I did some work with MAX at CNMAT and on my own a very long time ago, but since then I've led a MAX-free existence - don't know how that happened. Anyway, I finally have the opportunity to get back into it, so I just started to catch up and see how things have developed since I left. Holy cats - they certainly have changed. The fact that I'm on this forum shows you a bit about where my click trail has taken me.

Ok, so I have the opportunity to retool and almost start from scratch, and I'm faced with general questions about the ideal OS / software to commit to. I'm aware that questions like these can be inflammatory, so I'll give a bit more info and justification for being so general.

I'm primarily a Mac OS X dude, although in my main profession as web developer, I use all OpenBSD servers. For one particular project, I may be moving to Linux because of its clustering possibilities. But at home, I do all my work on my G5.

My training, such as it is, is in composition - with pen & paper even. But I'm eager to expand my toolset with various computer-assisted techniques, and I intend to develop my own, not just use what's out there, so programming is going to be a big part of my plans.

My immediate inclination was just to lay down the $495 for MAX/MSP (my copy is so old I don't think I can even upgrade). Maybe add jitter too - looks insanely cool. But then I clicked on a link to a company that does cycling74-based work, and followed a link to Pure Data. And things really opened up from there.

So it looks to me as if there's a bit of a conflict here...

• If go with what's semi-familiar and go with MAX on my Mac, I immediately also get an IRCAM Forum Pass. Quite a bit of $, but a lot of bang. Good support. Established community.

• But then there's Pd, which is open-source - that's a HUGE plus in my book, not just because it's free, and it'll run on my Mac...

• ...but from what I can tell, a lot of the most interesting development in this field would appear to be taking place on the Linux platform, AGNULA, dyne:bolic, etc.

So that's my question. MAX vs. Pd, OS X vs. Linux, and the natural combinations. Will Pd be a practical alternative to MAX/MSP, with its established user base and support, or even an improvement? Am I shortchanging myself with Pd on OS X? Might it be worth taking the plunge into Linux? I'll be interested to hear all advice and experiences.

Thanks...
Alex

The most interesting thing I read this morning was this interview posted on Digg,
http://sztywny.titaniumhosting.com/2006/07/23/stiff-asks-great-programmers-answers/
in which David Heinemeier Hansson says in his opinion "a strong sense of value" is essential to creativity.

That's at the nub of many a flame war - really differing sets of personal value judjements within a landscape of options that are't really that different when you look at it critically.

Learning Linux will teach you more than just Linux, it's a great way forward to understanding more about computing in general. Do you want to do that? Do you want to understand the machines at a lower level? I always reccomend Mac to "artists" even though I am a pure blooded Linux zealot and grand master of the Templars of the Church of the Holy Penguin, sworn to slit the throats of non believers on our crusade against the darkness. Sometimes it's just better to point people at what is right for them.

$400 to Cycling74 will help sustain a fine commercial software enterprise and it will buy you support. You need to value that, do you need the support and do you think the bargain is worthwhile (consider your own comment that you "cannot upgrade").
What are your past experiences of commercial software support? My own decisions were based on wasting $500 on Reaktor to find a years worth of my patches were scrap because I couldn't upgrade. That was a deliberate decision by Native Instruments to enforce a dongle based copy protection that they knew was screwing over their users, so I abandoned NI products forever.

On the other hand, Puredata is in a constant state of development. Do you need stability? Do you value that over cutting edge development? If so paying for a tightly version managed product may save you a lot of time and headaches. It may also create as many when they won't do what you want or need.

As a classical (writing) composer, you are certainly going to want integration with Logic or Sebelius or some kind of scoring package - so what is it about Pd that is attractive? There are no score interpretation marks other than MIDI I/O .

Anyway good luck figuring it all out.

I switched to PD 3 years ago from MaxMSP and never looked back, the two programs seem to go through periods of convergence and divergence - an innovation will often appear ported for it's neighbor in a relatively short time. Generally I suppose Max is a bit more juicy graphically which can be attractive to the Mac user. However, PD really comes into it's own when combined with linux as I have recently discovered:

As a recent convert to Linux I first tried the CD boo-table distros like dyne:bolic and DeMuDi from the AGNULA project which I found a little clunky. After further searching and testing I've settled on a Hard drive installation of Puppy Linux - compiling PD for this OS. At only 70MB or so Puppy is tiny and this frees the computer up for all those complicated patches - I even have PD running on a Windows95 spec computer relatively well. The other awesome Linux experience I've had is PpPod for iPodlinux - portable PD programmed instruments in my pocket!

Long story short I teetered on the brink of the same decision and PD+Linux has done me very well.

lead - thanks for the info. That pretty much aligns with my first impression.

obiwannabe - also thanks for the info, although I'll mention that I added those details about OpenBSD servers etc. to try to communicate the point that I'm not a "mac person" in the way the term is typically used, i.e. a technically retarded creative artist - I use OS X as much for programming as for PhotoShop, and I've worked quite a bit with Linux in the past, so learning it won't be an issue.

Re: scoring integration - I can't imagine I'm going to be outputting notation from Pd - my composition and my dynamic music experiments are pretty separate, although I can think of a few ways that it might be possible, e.g. MIDI -> FOMUS -> MusicXML -> Sibelius.

So far I'm inclining in the direction of Linux, although it's not an easy choice, as I have an awful lot invested in OS X... maybe set up a dual-boot system? I think I have a project for the weekend

Dual boot can be a royal PITA. Not because its technically problematic but because you find yourself split between two workspaces until you commit to one or tother.

Don't forget the liveCD distros like PureDyne, but the advice by Lead about picking a slim distro is spot on. Puppy and DSL are good candidates, but you will need to configure them a bit.

I roll a fresh Debian build using apt-get and just put on the basic essentials

X-Windows
Blackbox or XFCE wm
ALSA
Jack
Puredata (38-4 extended) from Hans site

Sorry if I seemed to labling you a "mac user", in fact I heartily encourage you to go with the Linux route - you will get a vastly superior system to either WIn or Mac and scope for unlimited experimentation. However, if you are a complete Linux noob perhaps it's best to start with a stock "complete" distro like Ubuntu or SuSe first time around. Best of luck.

axolotl still around? Which direction did you end up going?

I'm a Linux newbie who's looking to dual-boot on a MacBook with a complete distro and would like some advice as to which one will offer the best audio and Pd performance.

Eventually I'll roll my own but I want to start with something stock. I'll also be using ChucK and SC3 and an RME Fireface.

Thanks...