Teaching Linux/Unix Basics to Microsoft Junkies?
flupps asks: "I've been asked to hold a two-day crash course in a class of students that currently are studying to become MCSD certified. I'm looking for ideas how to set this up. I was thinking about starting with some general file system descriptions, where to find what files, the man pages, the tab-button, etc.
After that move on to some of the daemons and just explain what they do." He's got at least one idea to start with (below), but what must-have skills or demonstrations would you add?
I also plan to set a database program in VB (one of the certificates in the MCSD suite) against a MySQL or Postresql db and show that there are free alternatives that works as well as SQL server.
What would you think could be a good addition to teach them?
This is in no way meant to be a very advanced course, but I want to show some of the excellence of *nix and why you sometimes can save time and stability and maybe make them interested and read up more by themselves afterwards.
Any suggestions very welcome.
is teaching cat | grep . I don't think I use any command combo more than this other than ls -al. Piping and redirection is really important stuff for Microphiles to learn right away. It's a great way to show off the power of a CLI.
Make sure you teach them how to compile and install software. When I first learned *nix I learned how to navigate the file system, run things, edit files, move things around, etc. But it took me like a week to figure out how to install and set up new software. I remember having the hardest time with it because every single piece of software was different. There was no standard setup.exe or *.rpm all the time. I had to make, make install. And that didn't always work either. That, imho is one of the major differences and difficulties there is in moving from windows to *nix. In windows once you've installed one piece of software you've installed them all.
The GeekNights podcast is going strong. Listen!
instead of doing like MCSE and giving them fish, teach them how to fish.
/etc - it's where most of the config files are. /usr/bin - it's where most user programs live /usr/sbin - it's where most superuser programs live
/all - in unix we have ifconfig' some basic translations of basic stuff.
"This is
This is
This is
If you're interested in using a command and don't know how, use 'man command' and get them familiar with how to use commands. "
You've got two days - so some basic 'how to get info' and then examples of getting that info, would be good.
Possibly a run down of 'in Microsoft, you have IIS, in Unix there's apache, ftp, etc'. 'In MSFT, you have ipconfig
How about running through the 'Administrative tools/Common' menu in 2000 and showing them where all those toys live in *nix - or where they might be able to find them.
But make sure you teach them how to fish for themselves - I suppose MSFT has the help pages, but man pages are our best equivalent. Or homepages for the package in question where applicable.
Good luck!
-- There is no sig line, only Zuul.
I also plan to set a database program in VB (one of the certificates in the MCSD suite) against a MySQL or Postresql db and show that there are free alternatives that works as well as SQL server.
I would qualify that - you'll probably have at least one person in the group who's up on MySQL and/or PostgreSQL deficiencies (yes, they have them). Don't try to convince them that MySQL can be a drop-in replacement for SQL Server 2000. Both MySQL and PostgreSQL *can* be used in many situations, and should be considered along with other options re: price/performance, but don't go overboard and talk down to MS people saying MySQL is as good as (or better) than SQL Server. It does a disservice to everyone involved.
Covering RPMs and/or apt-get technology might be useful at the end of 2 day overview.
What would help more than anything else is showing people where/how to get help - online resources (RPMfind, for example) and whatnot. There's only so much you can cram in to two days - don't overdo it. Cover the basics in detail, and give resources to visit afterwards for people who want to learn more and/or experiment.
creation science book
Here are a few random suggestions, in no particular order.
* Open a relatively complicated page in MSIE, the same page in Mozilla-win32, and the same page agin in Mozilla-linux. Go to a bunch of annoying web sites, with Mozilla's pop-up/pop-down filters enabled.
* Use ssh to log in to a box halfway across the world. Demonstrate some simple system administration tasks, and the fact that anything you can do at the console you can also do remotely, via ssh.
* Run either Gnome or KDE. Change the themes, a couple of times, demonstrate the customizable UI. Switching between one of the mac Aqua-like themes, some star trek theme, and one of the Winxx-lookalike themes should be very effective.
* Install a distribution in server mode (no X11). Demonstrate the extreme modularization of Linux, such as you can complete get rid of all GUI support, and use only the disk/network services to turn a box into a network appliance.
* Install Windows and Linux on the same box. Boot into Linux; then mount and browse Windows partitions. Make a casual remark that Windows cannot browse Linux partitions in the same way.
* When the Linux box boots up, and is busy going through the initscripts, starting all the services, explain that if one service fails to start for some reason the boot process will continue and the machine should still be mostly usable. Ask if anyone experienced a situation where a Windows driver kept croaking during the boot process, and what happened alter.
I recall an incident about three years ago when UMAX shipped a buggy driver for their scanners. The driver was faulting on machines with USB ports, and CPU speeds over 400 Mhz (something about some timing loop), forcing a complete crash during the Windows boot cycle, with the subsequent reboot falling back into safe mode.
The Linux equivalent for this would be something like SANE, which runs completely in user mode, and therefore cannot crash the entire OS.
* Use samba to browse the local windows network neighborhood.
* If you have a fat pipe, forward X11 over ssh, and run remote X applications on the local terminal.
* Install a base distribution package right out of the box. I'll use Red Hat 7.2 as an example. Apply all the errata to bring the box up to date, except for the kernel, without rebooting. Even install a new version of glibc (the equivalent of msvcxxrt.dll) without rebooting the box. Install a new kernel on the remote machine, make sure that LILO or GRUB is all set up, then remotely reboot the box into the new kernel.
