MacResearch

I think no.
From a hardware point of view there's no difference at all (you can buy a PC with the same hardware), at least for computational tasks (CPU and/or GPU).
The OS makes a difference but not in "research" itself. It much depends on what tools you are dealing with.
Many tools are written in python/java/perl which allows you to run on any platform.
You may have some troubles in compiling C/C++ software because of some differences in libraries.
Apart from this, there is a bunch of software that is PC only (either Windows or Linux), not to mention all instruments that are attached to Windows computers...
There are some fields (microscopy and imaging in general) that take advantage of OS X features such as CoreAnimation, CoreImage and CoreVideo...

/*
Davide Cittaro
NGS data analysis

IFOM - Istituto FIRC di Oncologia Molecolare
via adamello, 16
20139 Milano
Italy
*/

Yes.
Literally because Macs have, on average, better hardware quality than run-of-the mill PC.
Meaning less downtime, less hassle. I know because I have to maintain them.

What about the OS argument? (Many tools Windows/Linux only.)
Here's where modern Macs with Intel processors shine: virtualisation!
Use VMware (or Parallels if you prefer, or Sun's free Virtualbox if it works for you, or BootCamp if you really, really must), and you can have any number of different OS configurations on one machine. I have XP software that won't run on anything else (certainly not Vista), Windows 2000 spectrometry software that's gone orphan (with the last Win2000 PC retired years ago, and no new versions available), several Linux flavors, and even Solaris for the odd project. All on different virtual machines on a FireWire HD alongside my MacBook Pro. Huge savings. Very comfortable.

And if some of the tools you need are natively available on Mac OS X, you'll never look back.
It can be so liberating when things just work.
So, even if the initial investment in a decent Mac is a little more expensive than a PC of comparable quality (not just numbers on paper!), it's really a near-universal toolbox.
Just make sure you get enough RAM — the temptation to run everything in parallel is huge, once you get the hang of it.

I've been using a Mac Pro for the last two years as my main compute machine. My opinion of this platform as compared to linux varies depending on whether I'm trying to install new software or not. I do bioinformatics research and I use a lot of open source software tools. Most programs can be installed with little or no problem on the Mac but some can be difficult to compile and install. Lately I've been using linux virtual machines running on VMWare Fusion for data analysis but a single virtual machine can only access 4 of my 8 cores and disk performance is pretty bad, compared to the host Mac OS.

On the other hand, Apple hardware is easy to buy. Companies like Dell and Sun offer a bewildering array of chassis, cpu and storage options and unless you're really familiar with server hardware you'll spend hours trying to configure a system on your own.

If you aren't tied to Excel, per se. You might want to look at DataGraph or even DataTank, but VisualDataTools. Both programs are awesome for data analysis. DataGraph in particular is pretty intuitive, does a lot of nice 2D plots and can handle a lot of data points. It's inexpensive and the developer is extremely responsive.

http://www.visualdatatools.com/DataGraph/index.html

Dave

I've been kind of stuck on a spreadsheet, as that's what I've mostly used in the past. I've seen the Macs in Chemistry list before, but forgot about it. I'll check it out--perhaps this is time to expand my alternatives a bit.

Thanks

It works better than Excel, meaning that it at least does the minimum, but still underperforms compared to my benchmark--a 5 year old PC laptop.

The other thing that hangs me up is that I have gotten used to a nice add-in in PC Excel that really helps in navigating and formatting, etc (Spreadsheet Assistant). It has lots of easy right-click options for selecting cells, making it easier to format tables for input into reports.

I'm checking out these other options right now.

If you run a lot of unix/linux scripts/applications and do a lot of scripting but don't necessarily want to spend time on adjusting/tweaking/fudging/getting-mad-at unix/linux, it is a definite yes.

I guess that if you are almost interested in MS Office products (Excel), a PC does the job just fine.

If it's a mix of both, I'd still recommend a Mac for the versatility and stability. You can still run Windows through Parallel/VMware or Bootcamp if you are not happy with the Mac version of Microsoft products.

