From your post, you not looking at this with the right perspective, not asking the right questions, nor asking them to the right people. You state that you have been put in charge of "maintaining" and never once mention anything about your company's predicted growth, development plans, future computation needs, near and long term service offerings, uptime requirements, security requirements or so forth. You have to do a requirements analysis that extends to between five and ten years and design a system that can grow seamlessly with your employer, meeting their current and expected needs in all pertinent areas.

If you can develop a system that does what is required on paper, the next step is to implement it in parallel with the existing system, and transition services and users over in phases. After all services have been transitioned, you can decommission the old infrastructure piece by piece.

Haha, to edit my posts...

I had to sit down and install a bunch of the shared decoding libraries for mpegs among other things, she couldn't play mp3s with the out of the box install. Seriously? 3 different apps complained about the lack of suitable decoders, but only one was smart enough to pull up a package management dialogue and give the option to actually do something about it. This is something pretty much everyone uses, why set it up so backasswardly? Every app installed by default that plays mp3s should work. You shouldn't have to guess correctly which app you need to use to play your music.

Oh, of note:

The network manager used in Ubuntu 9.10 has problems with reinitializing network interfaces that go dead for some period of time. When she boots windows the network interface gets automatically re-initialized after connectivity failure, however the Gnome network manager does not seem to get this right very often - I've instructed her to wait until other people come back online then unplug and re-plug her ethernet cable, which fixes the problem. I am on the same network and I have no problems with KDE's built in network manager.

Honestly, I like a lot of the stuff they're doing in Ubuntu, however having JUST set up a complete novice Linux user with Koala and watched the things they had an issue with:

1. The SMB mounting tool is nice, except it doesn't show shares in Gnome file dialogs! The connection it makes is not persistent. Nor is SMBFS installed by default. I had to install smbfs then go in and set up everything manually in fstab, which is ridiculous for a distro not to have covered in a cleaner way. That's not hard for me but come on!

2. Mime types are not properly set up in firefox. With a totally fresh install, a .doc downloaded from the web cannot be opened directly, even though it's listed as the type handler... She ended up going to the containing folder and opening it through the file browser, again this is pretty bad not to have working.

3. Sound settings are not properly saved by the mixer on reboot. In addition though pulse is installed by default it doesn't work nearly as well the way it is configured by default as in some other distributions I've used. I've had to sit down and fix various sound issues several times.

There are probably more things I'm forgetting as well, or that she has not seen fit to bother me with...

On the plus side, the regressions in 9.04 with full screen flash and some types of webcams seem to have been fixed (no more LD_PRELOAD shortcuts). That's positive.

Ultimately, the only thing at this point that is really keeping me considering Ubuntu/Kubuntu over SuSE is apt. YaST is pretty good, but apt is better and the package coverage is also better. I really dislike Canonical's insistence on making you jump through hoops to use "non free" software. I am very pro-free-software, however if anyone involved with high level decisions at Canonical is reading this right now, give me a freaking button I can click during the installation that says "I am a big kid, I can make my own choices regarding free/non free software, I'm not interested in making a big philosophical statement with this computer, please include non-free software in my basic installation".

I'm sorry, but Plone kills drupal when it comes to design expandability and security. I like Drupal a lot but unless you have a huge in house PHP team already or you're not interested in ever utilizing any enterprise level features on your CMS, it is a mistake to use Drupal over Plone or an enterprise level CMS. It's all about the right tool for the right job...

That's completely backwards on several levels...

1. Executives should ALWAYS be last in line for any forced migration like that... The right way to do it is:
A - deploy in test environment
B - staged rollout, starting with bottom teir employees in batches, gradually progressing up the organizational chart.
C - ???
D - Profit!!!

So obviously the accounting department would be transitioned first and that scenario would never occur if you have a halfway competent IT team. I feel sorry for your company you have any position of responsibility in the IT department.

2. Passing around excel spreadsheets is amateur hour... I grimace when I see companies with more than 20 employees do this sort of thing. There are plenty of nice financial projection suites that do out of the box stuff, if you want to do any sort of sophisticated prediction and modelling, use something like R which has robust mathematical and statistical capabilities and tons of finance modules available. If you have money and want to be professional, your financial department should be integrated into an ERP platform.

Microsoft office stopped being relevant after 2003. The whole reason office was such an immovable behemoth was backwards compatibility and the retraining curve for other office suites. As it stands now, openoffice is easier to migrate users to than office 2007/2010 and the whole docx/xlsx fiasco has caused a lot of headaches. Few people care about the added functionality in newer versions.

That's all wrong.

The accounting department should be switched first (execs should always be last in line for any sort of forced migration like that). In addition, if your accounting department is sending around a bunch of excel spreadsheets your shop is amateur hour anyhow... *cough* L2ERP n00b *cough*

It was a typo... what they meant to say was "geographies with a lower average (w)age"...

As a mathematician/statistician, I try to encourage people to approach the subject from a wholistic and philosophical standpoint. I prefer applied maths, and for most people I think that is the direction which tends to be most rewarding. The way that I make these interests dovetail is by trying to understand the foundations of various branches of applied mathematics and how they relate to each other. For instance, there is a great deal of overlap between statistical physics, machine learning, bioinformatics and mathematical finance. A rigorous understanding of probability theory, information theory and real analysis will take you a LONG way in any branch of applied maths. Stochastic modeling is at the core of everything applied, and being able to decompose a system into components (be they wavelets, sinusoids, eigenvectors, polynomials or what have you) is incredibly important as well.

Ultimately, if you are going to spend the next 4-7 years in a PhD program you should be really passionate about what you are doing. Several other people have mentioned this as well but it's important enough that I think it bears repeating.