Perhaps they're stuck with some from-a-box commercial Linux distro that is not as great after you've taken it home and used it for a while as it seemed in the store you bought it from? Perhaps the costs could be lowered by choosing a non-commercial Linux distribution; spending some money instead on hiring competent and productive administrators who are capable of joining and fruitfully contributing to the Debian Maintainer community.
With Debian, debconf, and something like FAI, they could perform huge numbers of customized and preconfigured (not merely "kickstarted", but preconfigured) installs all at once. And by using ideas from http://www.infrastructures.org/ and "Apt", it may be possible to maintain them with less work than it would take given any commercial distro. If it was my choice, I'd certainly choose Debian, since the Debian toolset and community support make adminstration tasks so much easier.
Of course it will require well trained full time staff to keep it all functioning. Perhaps that requirement for a Computer Science degree will turn out not to be superfluous after all!
(So did Mr. Stallman ever get to actually FIX that printer?)
Perhaps they're stuck with some from-a-box commercial Linux distro that is not as great after you've taken it home and used it for a while as it seemed in the store you bought it from? Perhaps the costs could be lowered by choosing a non-commercial Linux distribution; spending some money instead on hiring competent administrators who are capable of joining and fruitfully contributing the Debian Maintainer community.
With Debian and something like FAI, they could perform huge numbers of installs all at once. And by using ideas from http://www.infrastructures.org/ and "Apt", it may be possible to maintain them with less work than it would take given any commercial distro.
Of course it will require well trained full time staff to keep it all functioning. Perhaps that requirement for a Computer Science degree will turn out not to be superfluous after all!
(So did Mr. Stallman ever get to actually FIX that printer?)
Scheme is the best first language.
on
C
·
· Score: 1
I agree with "How to Design Programs" (MIT Press, available online with a great Scheme environment usable by beginners) as well as experts) that Scheme is a great first language.
I think that speech data entry is inefficient and not appropriate in most office environments. Think of how noisy it would be if everyone spoke to their computers!
What would be really wonderful is a Gregg Shorthand recognition system, for palmtop, laptop, and desktop digitizer pads. It would be a lot faster than the current text recognition systems, and maybe even faster than a keyboard for prose input. I don't think that Gregg is being taught as much as it used to be, but a freely available Gregg input system would bring it back for sure. There are already several gesture recognition programs out there. Gregg is something like that.
Sure, the current "boot-floppies" installer is not that great. Historically, the Debian installation system has not gotten much attention, since once you install Debian, you just upgrade after that. You never need to use the installer again.
Are you aware of the "debian-installer" effort? Development has been postponed until after the Woody (Debian 3.0) release, since it was too far from completion when "testing" was created, and many of the people working on it are the same people who work on "boot-floppies". They can only do so much in a day and one thing at a time.
"Debian-installer" promises to be quite an improvement over "boot-floppies". It deserves attention from developers and folks wishing to base their product on Debian GNU/Linux. It sports a wonderful modular architecture that should be easy to use as sort of an installer toolkit or framework. It will be easy to customize for totally automatic, or interactive installs, for over the network or CD use.
As for using Debian GNU/Linux on the desktop, I must say that many people out there are doing just that. The "menu" system is wonderful, and we have both Gnome and KDE working quite well. The X Window System is configured upon installation via "debconf", and is at least as good as Red Hat's XConfigurator was last time I ran it.
When you create a package of the software, the
library dependancies are gotten for you by the
packaging tools. It's just a matter of installing
everything it needs and building it. After the
packager does that, everyone else out there who
wants to use the software can just run "apt-get"
and Everything Just Works.
It is better to use libraries than "reinvent the
wheel" and thus have a lot of programs with
unshared mostly identical code in them taking up
RAM. At least on a decent Linux distribution
with good package management and up to date
development libraries available, you can get
a very high level of code sharing going on.
This is in contrast to the Windows environment
where every vendor has their own proprietary DLL
and very little code sharing can happen between
vendor's softwares. That plus there is not any
central package management system in place, afaik.
