So I was lucky then. I could bootstrap building GNU/tar using an existing compiler on our Sun system (gcc 2.81 for $DEITY's sake), and then use that to untar the GCC sources.
Yup. Correct remark. I used OS/2 from 1994 until 1999, when I switched over completely to Linux. I was already using a whole lot of Free Software on OS/2, of which LaTeX was the most important one for me.
And dual-booting for me meant booting between OS/2 and DR-DOS.
I worked for a small bank, and they used a mix of leased Bull mainframes, IBM mainframes and WANG VS minicomputers which were later replaced by HP/UX + Oracle.
If everything must 'Just Work (TM)', then you should not even need to install drivers on your system.
For this to work, one needs to be able to design an OS that can take into account the set of all hardware that exists and will ever exist, and to accomodate for that you should have a (very) limited set of hardware interfaces together with a limited set of protocols for each of these.
Of course, USB comes very close to this ideal. However, protocols depend upon their application area, and for every kind of functionality, you need some additions (eg. wireless LAN vs. Ethernet LAN).
I do think, however, that it is entirely possible to build a Linux desktop which can accomodate many types of users.
I have created a Linux system for my father, consisting of Debian with KDE.
He can mail, surf, uses QCad and OpenOffice, and I am able to do remote troubleshooting and repair.
For scanning we still use his old 100 Mhz Pentium system with Windows 98 and Ultr@VNC. That is because he has an old parallel port scanner, and also a SCSI slide scanner (the SCSI card is ISA). He is very satisfied with this system.
We even added a webcam once, but the GNOME application which connected to it was unstable and tended to crash.
I do not need to reinstall/repair his system every so often, which was the case when he ran Windows. This makes it possible for us to concentrate on finding out and learning functionality.
Every six months or so, I provide him with an update of Debian unstable (after I tested it) so that his system expands in functionality.
I added an account for my brother too, so that he can surf and check his email on Yahoo. He did not have any questions or problems at all.
I am fairly confident that I can apply this to other people also (a colleague of my wife has mentioned his interest).
Things which now have gotten better in Debian unstable vs. Debian 3.1 functionality : mostly that KDE now supports USB Mass Storage devices out of the box with hal and DBUS. That is the most basic reason that I need to upgrade my fathers system, he gets more and more requests to exchange things via memory stick.
Other things that I need to check out : printing from QCad and GIMP I still haven't tackled on his system (it works for me) and activation of sound on his desktop (but I haven't done that for myself either, mostly because I do not really need sound functionality on my systems).
What we need (and what I have planned, now only find time) is a public requirements specification on what the basic functionality of a desktop must be and how this can be filled in on most popular distro's.
Part of this public requirements should also be about adding new hardware, after a desktop has been configured and used. This should lead to public recommendations about what brands of hardware to choose.
Another part should also be a list of available software in stores and what alternatives there exist for them.
This must lead to a public integrator manual which should contain easy steps to configure a chosen distribution to the required functionality.
I stress that one of the main points should be that these documents should be public, easily reachable and easily readable, to make it clear to hardware and software companies which do not make it easy to use their products on Linux are not mentioned and to make it feel to them that they are not mentioned.
This manual should be usable by the less-than-average geek, which can then support his family and acquaintances, and make recommendations on hard- and software to use.
I think that the main barrier to adoption is just the lack of local, simple, easy to remember and easy to find knowledge for the average computer enthusiast who normally supports his neighbourhood.
I have an AMD Athlon P2400 rig with 2 Gb ram, 1 40 Gb drive and 3 120 Gb drives in RAID5.
In between 2004 and 2005 the thing had been continuously running about 250 days, and it added at the end of the year about 400 EUR extra to my electricity bill.
Currently, I now start it and stop it when I want to use it.
I need time to figure out if I can use some advanced power management on that system or if I need to move to a more recent motherboard with power saving features.
I have three collegues here, on CS Ph.d, the two other website developers and programmers, they all three hate Flash because it insists on doing things its own way. None of them really knows Linux.
I have the impression that the Colour of Magic started more as a continuation of Strata, with Strata clearly a parody of Larry Niven's Ringworld (something which you can best appreciate if you read Ringworld first, and then immediately Strata after it).
I think that the Color of Magic was more a way to set a starting point for further ideas and discoveries.
