. I just pick the tool I prefer for the job at hand. As far as I am concerned this thread is not about the GUI/OOUI/command line debate, but about the marketing focus, which can do a LOT of harm. If I go to a hardware store and by a drill to put up some shelves at home I have to know how to use a drill, what kind of wall I am going to be drilling, the type of rawl plugs I will use, etc. It is actually quite a complicated business, and the thing is most of us accept this as normal. Why make using a computer different. To be alble to write with pen and paper I had to learn about ink, hand position, types of paper (don't use a biro on bond because it smears easily, etc.), all things that have nothing to do with the actual activity of writing a letter, just the mechanics of writing a letter. Today I rarely use pen and ink, I use computer and printer, but does that mean that I don't have to learn the mechanics of my writing materials? Am I born with a mouse attached to my left hand? Is there an instinctive power-on reflex? Do I come into this world knowing that opening seventeen instances of a spreadsheet program on a 486 with 16MB RAM is not a good thing to do? All this is basic computer use skills. I HAVE to learn. If somebody tells me here is your computer it just works what do I expect? If someone says here is your computer, it's fairly simple to use, but you'll have to learn a few things to keep it running smoothly the effect is subtly different. Unix has been famous for cryptic commands and general user unfriendliness it most circles. This is not necessarily good. When the usage message for a command is in excess of 5 pages (tar --help) it makes you wonder. I like an easy life, the same as the rest of us. I also like my challenges, it adds spice to life. If you present someone with a challenge they will usually rise to it (as long as it is within their sphere of competence). So let's make life easier, not more difficult. Let's tell the truth Linux is like Unix. You need to know. We can help you, but your a big boy/girl now, and you have learn how to take care of yourself. Microsoft didn't and takes a bashing from the general users who complain that it doesn't live up to expectation, and it doesn't, you still have to know, despite what Microsoft or others may say. __ _ Debian GNU User / /(_)_ __ _ ___ __ Simon Martin / / | | '_ \| | | \ \/ / Project Manager / /__| | | | | |_| | Isys \/_|_| |_|\__,_/_/\_\ mailto: [EMAIL PROTECTED] 'I used to be schizophrenic, but now both of us are all right' - Original Message - From: Stephan Hachinger [EMAIL PROTECTED] To: Simon Martin [EMAIL PROTECTED] Cc: Debian User debian-user@lists.debian.org Sent: 18 September 1999 21:05 Subject: Re: To the Debian Project, IMHO Hello! Again some criticism about your opinions... Ok, ok, I sit corrected in several respects, but I am still adamant that any attempt to paint Linux as an out of the box solution with no prior knowledge is a real danger to the on-going comercial success of Linux. I worked in tech-support for Xerox for about 7 years (Xerox used to sell Apple Mac, IBM PS/2 and Dell in Latin America), and I would say that at least 70% of the problems we had we with users who not only did not know what they were doing (no problems with that) but who did not WANT to know what they were doing. Microsoft has fixed the image of it's OS as just use and ignore it. Let's not fall into that trap. Fixing the customers expectations is paramount for a successful install. If you fix the expectations as zero cost, zero learning then you are NOT going to have a successful install. I am fairly competent with Debian, but the last time I looked at RedHat, I did not want to do any real config changes until I had read the corresponding man pages and other documents, and these are both Linux based In short. If you use a tool you have to know it. If you want to use a tool well you have to learn how it works. You don't get something for nothing and you definitely don't want to tell your customers to expect the world for nothing. OK, I know what you mean. I've been using Debian for a short time and Windows for a quite long time now, and I worked together with many people on PC projects. I also think in a way that there are two extremes concerning PC users: There are some who want to dive deep into the secrets of the system, and others only want to use it for doing their work. And I (I belong to the first group) realized, whatever I worked on, that the second group of users also got very good results out of their work. Let's look on what the computer was invented for: Scientific calculations. But people had to have very good knowledge and time to use it. And what
Re: To the Debian Project, IMHO
Simon Martin wrote: I am still adamant that any attempt to paint Linux as an out of the box solution with no prior knowledge is a real danger to the on-going comercial success of Linux. I worked in tech-support for Xerox for about 7 years (Xerox used to sell Apple Mac, IBM PS/2 and Dell in Latin America), and I would say that at least 70% of the problems we had we with users who not only did not know what they were doing (no problems with that) but who did not WANT to know what they were doing. Microsoft has fixed the image of it's OS as just use and ignore it. Let's not fall into that trap. I agree. I also worked on Tech support for a large company using Windows and found the same thing. IMHO the real issue is that there needs to be *two* kinds of Operating Systems, as we have now - the Windows kind where the user can install and use it without too much knowledge, with the slick GUI's, and the GNU/Linux kind as found in Debian distribution which can be configured, and most importantly taken apart to see how it works, and then put back together again, so giving people the chance to learn and understand about Operating Systems. IMHO there are two important issues. First, what we must not do is allow GNU/Linux to just imitate Windows - we must have and maintain a GNU/Linux distribution which is for techies and people who want to become techies. Let Corel and others have a distribution which they can make easier, but let us This leads to the second important issue - free software. Debian is important because it is committed to the principles of free software, *and* because it gives us the chance to help develop it, to participate. I may be wrong but there seems to me to be a sense of community among Debian users/developers which you just don't get with Windows and the stuff produced by big corporations. As Corel and other corporate concerns develop GNU/Linux I am sure they will go the same way - it will be them, the corporation, and us, the users, plus restrictive license agreements which take away freedoms. To sum up - what IMHO is important is not the commercial success of GNU/Linux, but keeping GNU/Linux tweakable, learnable, interesting and composed of free software with all the advantages that brings in terms of community and development. Godric
