Is “Open Source stack” the new buzzword for 2005?

Matthew is blogging (again) about SpikeSource and Open Source support. I definitely share his view (well, that’s certainly not the first time) and I’m quite puzzled about all this “Open Source stack” thing.

As I understand it, the idea is pretty simple: you gather a few OSS “products” that potentially fit together to fullfil an infrastructure as a whole (say an OS, a web server, an application server and some middleware) and you either package and redistribute it with a nice name, box and manuals (the RedHat way), or you “just” claim that you can provide overall support on the whole thing.

Now, the last one is a serious claim to make, expecially when middleware becomes part of the game. I can certainly foresee some traditional support around software that has “just” a configuration and a log file: send me your configuration, send me your logs and I will be probably able to sort you out. So, talking about LAMP, this can be more or less easily done for the L, the A and the M. There are some important issue with the P (supporting a programming language is much harder: go figure a middleware slution). But hey, this is plain old support, has been done for ages, I was doing that back in 1992, when I started working in IT. What does it have to do with a “stack”? Marketing buzzword, or am I missing something?

All in all I still believe that supporting a full-fledged solution (OSS based or not) is much different than supporting the single pieces. A solution is like a jigsaw puzzle: what is the point in supporting the single pieces when the real value is knowing how they can fit together?

To me (to us) “support” really means a process that starts with an assessment phase (what do you want to do? what are the proper tools for that?), goes through a mentoring/consulting process (here is the best way to do it, and here is a helping hand for you) and finally lands to a real support line (I know your setup, I know your people, I can really help now). There is still room, of course, for bulkish traditional support (hardware, infrastructure software and so on), but still I don’t quite see how companies like SpikeSource can provide real value to their customers given the current state of affairs in software-land. Unless they plan to move towards a network-based business… but this is something for a future post.