Anyway, I knew Pratchett since Strata in the eighties, but rediscovered him only in the nineties, with Mort the first book I read from the Discworld series.
If I have to compare Pratchett with other humorous writers, he is clearly on top. If you reread a book of him, you will still appreciate his humor, while with Douglas Adams and Tom Sharpe, when you read them once, you laugh out loud, but a second time you ask, what did I laugh with ?
Comparing Pratchett with other fantasy authors is unfair, because all of them mostly rehash the same things, while Pratchett is able to give new life to all those old ideas and create new ones.
About Good Omens, if you at least don't know the storyline of The Omen, then Good Omens might be a little difficult to understand.
Hm, I saved this and sent it to my email address (jurgen.defurne at pandora.be). However, at the moment I am out of internet connectivity because I moved, so it will be probably at least a week before I can look into your questions.
My goal is to build a database with problems and questions, how to solve them and how to add them to Linux distributions. I find that the biggest problem for people to change over to Linux is that they cannot easily find someone with the right level of expertise, while with Windows they do have that.
That is the reason that I do not had much problems switching over my dad and my wife, because I am always at their disposal.
I have put my father on Linux fulltime, and my wife uses it too on my terminal server.
I have written a introductory and an advanced course on Linux, but it would be nice to have some input from someone who normally uses Windows and now tries to switch.
Slashdot Mirror
So I was lucky then. I could bootstrap building GNU/tar using an existing compiler on our Sun system (gcc 2.81 for $DEITY's sake), and then use that to untar the GCC sources.
Yup. Correct remark. I used OS/2 from 1994 until 1999, when I switched over completely to Linux. I was already using a whole lot of Free Software on OS/2, of which LaTeX was the most important one for me.
And dual-booting for me meant booting between OS/2 and DR-DOS.
Weird, you can have at most 100 comments on a page, but I would like a choice that I could have all the comments on one page.
Are you working at Philips ?
Have you ever tried COBOL ?
You should read this.
How small was this bank ?
I worked for a small bank, and they used a mix of leased Bull mainframes, IBM mainframes and WANG VS minicomputers which were later replaced by HP/UX + Oracle.
Scheme and this book.
If everything must 'Just Work (TM)', then you should not even need to install drivers on your system.
For this to work, one needs to be able to design an OS that can take into account the set of all hardware that exists and will ever exist, and to accomodate for that you should have a (very) limited set of hardware interfaces together with a limited set of protocols for each of these.
Of course, USB comes very close to this ideal. However, protocols depend upon their application area, and for every kind of functionality, you need some additions (eg. wireless LAN vs. Ethernet LAN).
I do think, however, that it is entirely possible to build a Linux desktop which can accomodate many types of users.
I have created a Linux system for my father, consisting of Debian with KDE.
He can mail, surf, uses QCad and OpenOffice, and I am able to do remote troubleshooting and repair.
For scanning we still use his old 100 Mhz Pentium system with Windows 98 and Ultr@VNC. That is because he has an old parallel port scanner, and also a SCSI slide scanner (the SCSI card is ISA). He is very satisfied with this system.
We even added a webcam once, but the GNOME application which connected to it was unstable and tended to crash.
I do not need to reinstall/repair his system every so often, which was the case when he ran Windows. This makes it possible for us to concentrate on finding out and learning functionality.
Every six months or so, I provide him with an update of Debian unstable (after I tested it) so that his system expands in functionality.
I added an account for my brother too, so that he can surf and check his email on Yahoo. He did not have any questions or problems at all.
I am fairly confident that I can apply this to other people also (a colleague of my wife has mentioned his interest).
Things which now have gotten better in Debian unstable vs. Debian 3.1 functionality : mostly that KDE now supports USB Mass Storage devices out of the box with hal and DBUS. That is the most basic reason that I need to upgrade my fathers system, he gets more and more requests to exchange things via memory stick.
Other things that I need to check out : printing from QCad and GIMP I still haven't tackled on his system (it works for me) and activation of sound on his desktop (but I haven't done that for myself either, mostly because I do not really need sound functionality on my systems).
What we need (and what I have planned, now only find time) is a public requirements specification on what the basic functionality of a desktop must be and how this can be filled in on most popular distro's.
Part of this public requirements should also be about adding new hardware, after a desktop has been configured and used. This should lead to public recommendations about what brands of hardware to choose.