Re: To the Debian Project, IMHO
On Wed, 15 Sep 1999, Simon Martin wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Ok, so this is my 2 cents worth. GUI utilities are ok, command line utilities are ok, simple dialogs are ok, this is not the issue. The real issue is not to fall into the Microsoft or RedHat paradigm. Please, do not compare the two. I run Red Hat, Debian, SuSE and FreeBSD, and I truly believe the Red Hat guys do not deserve these criticism. If you run Red Hat you can use both the GUI interface or the command line utilities, just like in any other flavor of Unix (and I am including Solaris here too). This is a flavour of Unix, Unix is not trivial, Unix is a fairly mature fully featured operating system and you have to know at least a little about what you are doing and be prepared to read the manual before you do something useful. That's fine. However, many other users may just want to run a stable, efficient OS that they have not had a chance to run so far. I do not see why we should limit ourselves to the techies or geeks of the world, especially since many corporations as well as small and midsize businesses could also benefit from running Linux as long as we also make it easy enough for them to get the basic configuration up and running. IMHO one of the reasons that there is a contest between NT and Linux is that Microsoft said that NT was so simple to install and use, unfortunately tuning and other administrative tasks can be a real pain. Linux never made any bones about the fact that you have to learn to be able to use it. It's not out of the box and run. I think that the efforts to dumb down the operating system and say that anyone can use it would hurt the Linux image, maybe irreparably. I disagree here. The main reason why there is a contest between WinNT and Linux is not due to customer disappointment at how difficult WinNT is to set up or run. Rather, the main cause is the lack of stability and scalability of that other OS. Add to that the strings attached to a commercial OS manufactured by a semi-monopolistic company, and you have all the ingredients you need for people to dislike WinNT. Apart from that all real computer enthusiasts are masochists any way. Tell someone that this is a system that only real men can use (sorry about the sexist remark but real people does not convey the same meaning) and you'll have them fighting to get at it (how do you think I started) and have a real sense of achievement when they manage to get the system to boot, and now PPP, and now bind, and now sendmail, and now ... Again I disagree here. I do not think we are able to decide who is and who is not a real computer enthusiast. This is starting to sound like those inquisitors back in the 1950's trying to decide who was a good American. In any case, I also doubt that the best route for Linux is to convert it into an exclusive property of computer hobbysts. No matter how much wemay dislike it, that is not our decision to make. In our societies, it is the market the one that usually makes those decisions. Let Microsoft take the bashing from users who do not WANT to know. Stay just that little bit above the rest. Have all the tools in any and every form. Drag'n Drop is good, but so is rcp. And now it's back to work. __ _ Debian GNU User / /(_)_ __ _ ___ __ Simon Martin / / | | '_ \| | | \ \/ / Project Manager / /__| | | | | |_| | Isys \/_|_| |_|\__,_/_/\_\ mailto: [EMAIL PROTECTED] -BEGIN PGP SIGNATURE- Version: PGPfreeware 6.5.1 for non-commercial use http://www.pgp.com iQA/AwUBN99PNSTe2wIotMrcEQLUpQCg0g4Y22gkoXHf51aeQt3upQ8qK6UAoOTS qWoAPRGuyBGgw6H5LnWQK1BQ =0P97 -END PGP SIGNATURE- -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: To the Debian Project, IMHO
Ok, ok, I sit corrected in several respects, but I am still adamant that any attempt to paint Linux as an out of the box solution with no prior knowledge is a real danger to the on-going comercial success of Linux. I worked in tech-support for Xerox for about 7 years (Xerox used to sell Apple Mac, IBM PS/2 and Dell in Latin America), and I would say that at least 70% of the problems we had we with users who not only did not know what they were doing (no problems with that) but who did not WANT to know what they were doing. Microsoft has fixed the image of it's OS as just use and ignore it. Let's not fall into that trap. Fixing the customers expectations is paramount for a successful install. If you fix the expectations as zero cost, zero learning then you are NOT going to have a successful install. I am fairly competent with Debian, but the last time I looked at RedHat, I did not want to do any real config changes until I had read the corresponding man pages and other documents, and these are both Linux based In short. If you use a tool you have to know it. If you want to use a tool well you have to learn how it works. You don't get something for nothing and you definitely don't want to tell your customers to expect the world for nothing. __ _ Debian GNU User / /(_)_ __ _ ___ __ Simon Martin / / | | '_ \| | | \ \/ / Project Manager / /__| | | | | |_| | Isys \/_|_| |_|\__,_/_/\_\ mailto: [EMAIL PROTECTED] 'I used to be schizophrenic, but now both of us are all right'
Re: To the Debian Project, IMHO
