Macos X in Science (Physics) vs. Linux Debian + Windows XP
Hi there,
I am working in University Science (Physics/ Theory and Biophysics) but I have to be quite compatible with Windows office apps for some non-scientific office work, too.
My IT department suggests using Debian Linux + Windows XP with office 2003 (!) in virtual or dual boot setup instead of Leopard + MSOffice 2008 for Mac. It claims hardware and software cost advantages, and a homogeneous network policy, denying any usability arguments for my own productivity. I can't quite follow....
Any suggestion / experience / unique apps for Mac to convince them (or me :( )? Or do you know any review article comparing pros and cons of Macos X vs. Linux or Unix + Windows for a similar working environment?
I appreciate your answers, thanks!
Joe.
Macs in Science
I guess all departments are different but the physics departments that I interact with seem to moving to dual boot Macs, they previously had a mixture of SG, Sun, Linux, Mac etc
I'm trying to put together a clearer picture of the use of Macs in universities, I've put an initial comment here and I'm in correspondence with a few people to try and get more information, but if anyone has another data points I'd be grateful. I'll put an article on MacResearch when complete.
Linux and Windows virtual
I used Windows with a virtual Linux system on it (or was it the other way around?) several years ago. I found this to be a HUGE pain in the neck, because I was always wanting some file from the Linux side in the windows side (or vise versa) and then I would have to go through the process of using winscp to move it. It could be done but it was a real pain.
While I had avoided using Macs for years, I went and bought myself a G4 mini to see if the combination of Unix and ease of use that Apple claimed would be acceptable. After an initial learning curve, I found that it was the perfect set up.
A further thing that I have found since then, is that anytime you get a new linux box, or update the system on the one you have, you have to hope that someone has written the drivers for it. We have a number of linux computers in my lab that don't run some of the programs that we rely (ie VMD) on because we can't find the drivers.
On further reflection
Dual boot is just out of the question, you need to switch between the two too often. The main problem with the virtualization choice is that it eats too many resources. Everything kinda drags when you need to run a calculation. If you only needed to run the virtualization every once in a while, then that would be ok. But, I had it running all day long every day.
You could tell them, that the only way you are going to use Linux and Windows is if you have two NEW computers. Then all of the sudden, the price difference is in the Mac favor.
You could also tell them that they would not have to administer the Mac, that you can do it all yourself. Finally, if worse comes to worse, you can just tell them that you think that the only reason that they won't let you get a Mac is that they are trying to keep themselves employed.
How about the homogeneity argument?
Thank you for your quick replies - I appreciate it.
Yes, you are right: In dual boot it won't be suitable to use MS office apps while running heavy scientific calculations under linux, because the virtual machine is eating up way too many system resources. With a Mac I could still do it, because MS office is just another not too big app.
Of course I could and would like to administer my Mac myself; won't be much effort anyhow and definitely less than waiting repeatedly for dual or virtual booting. My group IT wouln't mind, but the head IT says either it doesn't have own resources to do the admin or it says I am not being paid to spend time on admin'ing it - depending on what standpoint I take. Well, I am not commenting on that any further...
However. How about the network homogeneity argument? They seem to fear I might unknowingly and repeatedly cause local network interferences they will have to deal with or that I will be otherwise a security thread since they won't be admin'ing and thus updating my Mac themselfes.
Thanks,
Joe.
how do you prove something is a non-issue?
Their argument about a mac causing a network problem is something that I don't know how to answer. There are five of us in my linux/windows office that use our laptops for most of our desktop computing needs. Three of us have had no problems whatsoever, we all have Macs. The other two, both running windows have had problems. One of the windows users has had his laptop banned from connecting to the network because he had repeated problems that they couldn't solve-this computer was administered by the IT staff-he finally had to get a new computer so that he could work.
I don't have a shred of anti-virus software on my laptop (and neither do the other Mac users in my office) and none of us have had a problem.
I am sorry that this is a testimonial and not a argument to help you combat the IT stupidity.