Whats really a shame is seeing row after row of PCs running scrolling banner screensavers at a college computing lab. Here in Portland, OR, I've seen that at PCC (pcc.edu)... I asked the people in the lab why don't they turn on the DPMS? They guy said he'd ask his supervisor about it, but nothing happened. (Perhaps the supervisor was too busy blowing smoke.)
Those provide a PHP or ASP like templating system with full access to the Apache API and all of full fledged Perl (and thus the CPAN modules). It's way better than PHP!
Anything that can be done in PHP could be done with either of these systems. I think it would be worth doing to port a lot of that fancy PHP HTML magic over to either HTML::Mason or Embperl. (May the better engineered solution win!)
There is, additionally, an emacs extension called "mmm-mode" (found via the HTML::Mason site) that makes psgml-html mode and cperl-mode co-exist! The perl parts are in cperl-mode, and the HTML parts are in HTML mode. It's really cool!
I hope they send one to Richard Stallman and ask him to fix it for them.
Wasn't it a printer with firmware he couldn't have the source to what made RMS start the FSF?
8-)
DEC SRC has a Virtual Paper web site. They've got a thing called "Lecturn" that takes a scan and turns it into a document consisting of a series of compressed bitmaps, and, if your arch has the supported OCR library available, the text annotated with bounding box information so that a search can highlight the right spot on the bitmap. The source is available as a Modula-3 program. Lecturn documents are HUGE. If the similar thing could be done with PDF, and the document files are much smaller (better compression) it would be a good thing to have. I also recall seeing a (binary only) scheme system that was up for free download that had OCR stuff built into it. It was kind of like a `festival' for OCR, I guess.
Slashdot Mirror
Perhaps they're stuck with some from-a-box commercial Linux distro that is not as great after you've taken it home and used it for a while as it seemed in the store you bought it from? Perhaps the costs could be lowered by choosing a non-commercial Linux distribution; spending some money instead on hiring competent and productive administrators who are capable of joining and fruitfully contributing to the Debian Maintainer community.
With Debian, debconf, and something like FAI, they could perform huge numbers of customized and preconfigured (not merely "kickstarted", but preconfigured) installs all at once. And by using ideas from http://www.infrastructures.org/ and "Apt", it may be possible to maintain them with less work than it would take given any commercial distro. If it was my choice, I'd certainly choose Debian, since the Debian toolset and community support make adminstration tasks so much easier.
Of course it will require well trained full time staff to keep it all functioning. Perhaps that requirement for a Computer Science degree will turn out not to be superfluous after all!
(So did Mr. Stallman ever get to actually FIX that printer?)
Perhaps they're stuck with some from-a-box commercial Linux distro that is not as great after you've taken it home and used it for a while as it seemed in the store you bought it from? Perhaps the costs could be lowered by choosing a non-commercial Linux distribution; spending some money instead on hiring competent administrators who are capable of joining and fruitfully contributing the Debian Maintainer community.
With Debian and something like FAI, they could perform huge numbers of installs all at once. And by using ideas from http://www.infrastructures.org/ and "Apt", it may be possible to maintain them with less work than it would take given any commercial distro.
Of course it will require well trained full time staff to keep it all functioning. Perhaps that requirement for a Computer Science degree will turn out not to be superfluous after all!
(So did Mr. Stallman ever get to actually FIX that printer?)
The authors use Emacs, and only know emacs. It would take too long to learn VI.
The Teach Scheme Project and the "How to Design Programs" web sites have some very good arguments in favor of using Scheme as a first computer language.
I agree with "How to Design Programs" (MIT Press, available online with a great Scheme environment usable by beginners) as well as experts) that Scheme is a great first language.
Helping out on the Debian project will provide that sort of experience. Check the website for information about how to join the team and help out.
I think that speech data entry is inefficient and not appropriate in most office environments. Think of how noisy it would be if everyone spoke to their computers!