Hello! Again some criticism about your opinions... Ok, ok, I sit corrected in several respects, but I am still adamant that any attempt to paint Linux as an out of the box solution with no prior knowledge is a real danger to the on-going comercial success of Linux. I worked in tech-support for Xerox for about 7 years (Xerox used to sell Apple Mac, IBM PS/2 and Dell in Latin America), and I would say that at least 70% of the problems we had we with users who not only did not know what they were doing (no problems with that) but who did not WANT to know what they were doing. Microsoft has fixed the image of it's OS as just use and ignore it. Let's not fall into that trap. Fixing the customers expectations is paramount for a successful install. If you fix the expectations as zero cost, zero learning then you are NOT going to have a successful install. I am fairly competent with Debian, but the last time I looked at RedHat, I did not want to do any real config changes until I had read the corresponding man pages and other documents, and these are both Linux based In short. If you use a tool you have to know it. If you want to use a tool well you have to learn how it works. You don't get something for nothing and you definitely don't want to tell your customers to expect the world for nothing. OK, I know what you mean. I've been using Debian for a short time and Windows for a quite long time now, and I worked together with many people on PC projects. I also think in a way that there are two extremes concerning PC users: There are some who want to dive deep into the secrets of the system, and others only want to use it for doing their work. And I (I belong to the first group) realized, whatever I worked on, that the second group of users also got very good results out of their work. Let's look on what the computer was invented for: Scientific calculations. But people had to have very good knowledge and time to use it. And what the computer can do now (at least sometimes): Help at work and calculate things without needing much time for administration any more. Although I like to administrate a system properly, IMHO it is very admirable that people can now work with computers almost with zero administration. And it makes life easier and makes the computer work for what it was created: Solving problems. So, IMHO, it is not good to think that people that can only work with GUI interfaces and dumb OSes are stupid or worth less. I think it's very understandable if someone wants to concentrate on his problems and not on his systems. Now, what do this thoughts end up to? I think there should be both kinds of OSes: The just use and... ones and the ones which require proper administration and can be tuned properly. Why should we let MS control the market of the just use and ignore it OSes? I there should be a Linux for EVERY person, and distributions like the coming Corel-Linux, which are almost completely GUI-based, will improve the success of Linux. But Linuxes also have to be tweakable. I think that's why there are different distros: That different users can reach different aims. Admins who want to tune it thorougly as well as home users who want to use it without reading any manual, just to show to extremes. That's what Linux should be, I think. And if Debian should become easier to install or something like that, there should be discussions about how the future Debian can solve problems in a better way than last versions. But there should not be a flame war about UNIX philosophy because it's no use starting such a war. Time and users will tell where to go finally and which philosophies are right. Why can't there be just peaceful coexistence between different users and philosophies??? And: About the last paragraph: I think good software can be used without knowing any manual. It has some help functions that quickly guide to the required functions/params. You can see what command you must call. That's good software. I want to dive into internals of OSes. But, personally, I don't read manuals very often. And I succeeded with WIN and with Debian this way. So that philosophy can't be that bad. That was my opinion about this. I already have used WIN and LINUX, and DOS, and the GEM UI, and DR DOS, and CALDERA DOS, and so on. Intolerance brings no solution. Did not want to offend anyone:). Kind Regards, Stephan Hachinger
Re: To the Debian Project, IMHO
In article [EMAIL PROTECTED], Simon Martin [EMAIL PROTECTED] writes: SM dialogs are ok, this is not the issue. The real issue is not to SM fall into the Microsoft or RedHat paradigm. This is a flavour of Hmmm... don't think so. SM Unix, Unix is not trivial, Unix is a fairly mature fully featured SM operating system and you have to know at least a little about what SM you are doing and be prepared to read the manual before you do SM something useful. OK, if i want to design a big network, i have to know many things about networks. But that has nothing to do with this or that os. It should be plain and simple to setup an os. An os is a tool - it should help me in getting my job done. Some tasks are complex, ok, then you have to learn something. But the UI of the os and the tools to fullfill the task should be as simple to use as possible (and still be flexible). SM IMHO one of the reasons that there is a contest between NT and SM Linux is that Microsoft said that NT was so simple to install and SM use, unfortunately tuning and other administrative tasks can be a The main problem with NT is not problems in tuning it or some other nice to have features. The main problem is, that NT is not scalable and most important it is not reliable and stable. If NT would be as stable as Solaris or even Linux, than NT would be a killer system, because MS knows that even experts like it, if complex tasks are simple to solve. I work with NT once in a while and at home i work with Linux. Linux is very flexible and i like this. But there are many tasks which are much simpler to do in NT. For example the registry of NT is no problem - but missing tools to edit the registry even from a boot disk is a problem. SM real pain. Linux never made any bones about the fact that you have SM to learn to be able to use it. It's not out of the box and run. I Here everything goes wrong! As i said above: If i plan to make a (big) network i have to learn about network-basics and network-theory. It's a problem when MS claims with NT you have not do know anything about network-basics. But it is very good, if you can do your tasks with easy to use tools. Let's take some network-settings as example. If you want to use DHCP with NT or Win, it's quite easy - go to network settings an say to get IP address automatically. If you want to use DHCP with Linux you habe much more to do. Why? Is there anything good if activating DHCP is not as simple as for NT? One pain with NT is, that often there is only one way to do this or that and this way gives you not all possible options that should be available. Here Linux is better. But a tool have not to be a pain to use in order to be flexible. -- Until the next mail..., Stefan.