Personally, I switched to Mac from a Linux/Windows dual-boot habit a few years ago and I wouldn't go back to it to save my life. You may call that Applephilia. :-) But, more seriously, my productivity increased drastically ever since.

I'm a researcher in volcano geophysics. I can run all my codes/applications/workflow on one single machine using Mac OS X. Scripting is a significant part of my work and requires *nix platform. But I also work in a Microsoft environment so run Windows through virtual machine for day-to-day admin stuff. When it comes to manipulate and store data (actual data or references), I found Mac OSX far faster and easier to use than MS Windows (e.g. my workflow includes a mix of DevonAgent, DevonThink, Papers and Applescripts). For producing and manipulating equations and graphs (e.g. postscripts and pdfs for publications), Mac is also, to my opinion, superior to Windows. An important point is that Mac OS X comes with a lot of scripting languages and various servers out of the box which make scripting and remote connection to my work machine a piece of cake. I guess my point here is that I can do far more with Mac OS X than with Windows without installing third-party applications. When it comes to presenting research results, I also find Keynote to be far superior to Powerpoint.

Then, there is a bunch of applications that I use which run on both Mac/PC (e.g., PersonalBrain, LaTeX, Matlab). But even for those, I found Mac OS to be more to my taste.

A possibly futile point is that working on a visually attractive environment also makes work nicer and somewhat "easier" (at least to me). In other words, I am happier to script/crush some numbers on a Mac rather than under Windows. This, obviously is entirely personal.

Example of workflow from a few weeks ago:
1. GPS processing using shell and perl scripting and GAMIT/GLOBK (*nix only)
2. Mass-plotting with GMT (hundreds of plots)
3. Mass-conversion of ps to image format using Workflow
4. Integration of the plots into a movie with iMovie.

No way I could do that with Windows. And doing it with Linux would have taken me far longer.

Hope this helps

Nico

www.volcanoloco.org
www.gns.cri.nz

> they use the same hardware, don't believe the lies, you are paying for the name and the OS

This is only true in theory. In practice the MacPro comes with a server-grade Intel-brand motherboard which costs several hundred dollars, not the $50 commodity stuff from Via. People have added up the component costs, and come up with a total of $2200 in parts for the $2400 2008 MacPro. So you are paying $200 for the design, assembly, name, and OS. Seems reasonable to me.

I'm primarily a computer science guy, however I've done a lot in other fields using Windows, Linux, and Mac OS X. I firmly believe that Linux and OS X are in a different league when it comes to data processing, with a plethora of academically-founded open source scientific applications, libraries, etc. that are relatively simple to set up. Yes, Windows has Cygwin, but just about every experience I've had with it has been awful.

If you only process data using a spreadsheet, Windows is probably the best option; as my girlfriend who works in a neutrino lab recently found out, the spreadsheet options for Linux and OS X are seriously lacking. Numbers.app is a very smooth and useable spreadsheet app, but it simply will not deal with large data sets. Open Office's Calc did not do any better, and the interface is pretty rough. She was surprised to find that Excel for Mac also did no better, which after some thought made sense to me, since the lack of spreadsheet capabilities on OS X helps keep labs and businesses on Windows.

Mathematica, Matlab, Maple, etc. are bearable on OS X, but I haven't used them much (and not in a couple years). As I recall, they were pretty slow on all platforms.

Honestly, hacking something up in your choice of programming language (java, perl, python, ocaml, haskell, etc.) is the best way to go on any platform. This does not "lock you" in OS X as a previous poster has claimed-- far from it! I have simply and easily ported all sorts of scientific software between platforms (granted sometimes it's a pain, and generally I do not relish porting software). Conversely, many Windows libraries are only available as a binary-only distribution or otherwise have dependencies on binary-only software, and as such those will limit your cross-platform ability.

That said, I find that many labs are not so adventurous when it comes to software, which means that 99% of them use one or two of Excel, Matlab, Mathematica, Maple, or one other I'm forgetting. This is a shame, but the reality is that those will probably not be satisfied with OS X or Linux.

However given all options, I am very satisfied with OS X. The hardware is reliable, the amount of system maintenance I have to do is negligible, setup is much faster than Linux, and the end result is more usable.