My mother always told me never to discuss religion or politics at a dinner party.
I was never much for religion, but I was always happy to let my friends know why Paul Keating was the best ever Australian Prime Minister. One thing mum never warned me about was the new taboo dinner party topic -- open source. Just like religion and politics, you just never know which side people will be on.
Of course, this hasn’t stopped me. Having tested my thesis on a number of knowledgeable friends, I can confidently state that a lot of the arguments in the open source debate are very wide off the mark, and do not do any justice to the real issues.
Open source presents a number of benefits to the IT community as a whole, particularly through the requirements of the general public licence (GPL). Kept simple, if you want to use source code published under the GPL, you must make any code you develop when using that code equally available to the general public.
It is also important to be clear on what open source licences signify, as much of the argument’s heat is generated through misunderstanding in this area. There are a number of "open source" licence types ranging from "public domain" at one extreme to Microsoft's "shared source" or the licences available from vendors such as SAP. SAP will allow the user to see some of the source code, which may be necessary to modify so as to implement specific organisational business rules. Also, not all open source licences require (or even permit) redistribution.
One side of the open source argument correctly states that GPL speeds up development lifecycles, which does not necessarily equate to innovation. To demonstrate the fragile balance between these two concepts let’s review the interesting technology race between Red Hat and Novell – both deservedly leaders in the Linux marketplace (note that I said “marketplace” not community – I am referring specifically to commercial flavours of Linux).
In February, Red Hat unveiled a new threading system that lets multiple functions occur at once on a single processor. Novell, which sells a Linux operating system called SUSE, legally grabbed Red Hat's code off the Web and had the same threading system out in a week. Red Hat can do the same with any of its rival's innovations.
The rise of Linux threatens to introduce brutal commodity economics to an industry accustomed to fat margins and customer lock-in. Any programmer can access and legitimately copy the building blocks that make the software work.
Unfortunately, those fat software margins fund R&D – Microsoft spends US$7 billion alone in this space. With much lower margins it is difficult to answer as to how much innovation will be funded. Certainly, in the infrastructure space, I believe Linux will speed up innovation for a time. In the application space, however, if a company spends a year writing code to deliver an IT solution to a significant business problem the last thing they want to do is put it out in the open market for their competitors to pick up.
Personally, I would like to see more focus on open format rather than open source in the debate. I don’t particularly care if Microsoft publishes or doesn’t publish its code for Office – but I believe whole heartedly that it should cooperate with the industry to develop standard formats so that any application potentially has access to the information stored in Office files. We are already seeing this trend in Microsoft’s .NET strategy.
This approach would oblige vendors to differentiate on true functionality and innovation, not on proprietary format lock-in. While plenty has been written about Microsoft’s faults in this area, they are by no means unique. The story equally applies to hundreds of vendors such as IBM with SNA, or even AOL with its Instant Messenger protocol.
Innovation protection, or socialism?
In these cases, the companies mentioned came up with something independent of the standards bodies. They truly innovated, and could justifiably say: "We invented this, so why should we open it up to everybody? We deserve the right to reap the rewards of our efforts."
However, protecting their innovation hasn’t stopped other vendors from reverse engineering their formats or protocols, which in turn creates an "arms race" -- Microsoft releases a new Word version with a new file format that breaks other applications’ ability to open Word docs, IBM adds a new feature to SNA that, for a while at least, only its controllers support, or AOL adds code to AIM to prevent third party IM clients from connecting. With AOL, back in 2002, Trillian -- a popular IM alternative to AIM -- was being forced to release patches daily to keep up with AOL’s changes to its software.
The argument against this sort of behaviour smacks a little of socialism, which tends to be shot down in America. It basically states that if the thing you invented has become an important de facto standard, then you should release it as a standard for the good of society.
I think, however, that you can make another argument that is based on economics and is therefore a little more palatable to the capitalist world. There are significant additional costs to production associated with the effort required to lock out your competitors, in addition to the negative experiences and market backlash from people with older versions of software trying to open incompatible new file formats. (The cynic in me says that vendors actually like this incompatibility because it drives the upgrade cycle.) When a compelling new feature is added to an application, the upgrade may cost end users $600 and have taken six months to develop. If the vendor eliminated the time and cost taken during the development lifecycle in its effort to lock out competitors, that application upgrade may have been on the shelves in only three months with a $300 price tag. This is the potential scenario if the vendor had accepted that the format is an open standard.
I find it interesting that the hardware industry (where there are some concrete marginal costs) is better than the software industry at agreeing on standards and handling commoditisation. People wouldn't accept the situation where their Dell PC can’t connect to a Cisco wireless base station. Even something as humble as home entertainment systems or the analogue phone line where de facto standards have become formal open standards. There still seem to be plenty of telecommunications companies competing to sell you a cordless phone or even more interesting phone services, such as Telstra’s 1# and 101 services, bundled broadband or SMS to home.
Standards today are lowest common denominator. Typically weak on functionality and low on innovation, standards give every vendor the perfect excuse to bend the requirements, or add an innovation as a proprietary extension but still claim standards compliance.
It’s not too hard to see how this occurs. Standards bodies are usually full of politics, dominated by vendor representatives who often make Yes Minister’s Sir Humphrey look positively philanthropic.
We need government-sponsored bodies who drive standards innovation and compliance and to act more like Standards Australia or the CSIRO (Commonwealth Scientific and Industrial Research Organisation), or, even more radically, the ACCC (Australian Competition and Consumer Commission). Perhaps governments should worry less about who the dominant technology vendors are and start making open standards and open formats count for something.
Even if you don’t care about standards and interoperability, consider the archiving perspective. Documents from hundreds of years ago can still be "accessed" because they use IOP (ink on paper) technology. In the computer age we are faced with obsolete formats all the time. Neither of my laptops can read a floppy disk. There is debate as to how long CD-Rs will last, and for how much longer software will be able to read Word Perfect, Ami Pro or Word 2.0 documents -- and these were the popular formats. At least if there is a document that describes the data format it will be possible to reconstruct software in the future to read that format, even if the original software has been lost.
Open formats, coupled with strong market pressures, will drive innovation for the benefit of the customer not the benefit of the vendor -- and this can only be good for the software industry as a whole.
Scott Petty is chief operating officer at Dimension Data Australia, a Microsoft Certified Solution Provider Partner.