Re: To the Debian Project, IMHO
In article [EMAIL PROTECTED], Jason Wright [EMAIL PROTECTED] writes: JW Also, I spend a lot of time in my job working on servers on the JW other side of the world, hidden behind slow, overloaded WAN links. JW Sure I *can* (and sometimes do) use ssh to run GUI apps on these I know, what you mean. But if a program is good designed, then it's easy to give it a textmode and a graphic mode UI and last but not least even a command line interface. Even a command line driven interface can be easy to use and a nice and pretty GUI can be hard to use. JW Or if your box is hosed and X won't come up and you can't get it JW out of single user because, say, /usr is trashed so you HAVE to JW fix it from the command line. If you've learned on a GUI and JW never learned the formats of the underlying config files, you're JW hosed and quite possibly out of a job. Where is the problem? A GUI is nothing bad - it's very good! If i'm good in my job, i'm happy if i can easy do my job with a nice and easy to use GUI. And if there are problems i'm able to edit config files by hand with a text editor. If you use a GUI you don't become dump! JW Granted, none of these things are common for the casual user, but JW such situations are very common for me. I find that command line It's very good, if this is possible to do. And here Unix is much better than NT. But it's not very good if all you have are simple and not very easy to use tools. Best is, if both are available. -- Until the next mail..., Stefan.
Re: To the Debian Project, IMHO
On Tue, Sep 14, 1999 at 09:56:24AM -0700, Ian Zimmerman wrote: Info keybindings are virtually identical to Emacs, and thus don't feel at all stupid to this Emacs user. In fact, I would be VERY annoyed if But consider what it used to be like - the curses info browser was very confusing because the motion keys weren't bound quite in the same way as those in Emacs (or anything else for that matter). the info maintainers heeded Havoc's advice and redid the keybindings in the image of, say, vi. I'd be surprised if any effort to improve ease of use chose to standardise on vi (unless nethack has a sudden burst of popularity, anyway). Keith Dselect? Nice in many ways; yes, it's the non-standard Keith keystrokes and lack of visual feedback (has my search finished Keith yet?) that are probably the worst things. Other than that *I* Keith can't think of a much nicer way to address package installation Keith in ncurses. That's right. And I have to disagree that the general flow of control in dselect doesn't make sense and this has _zero_ to do with the power of the tools. Change the flow of control and I bet you'll lose some flexibility somewhere. Better prompting and visual feedback would probably go a long way to making the flow of control seem better - one common source of problem is your idea of what's going on diverging from that of dselect. It would also be nice to be able to defer conflict resolution sometimes, not to mention the *!?# recommends handling and the forced display of the help screen. -- Mark Brown mailto:[EMAIL PROTECTED] (Trying to avoid grumpiness) http://www.tardis.ed.ac.uk/~broonie/ EUFShttp://www.eusa.ed.ac.uk/societies/filmsoc/ pgp6UPI8jQA8E.pgp Description: PGP signature
Re: To the Debian Project, IMHO
On Tue, Sep 14, 1999 at 10:53:52AM -0400, Raymond A. Ingles wrote: 1) Allow quick install from floppies It'd be really nice to have a system for having a single package installed on a floppy, or set of floppies in a way where a user can stick the disk in, and dselect would read and install whatever What advantages would this have over dpkg -i other than checking dependancies before installation? 2) Simple install When trying to install debian, I am simply overwhelmed by the number of packages I have to deal with, and I consider myself to have above average knowledge of UNIX programs! When first installing Debian, a simple walkthrough of Q/A would solve most of these problems. This sounds like what the profiles/task lists are trying to do. 3) Insist on better package descriptions When digging through packages, I'd frequently come across a package I don't recognize. I can't even tell if it would be something I'd want, because there's no description! You can do this yourself already by report a bug against the packages or mailing the maintainers. For whatever reason the description probably makes perfect sense to them and unless someone says something they won't realise that people are having trouble with it. Generally, if an individual package doesn't do something you'd expect or could be improved the best thing to do is talk to the maintainer - most will welcome useful suggestions, and it's much more likely to achieve something than pointing out general trends. -- Mark Brown mailto:[EMAIL PROTECTED] (Trying to avoid grumpiness) http://www.tardis.ed.ac.uk/~broonie/ EUFShttp://www.eusa.ed.ac.uk/societies/filmsoc/ pgpHEgxYAfC4F.pgp Description: PGP signature
Re: To the Debian Project, IMHO
I think most people miss some important points: - A text or graphic mode UI is some times the most effective user interface, some times a command line driven interface is more effective. - Even the best of the experts is very happy if a good tool is easy to use, so it costs less time to do a job. - Ease of use is not only for lusers or beginners! I'm a system administrator and programmer and i like programs which are easy to use, cause i like to have my job done instead of learning how to use a tool. - What advantage has a command line tool with 3000 options, extremly flexible, but no one is ever able to remember all these options? The user interface and the ease of use is one of the most important parts of every program - only this way a tool can really help in doing a job more effective and less time consuming. Take samba as an example? If all you do every day is working with smb.conf then you know one day nealy all the options. But if not? If you have to tweak smb.conf only once in a while? Then every time you have to work through the manpages - but hey, there is swat. That little small tool really helps you to just do your job. -- Until the next mail..., Stefan.
