A recent article caught my eye and turned it a nice shade of red. It discussed the -- hardly new -- idea that the future of software usage must involve a mixture of free and proprietary products -- something the writer refers to as "mixed source". The piece was entitled "Mixed source - the best of both worlds" which may give you a clue as to where I disagree with it.
The article was an opinion piece by Steve Harris, senior director for open source products at Novell in issue 78 of Linux User & Developer magazine. Sadly it's not yet available on-line and I don't honestly know if it will be. If it is I'll post a comment with a link here so you can read it for yourself.
Nothing new under the sun
As I said, the idea that software stacks will become a mixture of free and proprietary products is nothing new. Indeed lots of people are already using such stacks. Personally I believe that once freedom is introduced into a "market place" it will become harder to suppress until eventually it becomes the dominant licencing strategy. This is evident in the fact that a company like Novell not only bought a free software company (SuSE), but bought into the free software philosophy -- well partly anyway. So while proprietary software may not entirely die out (more's the pity) I feel (and hope) it will become the de-facto NON-standard way of licencing software.
"Mixed source" is a bad name for this -- er -- mix though. The source or openness of it is largely irrelevant if you ask me. When you mix free and proprietary systems in one application stack -- like it or not -- the entire stack has a proprietary effect. Obviously the degree of that proprietary effect will depend on how vital the proprietary software is to the stack. Use a free software database back-end with a proprietary front-end and your stack is largely subject to the whim of front-end's vendor.
The best of one world, the worst of the other
What particularly bugged me was a comment about where the split in this twin-licenced software stack would come. I quote:
"A mixed source approach enables customers to keep costs low with open source software while also minimising threats, protecting key corporate data, reducing network administration costs and complying with regulations with the help of steadfast proprietary solutions."
If you'll pardon the pun, that implies that open source/free software has one value -- cost. It also implies that issues like security, networking and regulatory compliance are only really met by proprietary software. Both of these implications are wrong, plain wrong. The (in)security of proprietary products is dependant upon secrets kept not by you but by your supplier. If you "protect" your corporate data with proprietary systems (and by implication proprietary protocols) your data belongs to your supplier, not you. By contrast if you use free software in your security, network administration and data protection you are as (if not more) secure as proprietary offerings but you are not reliant upon your supplier's ability to keep their product secure. He goes on...
"when adding additional security functionality to their open systems, many customers are actually reverting to a mixed Linux/proprietary solution."
This is the problem with the word "open": it mean openness and open windows
Again the implication here is that "open" means less secure. See this is the problem with the duality of the word "open": it can mean openness (in terms of sharing) and open windows (in terms of unwanted access).
Square plugs in round gaps
Elsewhere Mr Harris speaks about addressing the "gaps in the capabilities of open source" by further open source development or proprietary solutions. Yet again the message is that FOSS has gaps which can only really be plugged by proprietary software. Yes, he does mention further FOSS development, but the thrust of the piece is essentially that free software can't cut it and requires bailing out by proprietary systems.
What good does it do you to free your legs from the shackles of proprietary software only to voluntary tie one hand behind your back with a proprietary plug?
It's true that free software probably can't cut it in every fields (yet) but you know what? Neither can proprietary software; implying that an average business could not equip itself entirely with free software is misleading. The solution to any gaps in the free software stack is not to develop proprietary plugs. If the base software is free and the customer is enjoying the benefits of that, then why should they then rely upon a closed piece of glue? Would that plug grow with the main software? What if the base software went in a direction which was counter to the proprietary vendors aims? If the customer is going to be spending money on these plugs would they not be wiser to spend it getting free plugs, created so the whole community benefits. Perhaps they could jointly fund develop with other customers, making the whole thing more cost-effective for all. Honestly, what advantage is there is having a small piece of glue technology created just for you when the underlying system is free? What good does it do you to free your legs from the shackles of proprietary software only to voluntary tie one hand behind your back with a proprietary plug?
I'm speak from experience here. When we wanted some custom modules added on top of our free software CMS (ExponentCMS) I deliberately had the code written under a free licence by my third-party developers. We've not yet deployed the new features; however, once we're sure they're good to go we'll release the modules to the ExponentCMS community. Had we got it done it in a proprietary licence, the modules would have remained static -- we probably won't have the resources to fund future development other than in-house stuff. By freeing them up they can be further developed, learned from and built upon in a way that proprietary modules could not.
Not so hidden agendas
I should acknowledge that the real and, given the position Mr Harris holds, somewhat understandable goal of the piece seemed to be promoting the fact that if this "mixed source" world grows then end-users will "need support from vendors and service providers who understand how to connected(sic) open source and proprietary technologies". Hmm I can't think who he might have in mind there, can you?
The closing argument of the piece is:
"Vendors who are well-connected with the open source community yet also able to provide and support proprietary solutions are the ones who are most likely to prevail."
Time will tell if that happens but I very much doubt it will, especially if those vendors promote the idea that you can't have the former without the latter.