As for not getting paid to administer your computer, I have found that administering my own computer (Windows or Mac-Linux is a different matter) takes a whole lot less of my time than trying to get IT people do it for me. What you could do, is keep track of how much time you spend on asking them for help, waiting on them to actually do it, and how long you have to twiddle your thumbs while they play with your computer.
homogeneous network
I did a quick search of the web and I found a good answer, given by jwelch, to the argument about wanting a homogeneous network. I give a short part of it here.
"When you have a perfectly homogenous network, any malware that gets past your protection can propagate as fast as the network and the hardware will allow. You have no firebreaks, no protection whatsoever until a patch can be released."
Re: How about the homogeneity argument?
Hi Joe,
I can speak to all of these things as I am a scientist, a programmer and I manage an IT group for a good sized department that has probably the most heterogenous computing environment I've ever come across:
1) Efficiency. I'm far more productive on a Mac than any other platform I've used in the past (and coming from a crystallography/enzymology background I've used a lot). Having the ability to run Unix apps and desktop productivity apps without doing anything to the config of the system is a big win. That was before the switch to Intel. Now, when I need to run a Windows app or Linux app, firing up VMWare means I get all of the benefits of a full fledged desktop computer (meaning access to the desktop without jumping through hoops) without having to switch boot contexts or disrupting what I'm doing on the Mac side.
2) Cost difference. Good, fast, cheap. Pick two. This ridiculous notion that Macs are more expensive doesn't seem to go away. I can get a $400 white box or $1000 Dell, sure. But when I used Window's and Linux I was constantly fighting the system. Factor in that I can get a great piece of hardware, with a great OS AND still run Windows, Linux, OS/2 (if I felt like it) for a few dollars more and I don't see how this argument holds water.
The more important question isn't what is convenient for them, but how much is your time worth. My time is too valuable and responsibilities are so sparse, I can't afford to be inefficient. And my employers can't afford for me to inefficient either. The amount of money they would save on a computer would be eaten up in two days of lost productivity. Since switching to a Mac I spend less time dealing with the computer and more time doing work and science.
3) Network interference. Huh? We have on our network, SGI's, Linux boxes, Windows (running Windows 95 - Vista), Sun Solaris boxes and Macs (OS 9 - 10.5) and we've never (NEVER EVER) seen any kind of interference on the network. This includes running all sorts of protocols from NetBeui to AppleTalk, Samba, NFS, AFP, you name it we have it. We do see a lot of cross talk in non-routable protocols, but Windows can do this as much as OS X. There isn't much to counter in this regard because it's nonsense and you should state as much. Now if you said you wanted to run your own PDC or mail server I might be more inclined to agree.
4) Security. I can take any computer, do something stupid on it and make it an attack vector on the network. The question is, do you practice safe computing and are you sensible about how you use the system. My policy (regarding others administering their systems) is: We'd prefer to admin it, but if you insist on doing so, when there is a problem we detect (hacked, spewing viruses etc...) we'll knock the computer off the network without notice. Do computers get broken into? Sometimes. And we deal with it. If a computer gets hacked twice, we take over it or it stays off the network. So far, it's never happened twice.
Similarly, we recognize through reasonable security practices, that there are vulnerabilities that can be exposed in core services when someone is an admin on their computer. So we have policies such as no NFS mounts to computers we don't fully administer. Simple. Is it perfect? No. Can things still happen? Yes. And we deal with those things when they come up.
5) In terms of WHO should admin what, beyond getting the computer on the network or accessing shared resources/servers, I fail to see how you are spending your time is the business of the head IT person. Again, our policy is, if you want us to admin something, we need to have the resources and it needs to be in a particular config. If you want to admin it, we'll limit what resources (on the network) of ours you can access, but do as you like. See point number 4 as well.
Regards,
Dave
Thank you!
Thank you so much for taking the time to fire up my arguments - excellent job!
Am I overseeing any answer rewarding system similar to apple's discussion forum?
Joe.
Rewards?
Hey, this is a community effort. The reward is knowing that other scientists can use their platform of choice.
As far as I'm concerned, that goes into the "benefit to society" part of my job.