Re: To the Debian Project, IMHO
On Tue, Sep 14, 1999 at 10:19:37AM -0500, Keith G. Murphy wrote: It would also be sad to leave behind the Unix heritage of simple utilities that can pipe to each other to do useful things. This is more than a heritage, this is a cornerstone of a useful computing environment. I thought VMS was annoying because I had to explicitly read and write temp files to 'pipe' programs together. Then I found out how good I had it when I was condemned to Windows NT. No doubt that a feature-rich GUI has its place. But don't forget the (hopefully) many of us that are irritated when we experience: $ type sed sed is /bin/sed instead of $ type sed /bin/sed - Programs don't create data. People create data. Every program is a filter. -- Mark Gancarz, 'The Unix Philosophy'
Re: To the Debian Project, IMHO
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Ok, so this is my 2 cents worth. GUI utilities are ok, command line utilities are ok, simple dialogs are ok, this is not the issue. The real issue is not to fall into the Microsoft or RedHat paradigm. This is a flavour of Unix, Unix is not trivial, Unix is a fairly mature fully featured operating system and you have to know at least a little about what you are doing and be prepared to read the manual before you do something useful. IMHO one of the reasons that there is a contest between NT and Linux is that Microsoft said that NT was so simple to install and use, unfortunately tuning and other administrative tasks can be a real pain. Linux never made any bones about the fact that you have to learn to be able to use it. It's not out of the box and run. I think that the efforts to dumb down the operating system and say that anyone can use it would hurt the Linux image, maybe irreparably. Apart from that all real computer enthusiasts are masochists any way. Tell someone that this is a system that only real men can use (sorry about the sexist remark but real people does not convey the same meaning) and you'll have them fighting to get at it (how do you think I started) and have a real sense of achievement when they manage to get the system to boot, and now PPP, and now bind, and now sendmail, and now ... Let Microsoft take the bashing from users who do not WANT to know. Stay just that little bit above the rest. Have all the tools in any and every form. Drag'n Drop is good, but so is rcp. And now it's back to work. __ _ Debian GNU User / /(_)_ __ _ ___ __ Simon Martin / / | | '_ \| | | \ \/ / Project Manager / /__| | | | | |_| | Isys \/_|_| |_|\__,_/_/\_\ mailto: [EMAIL PROTECTED] -BEGIN PGP SIGNATURE- Version: PGPfreeware 6.5.1 for non-commercial use http://www.pgp.com iQA/AwUBN99PNSTe2wIotMrcEQLUpQCg0g4Y22gkoXHf51aeQt3upQ8qK6UAoOTS qWoAPRGuyBGgw6H5LnWQK1BQ =0P97 -END PGP SIGNATURE-
Re: To the Debian Project, IMHO
Mark Brown wrote: On Tue, Sep 14, 1999 at 09:56:24AM -0700, Ian Zimmerman wrote: Info keybindings are virtually identical to Emacs, and thus don't feel at all stupid to this Emacs user. In fact, I would be VERY annoyed if But consider what it used to be like - the curses info browser was very confusing because the motion keys weren't bound quite in the same way as those in Emacs (or anything else for that matter). On the other hand, I wouldn't be too tough to make the key bindings depend on the EDITOR environment setting (I always wondered why I found the dselect keys so difficult ;-) THAT would be the right way to do it INNVHO, but I'm not going to volunteer because it works great whatever the keybindings are, and it's just not that hard to deal with, anyway. the info maintainers heeded Havoc's advice and redid the keybindings in the image of, say, vi. I'd be surprised if any effort to improve ease of use chose to standardise on vi (unless nethack has a sudden burst of popularity, anyway). Keith Dselect? Nice in many ways; yes, it's the non-standard Keith keystrokes and lack of visual feedback (has my search finished Keith yet?) that are probably the worst things. Other than that *I* Keith can't think of a much nicer way to address package installation Keith in ncurses. That's right. And I have to disagree that the general flow of control in dselect doesn't make sense and this has _zero_ to do with the power of the tools. Change the flow of control and I bet you'll lose some flexibility somewhere. Better prompting and visual feedback would probably go a long way to making the flow of control seem better - one common source of problem is your idea of what's going on diverging from that of dselect. It would also be nice to be able to defer conflict resolution sometimes, not to mention the *!?# recommends handling and the forced display of the help screen. -- Mark Brown mailto:[EMAIL PROTECTED] (Trying to avoid grumpiness) http://www.tardis.ed.ac.uk/~broonie/ EUFShttp://www.eusa.ed.ac.uk/societies/filmsoc/ Part 1.2Type: application/pgp-signature -- - Ralph Winslow [EMAIL PROTECTED] The IQ of the group is that of the member whose IQ is lowest divided by the number of members.