It helped me over the hump when I became serious about learning Linux, and I use it as a resource still today. Even if you don't use it as a guide for your class, I'd highly recommend that you mention it to your students.
--SC
You read fiction? I write it! Lemme know what you th
Show them the ease of modprobe, the powerfulness of iptables,[...] mounting of ISO-images(!). [...]Inprint in their heads that the machine does not need to be rebooted after updates and installations, ip-adress-changes and change of configuration.
/etc, /etc/sysconfig, /etc/$SOFTWARE_NAME, /opt / SOFTWARE_NAME, /usr/etc, /usr/local/etc, ~/.$SOFTWARE_NAME, /usr/share/$SOFTWARE_NAME ... but no registry. Woot!
They are developers, for crying out loud. They'll start running after 5 minutes of iptables; why not show them gcc?
Show them the [...] the geniality of config-files, and NO REGISTRY.
Look, config files in
Don't get me wrong, I'm a 100% unix guy; but it seems to me that exactly this kind of arguments makes people stay away from linux. You don't have to crush Windows, you have to give them reasons to make them beleive in linux, and to want to hop in the wagon.
-- No sig today
Comment removed based on user account deletion
Most of the comments I've been reading have been going off about administration and configuration. These are people going for a developer certification. That means they probably already know a lot about programming, and they probably already know a fair bit about programming under Windows.
Going in there and giving them a Linux sales-pitch would be a waste of their time.
Database connectivity sounds like a cool thing to demonstrate, you might want to demonstrate the basic development tools and documentation available at their disposal. Show them an easy editor to use.. something consistent with the editors used in the Windows world, show them gcc and some neat stuff like xxgdb. The ones who are clueful enough to care will pick it up when they leave.
IMHO the most important thing to explain to them is software licensing. It is quick, but when they realize that if they like to develop software, they'll clue in that developing their apps for Linux is easier.
A few tiny things like that would probably take up all the time you have. Cygwin might help them know how to develop apps from Windows to target Linux boxes.
My experience with MCSDs and other Windows developers is that they don't really care about the OS, they just care about writing apps and using OSes to make money by solving problems.
You need to back up and do something more fundamental before you start showing them filesystems and daemons. You need to compare the two competing philosophies that drive Windows and Unix cultures.
After this balancing act, then you can begin to lead them down your path of showing them practical items. At each point you can refer back to these fundamentals. For example, when it /etc, you can explain why Unix admins think text file configuration is inherently more stable and powerful than registry keys, because without such an explanation the Windows admins will typically see it as quaint and backward. Again, when you get to /dev, you can show the inherent debugging power of being able to do things like "tail /dev/midi00" to debug a connector on the computer, even if that data is not useful immediately. You can show how grep, awk, and perl can be chained together to do advanced data processing (on text) that would not be possible on Windows without a specific feature to make it happen. The key is to refer back to a specific philosophy for each exercise, so they can see the big picture.
None of them will watch a hands-on lecture and run out screaming "I've got to convert to this immediately! He broke out this thing called grep and it was.... it was.... AMAZING!" :-) Rather, you want to give them a clear understading of our culture, and just like how a high school senior goes to a college campus and says, "Yeah, I can see myself here" you might kindle an interest in some of them to find out more about how we *nix people think.... and that would be the first step to bringing them over.
Use the right tool for the right platform.
Sure, DOS has had scripting and pipes from day one (well, unless you tried MSDOS 1.0). Were they as useful as their Linux counterparts? No freakin' way.
Why does TYPE not take stdin? Why is "copy con" equivalent to "copy con:"? (And, why is "copy con.txt" ambiguous?)
How can a batch file determine if a directory exists? Hint:
if exists c:\foo\con
yields different results in different DOS versions
DOS for the longest time failed the basic tests. And for the longest time, I was working with the MKS toolkit, replacing the ones that didn't quite do what I wanted them to do with copies ported from comp.sources. But it never became UNIX.
NT is still rife with inconsistencies in the CMD shell, and I don't know (nor care to know) if or when they get partially fixed.
The point is: if you want to use Windows, use Windows tools. Learn how to use VB Script to its effect. Learn MSVC if you must. Prentending that it's another UNIX if you squint right will hurt you. Windows is not designed to be UNIX.
Every time I use Windows on the premise that an OS is an OS and a command shell is a command shell I get hurt. I should have learned that lesson from VMS years before.
Does anyone knows if the Posix subsystem still exists in Windows XP? That was the worst checkmark compatibility I ever saw. You could run Posix code on NT, to allow NT to be purchased by the federal government. And unless you wanted to do actual work with it, the compatibility was fine.
It is completely beyond me why people are porting Apache to Windows. NT comes with a perfectly functional web server, why bother replacing it? Don't get me wrong, I hate IIS with a vengeance, but the loopholes in the underlying operating system (like the $::DATA bug) will have to be special cased in Apache too. And the $DEITY like privilege issues that plague the IIS indexing server will plague Apache just as well.
Possibly even worse, because code ported from UNIX will have to be modified to suit NT's security model, a redesign from scratch really is the only appropriate way to deal with such huge gaps in design philosophy.
Bert Driehuis -- All I asked was a friggin' rotatin' chair. Throw me a bone here, people.