Another part should also be a list of available software in stores and what alternatives there exist for them.
This must lead to a public integrator manual which should contain easy steps to configure a chosen distribution to the required functionality.
I stress that one of the main points should be that these documents should be public, easily reachable and easily readable, to make it clear to hardware and software companies which do not make it easy to use their products on Linux are not mentioned and to make it feel to them that they are not mentioned.
This manual should be usable by the less-than-average geek, which can then support his family and acquaintances, and make recommendations on hard- and software to use.
I think that the main barrier to adoption is just the lack of local, simple, easy to remember and easy to find knowledge for the average computer enthusiast who normally supports his neighbourhood.
You could always try this.
But I threw away all my 16-bit related motherboards and stuff at the end of last year, when I moved to my new home.
I suppose the referenced project is practically dead, due to such reasons as mine.
Seriously, and very probably.
I have written a small benchmark program which does random things strings of random size.
I implemented this in C and in LISP.
The LISP code compiled by SBCL has the same speed as the C program, and compiled by CMUCL it is even a little bit faster.
At first sight I thought you said 7000 nanometer, but I presume you mean nautical miles.
I have an AMD Athlon P2400 rig with 2 Gb ram, 1 40 Gb drive and 3 120 Gb drives in RAID5.
In between 2004 and 2005 the thing had been continuously running about 250 days, and it added at the end of the year about 400 EUR extra to my electricity bill.
Currently, I now start it and stop it when I want to use it.
I need time to figure out if I can use some advanced power management on that system or if I need to move to a more recent motherboard with power saving features.
I have three collegues here, on CS Ph.d, the two other website developers and programmers, they all three hate Flash because it insists on doing things its own way. None of them really knows Linux.
This strengthens my beliefs that there is no such thing as an invention, there are only designs.
This strengthens my beliefs that there is no such that as an invention, there are only designs.
I have tried this.
I only get views of devices which are defined in /etc/fstab.
No, both on the same system using Debian unstable.
What I really like about GNOME, and what I haven't found yet in KDE, is its volume management.
For KDE to work with USB Mass Storage devices, I need to configure udev via udev.rules.
With GNOME, I just plugin the device and it appears on my desk, with a nice default name.
This makes GNOME a much better sell for ordinary users.
Pity I do not have mod points any more.
You are the first one to clearly state what I have known for some time intuitively, but could not state myself.
I have the impression that the Colour of Magic started more as a continuation of Strata, with Strata clearly a parody of Larry Niven's Ringworld (something which you can best appreciate if you read Ringworld first, and then immediately Strata after it).
I think that the Color of Magic was more a way to set a starting point for further ideas and discoveries.
Anyway, I knew Pratchett since Strata in the eighties, but rediscovered him only in the nineties, with Mort the first book I read from the Discworld series.
If I have to compare Pratchett with other humorous writers, he is clearly on top. If you reread a book of him, you will still appreciate his humor, while with Douglas Adams and Tom Sharpe, when you read them once, you laugh out loud, but a second time you ask, what did I laugh with ?
Comparing Pratchett with other fantasy authors is unfair, because all of them mostly rehash the same things, while Pratchett is able to give new life to all those old ideas and create new ones.
About Good Omens, if you at least don't know the storyline of The Omen, then Good Omens might be a little difficult to understand.
Hm, I saved this and sent it to my email address (jurgen.defurne at pandora.be). However, at the moment I am out of internet connectivity because I moved, so it will be probably at least a week before I can look into your questions.
My goal is to build a database with problems and questions, how to solve them and how to add them to Linux distributions. I find that the biggest problem for people to change over to Linux is that they cannot easily find someone with the right level of expertise, while with Windows they do have that.
That is the reason that I do not had much problems switching over my dad and my wife, because I am always at their disposal.
Just what is it that you need to do ?
I have put my father on Linux fulltime, and my wife uses it too on my terminal server.
I have written a introductory and an advanced course on Linux, but it would be nice to have some input from someone who normally uses Windows and now tries to switch.
When I install GNOME out of the box in Debian, it automatically mounts my USB drives. I still haven't this functionality with KDE.
For looks, I prefer KDE.
climacs and hemlock come to mind, but are less known because they come more from the Common Lisp crowd.