To the Debian Project, IMHO
I have been a Debian user for several years now. The first time I loaded it was before the Bo version. I compared it to what was available at the time and chose Debian mainly because of the installation control it offered via dpkg and dselect. But I also preferred just about everything else specific to Debian. I felt Debian provided me with a distribution oriented towards the power user and have stuck with it since. My system works exactly the way I want it to and only with Debian could I have tailored it so thoroughly. It pleases me greatly every time I work with it. Although dselect is becoming unruly with the huge number of packages that have become available in recent versions, I still prefer it over other package tools because of the two things that are becoming more scarce every day with modern software, that is control and flexibility. There seems to be a wave of opinion, in the software industry, that the value of software should be primarily gauged by its level of grandmatization (the unrealistic practice of engineering software so that even grandma can install and use it). The important measures of its value such as content, stability, capability, and flexibility are being placed secondary. This counter productive philosophy, started by the biggest and best example of what software should not be and you know who I'm talking about, is propagating at the expense of the truly important measures. Extrapolated to the end result, it's computer fascism. I will have no choices when I install software and will not be able to change anything after I do. Some operating systems and applications are almost to that point already. After reading the last few months worth of newsletters at debian.org, I am concerned that the Debian project may be buying into this foolishness. My hope is that Debian sticks to their guns as a distribution for power users and does not jump on the grandmatization bandwagon. To the Debian organization; please don't worry about the unimportant aspects and concentra te on the important ones. I want control, flexibility, stability, and content. I DO NOT care how difficult or time consuming (barring problematic) an installation process is. If my refusal to compromise any of these important aspects means I have to spend more time answering questions and entering configuration choices during an installation process or even editing configuration files by hand, then so be it!
Re: To the Debian Project, IMHO
On Mon, Sep 13, 1999 at 04:53:04PM -0700, Craig H. Block wrote: After reading the last few months worth of newsletters at debian.org, I am concerned that the Debian project may be buying into this foolishness. My I think that the debian project will have the best of both worlds. There will always be the community driven distribution that we all use. The corporates (i.e., Corel) will also have dumbed down packages for the newbie types. Conceptually it works in a traditional manner. First, there is the console app with all its command-line flags and options. Then there is the graphical front-end (the X version). This version is more or less about as hard as the command-line version depending on experience, but equally as powerful and scriptable. Then there is the dumbed-down KDE or GNOME versions that are aimed at newbies. These versions may have all the options of the earlier versions but the basic functions that a newbie would want are right there on the toolbar. I don't see this as a problem. Use of the KDE/GNOME version does not negate use of the console version or the X versions. What I would like to see for all apps is the following. One console version package. One X package. One Gnome and one KDE package. Then, there could be the really big 'all-in-one' package. This would come in two versions. One ends in KDE; one ends in GNOME. This is for people who like to use different things at different times, and it would install the console, X and (DESKTOP ENVIRONMENT) versions simultaneously. This is sort of how the gmc package works. It installs the console Midnight Commander and the GNOME/X Midnight Commander at the same time. This gives you the flexibility of the console MC, and if you want to toss icons around you have that option (sometimes it's just easier). One the one hand, you can say 'damn that's just way too many packages to support.' But wait, we already support multiple versions in this exact manner. This will be an effort in consolidation and collaboration. It will also keep the newbies talking to the vets (i.e., the student/guru relationship) which is good for everyone. hope is that Debian sticks to their guns as a distribution for power users and does not jump on the grandmatization bandwagon. To the Debian organization; please don't worry about the unimportant aspects and concentra te on the important ones. I want control, flexibility, stability, and content. I DO NOT care how difficult or time consuming (barring problematic) an installation process is. If my refusal to compromise any of these important aspects means I have to spend more time answering questions and entering configuration choices during an installation process or even editing configuration files by hand, then so be it! This addresses another issue, configurability. Just because a package is installed in a default state (for the newbie) does not prevent you from getting into the docs/HOWTOs and learning how to edit the config files by hand. I would prefer a default installation for newbies, because in general the person who wrote the default config knows what he/she is doing and can give a newbie something secure to work with. After that, I don't have to make innumerable changes to the lines in the files. Perhaps I would have to change just a few. That's ok by me. 'Configurability' and 'security' issues brings up a recent thread on /. about whether we should distribute linux in server and workstation versions. Rather, the server version would just like the workstation version except with nothing omitted. Another way to say this is that the workstation version is the newbie version with some of the more difficult and security risky server packages omitted. (just my $0.02) -- NatePuri (natedawg) Certified Law Student McGeorge School of Law Sacramento, CA [EMAIL PROTECTED] [EMAIL PROTECTED] http://www.ompages.com http://office.ompages.com/~natedawg PGP: http://www.ompages.com/PGP.html UIN: 43504034 IRC: office.ompages.com #ompages pgp9qEAug0xql.pgp Description: PGP signature
Re: To the Debian Project, IMHO
Hi, As someone who's written lots of code toward making free Unices easier to use, let me just say that I think this supposed conflict between power and ease of use is total nonsense. Even if a tool is for power users, it can be pleasant for those users to learn and use or it can have an atrocious interface from hell. Atrocious interface from hell does not equate to powerful, just annoying. For example, look at info or dselect - regardless of whether you personally like them, the many people that _don't_ like them don't like them because the keystrokes are really stupid and the general flow through the programs doesn't make sense. However, this has _zero_ to do with the power of the tools. On a programming level, command-line tools are often written in such a way that adding a GUI frontend is difficult and requires changes to the command-line stuff. But this is not a fundamental conflict, just poor planning when writing the command line tool. Co-existence is very possible, and taking a position on one side or the other is just pointless. Havoc