What would be really wonderful is a Gregg Shorthand recognition system, for palmtop, laptop, and desktop digitizer pads. It would be a lot faster than the current text recognition systems, and maybe even faster than a keyboard for prose input. I don't think that Gregg is being taught as much as it used to be, but a freely available Gregg input system would bring it back for sure. There are already several gesture recognition programs out there. Gregg is something like that.
Hoe a row.
No joke, I'd warrant.
Sure, the current "boot-floppies" installer is not that great. Historically, the Debian installation system has not gotten much attention, since once you install Debian, you just upgrade after that. You never need to use the installer again.
Are you aware of the "debian-installer" effort? Development has been postponed until after the Woody (Debian 3.0) release, since it was too far from completion when "testing" was created, and many of the people working on it are the same people who work on "boot-floppies". They can only do so much in a day and one thing at a time.
"Debian-installer" promises to be quite an improvement over "boot-floppies". It deserves attention from developers and folks wishing to base their product on Debian GNU/Linux. It sports a wonderful modular architecture that should be easy to use as sort of an installer toolkit or framework. It will be easy to customize for totally automatic, or interactive installs, for over the network or CD use.
As for using Debian GNU/Linux on the desktop, I must say that many people out there are doing just that. The "menu" system is wonderful, and we have both Gnome and KDE working quite well. The X Window System is configured upon installation via "debconf", and is at least as good as Red Hat's XConfigurator was last time I ran it.
Sounds like marketroidese FUD to me... an anon cow too.
Debian is more popular than the for-sale (don't
upgrade; reinstall for another $50) distro
marketting departments would have you believe, IMO.
We are deB.ORG; you will be freed.
When you create a package of the software, the
library dependancies are gotten for you by the
packaging tools. It's just a matter of installing
everything it needs and building it. After the
packager does that, everyone else out there who
wants to use the software can just run "apt-get"
and Everything Just Works.
It is better to use libraries than "reinvent the
wheel" and thus have a lot of programs with
unshared mostly identical code in them taking up
RAM. At least on a decent Linux distribution
with good package management and up to date
development libraries available, you can get
a very high level of code sharing going on.
This is in contrast to the Windows environment
where every vendor has their own proprietary DLL
and very little code sharing can happen between
vendor's softwares. That plus there is not any
central package management system in place, afaik.
Whats really a shame is seeing row after row of PCs running scrolling banner screensavers at a college computing lab. Here in Portland, OR, I've seen that at PCC (pcc.edu)... I asked the people in the lab why don't they turn on the DPMS? They guy said he'd ask his supervisor about it, but nothing happened. (Perhaps the supervisor was too busy blowing smoke.)
The author of that article obviously does not know about Apache mod_perl (http://perl.apache.org/), HTML::Mason (http://www.masonhq.com/), or Embperl (http://perl.apache.org/embperl/).
Those provide a PHP or ASP like templating system with full access to the Apache API and all of full fledged Perl (and thus the CPAN modules). It's way better than PHP!
Anything that can be done in PHP could be done with either of these systems. I think it would be worth doing to port a lot of that fancy PHP HTML magic over to either HTML::Mason or Embperl. (May the better engineered solution win!)
There is, additionally, an emacs extension called "mmm-mode" (found via the HTML::Mason site) that makes psgml-html mode and cperl-mode co-exist! The perl parts are in cperl-mode, and the HTML parts are in HTML mode. It's really cool!
I hope they send one to Richard Stallman and ask him to fix it for them. Wasn't it a printer with firmware he couldn't have the source to what made RMS start the FSF? 8-)
DEC SRC has a Virtual Paper web site. They've got a thing called "Lecturn" that takes a scan and turns it into a document consisting of a series of compressed bitmaps, and, if your arch has the supported OCR library available, the text annotated with bounding box information so that a search can highlight the right spot on the bitmap. The source is available as a Modula-3 program. Lecturn documents are HUGE. If the similar thing could be done with PDF, and the document files are much smaller (better compression) it would be a good thing to have. I also recall seeing a (binary only) scheme system that was up for free download that had OCR stuff built into it. It was kind of like a `festival' for OCR, I guess.