Re: To the Debian Project, IMHO
On Mon, Sep 13, 1999 at 16:53:04 -0700, Craig H. Block wrote: This counter productive philosophy, started by the biggest and best example of what software should not be and you know who I'm talking about, is propagating at the expense of the truly important measures. Extrapolated to the end result, it's computer fascism. See http://www.tuxedo.org/~esr/jargon/html/entry/Godwin's-Law.html . After reading the last few months worth of newsletters at debian.org, I am concerned that the Debian project may be buying into this foolishness. The project now has a couple of people in it who know to do PR. That doesn't in any way mean the project is sacrificing its values to marketing. The Debian project is still what the developers make it. If you disagree with a direction the project is taking, become a developer and do something about it (see http://www.debian.org/devel/help). HTH, Ray -- Tevens ben ik van mening dat Nederland overdekt dient te worden.
Re: To the Debian Project, IMHO
Well... my personal oppinion surely agrees with Havoc's, and I'm just addind my $2c... I sure don't want to see ONLY X-friendly configuration programs, specially because after you get some insight on the matters this proves to be a slow, sometimes unreliable and, sometimes, even impossible way to configure what you want (just remember the problem with Win95's hardware configuration... if your hardware is misconfigured and you can't run the GUI, then you can't hace access to the tools that let you correct the mistake - so you're stuck!)... On the other hand, the GUI-versions of those tools are of great value to the user that has little time to read the whole documentation about some feature (say, SAMBA) and want to do his learning 'as time allows it', going from a basic setup to an advanced and optimized one setp over step, in a reliable way... For this purpose, you can have either a well-commented configuration file (and, in the case of SAMBA, I've never seen one that's 100% complete - RedHat's one is good, Debian's one is better, but each one has it own tweaks, and I may want to try a mix of both) OR a well-documented GUI-tool, like gtksamba, which organizes the parameters in groups and calls the man file for each parameter when asked... This is a great way to do your learning and to walk over the gap that separates the regular user from the power user... This was the road I have chosen years ago in respect to Windows (using the excellent documentation that the Norton Utilities used to provide for its tools), and that's the road I've chosen to learn more about Linux. In the perfect world, the GUI tools and the console tools will both be available, and completely independent - i.e., you can choose to install and/or uninstall each of them whenever you want to w/out messing with the other. ...just my $2c... Guilherme Zahn (still a 'pseudo' power-user)
Re: To the Debian Project, IMHO
I'd like to repeat a suggestion made to this list over two years ago. I loved it when I saw it, and I think it bears repeating. I saved it because it gave what I thought was a wonderful example of user-interface design. -- To: debian-user@lists.debian.org Subject: DEITY TEAM: Ideas (Long) Date: Wed, 16 Apr 1997 14:52:53 -0600 From: Mike Patterson [EMAIL PROTECTED] Ok guys, ready for my comments on dselect? There are some doozies in here, and some not-so-doozies. Also, there may be many points that can't work simply because of the way packages are set up. The ideas are in no particular order. 1) Allow quick install from floppies It'd be really nice to have a system for having a single package installed on a floppy, or set of floppies in a way where a user can stick the disk in, and dselect would read and install whatever is on it. The main reason for this is for the user's ease. If the package being installed has dependancies, they should be listed with some sort of You need XXX to install this, where can I find it? message. 2) Simple install When trying to install debian, I am simply overwhelmed by the number of packages I have to deal with, and I consider myself to have above average knowledge of UNIX programs! When first installing Debian, a simple walkthrough of Q/A would solve most of these problems. EXAMPLE: (inaccurate, but you get the point) Are you going to access the WWW from this system?[Yes] What is your prefered Reader?[NETSCAPE] Are you going to read email on this system? [Yes] What is your prefered Reader?[ELM] Do you want X installed on this system? [Yes] Will you be writing programs for X? [Yes] If all questions are displayed at once, some questions can grey out, or options become unavailable as each question is answered. 3) Insist on better package descriptions When digging through packages, I'd frequently come across a package I don't recognize. I can't even tell if it would be something I'd want, because there's no description! 4) Graphical Install It is possible to make a GUI work in text, but often you'll end up in a situation where too much information is displayed in a manner that the user just can't comprehend because of all the text on the screen. I also understand that X may not be an option. Mid-way, you can use text with IBM-ASCII symbols to make nice borders around everything to seperate them out evenly. 5) Better keystroke selections I dont' know how many times I got upset when I selected a package with return only to exit the selection process. d to read more of the description? + -? Here's an idea of what I'm thinking of: (done in non-ibm ascii, because I'm not on an IBM.. and done without any planning, so I can do better :) __ ___ _ / Location \/ Selection \/ Install \ ---/ \ || | - Base X11 *Communication* math development games design - | || | PACKAGE Description Status| | - | | Elm A mail reader Installed | |-Netscape A WWW Browser SELECTED | | tin A News reader | | ppp A communication protocol Unconfigured | | slip A communication protocol FAILED| | dip Manager for SLIP DE-SELECTED | || || | Press space on any item to select or de-select it. Press Return on any | | Item to get more detailed information on it. | -- In this particular design, you use tab to rotate through the tabs, left and right arrows to select the groups of packages you are dealing with, up and down to scroll through the possible selections, and space to select. Return gives the description of the package, as well as a reason for its status. In the case of slip above, it would search the package and say something like: file /usr/bin/slip is missing.
Re: To the Debian Project, IMHO
Havoc Pennington wrote: Hi, As someone who's written lots of code toward making free Unices easier to use, let me just say that I think this supposed conflict between power and ease of use is total nonsense. Even if a tool is for power users, it can be pleasant for those users to learn and use or it can have an atrocious interface from hell. Atrocious interface from hell does not equate to powerful, just annoying. For example, look at info or dselect - regardless of whether you personally like them, the many people that _don't_ like them don't like them because the keystrokes are really stupid and the general flow through the programs doesn't make sense. However, this has _zero_ to do with the power of the tools. On a programming level, command-line tools are often written in such a way that adding a GUI frontend is difficult and requires changes to the command-line stuff. But this is not a fundamental conflict, just poor planning when writing the command line tool. Co-existence is very possible, and taking a position on one side or the other is just pointless. I absolutely agree, but... Developer time is limited, especially free time. If there's a choice, any administration tool needs to be command-line oriented for maximum flexibility. And please don't get into the RedHat situation where it's hard to change a configuration file without messing up the GUI administration tool. It would also be sad to leave behind the Unix heritage of simple utilities that can pipe to each other to do useful things. Dselect? Nice in many ways; yes, it's the non-standard keystrokes and lack of visual feedback (has my search finished yet?) that are probably the worst things. Other than that *I* can't think of a much nicer way to address package installation in ncurses.
Re: To the Debian Project, IMHO
Keith == Keith G Murphy [EMAIL PROTECTED] writes: Keith Havoc Pennington wrote: Havoc For example, look at info or dselect - regardless of Havoc whether you personally like them, the many people that _don't_ Havoc like them don't like them because the keystrokes are really Havoc stupid and the general flow through the programs doesn't make Havoc sense. However, this has _zero_ to do with the power of the Havoc tools. Info keybindings are virtually identical to Emacs, and thus don't feel at all stupid to this Emacs user. In fact, I would be VERY annoyed if the info maintainers heeded Havoc's advice and redid the keybindings in the image of, say, vi. Keith Dselect? Nice in many ways; yes, it's the non-standard Keith keystrokes and lack of visual feedback (has my search finished Keith yet?) that are probably the worst things. Other than that *I* Keith can't think of a much nicer way to address package installation Keith in ncurses. That's right. And I have to disagree that the general flow of control in dselect doesn't make sense and this has _zero_ to do with the power of the tools. Change the flow of control and I bet you'll lose some flexibility somewhere. dselect is the No. 1 reason I prefer Debian to Red Hat, period. -- Ian Zimmerman Lightbinders, Inc. 2325 3rd Street #324, San Francisco, California 94107
Re: To the Debian Project, IMHO
snip When trying to install debian, I am simply overwhelmed by the number of packages I have to deal with... snip I have a solution for that problem using a setup script as follows. I simply copy this script and a selections file into ~/deb21 after bypassing dselect during initial install. The selections file (dpkg-selections) can be generated from an existing install with dpkg --get-selections. A person could design a custom interactive interface to generate the selections file and avoid dselect's selector altogether. This script is a good example of dselect's and dpkg's power and flexibility. #! /bin/bash grep -q /dev/hdc /cdrom /etc/mtab if [ $? -ne 0 ]; then mount /dev/hdc /cdrom if [ $? -ne 0 ]; then exit; fi fi grep 2/4 /cdrom/.disk/info if [ $? -ne 0 ]; then echo Wrong CD; umount /dev/hdc; exit; fi umount /dev/hdc if [ ! -f /var/lib/dpkg/methods/multicd/shvar.multi_cd ]; then dselect access fi if [ ! -f /var/lib/dpkg/methods/multicd/available ]; then dselect update; fi dpkg --set-selections ~/deb21/dpkg-selections if [ $? -ne 0 ]; then echo Could not set selections; exit; fi dselect install dselect config dselect remove