My blog has moved!

You should be automatically redirected in 4 seconds. If not, visit
http://chrisjhorn.wordpress.com
and update your bookmarks.

Saturday 10 November 2007

Open Source, China and Microsoft

Confucius was born in the state of Lu. When he received news that the powerful state of Qi was preparing to attack his homeland, he sent his gifted disciple Zi Gong to talk to the rulers of the surrounding states. Zi Gong went first to the state of Qi and observed to the military generals the flaws in a strategy to attack Lu. He succeeded in persuading the generals to first attack the state of Wu instead. Zi Gong subsequently went to Wu and instigated the king of Wu to attack Qi…Thus, Confucius saved Lu.

From ninth chapter of Chang Duan Jing.


At first sight, it would appear that there should be an excellent cultural fit between the open source movement in the West, and Chinese values. Open source emphasizes collective knowledge and sharing of competence. In China, the loyalty to the group is strong; communist philosophy emphasizes sharing, and Confucian teaching emphasizes the latent potential of the individual to attain skilled judgment from the experience of others.

The use of the web is growing fast in China – exceeding the US and growing much faster than the US - as reported in Forbes. Open source collaboration uses the web as a collaboration platform, so you would also expect this to add to the momentum of open source in China.

The Economic Intelligence Unit recently reported a ratio of 100 jobs for every computer science graduate in China, with this number expected to sky-rocket; and Duke University reported 60,000 computer science graduates from 4-year degree programmes in China in 2004, and 292,000 from 3-year programmes. This huge domestic demand, and huge output, of software developers might lead you to expect further momentum for open source in China.

A recent Eclipse Members meeting noted that China has the most number of the downloads globally (over a recent 18 month period) at 21%, followed by the US at 18%, and both Germany and Japan at 8% each.

So: what is the status of open source in China ?

Well, there are a small number of open source projects in China, but apparently not as many as you might expect. XOOPS (a content management framework) is quite well known and Stephen Walli’s blog contains an interesting presentation by Tiaiwen Jiang, the community leader in China, on the project from the Chinese perspective. Huihoo is a leading open source middleware project including a J2EE implementation, JFox. Qianqian at Harvard Medical School initiated, in 2004, a collaborative project Wen Quan Yi to develop an open source font set for the 70,000 Han characters encoded by Unicode.

An interesting development is the merger of ObjectWeb in France and Orientware in China at the end of 2006, to form OW2. They are sharing open source contributions in a variety of middleware technologies and their deliberations are documented in their Board of Directors minutes...

But, perhaps predictably, the main interest in open source in China is Linux: just google ‘china open source’ and you’ll see! In 2003, China enacted the Software Government Procurement Regulation (SGPR) which excluded foreign companies from the federal software market. As recently as the end of 2005, CIO Magazine was discussing China’s federal commitment to Linux – for multiple reasons at the time, including suspicions of US intelligence agencies “trojan horses” in US proprietary code (remember the B-767 government jet delivered in 2002 to Beijing, but with eavesdropping devices discovered on delivery ?); overcoming WTO IP concerns by promoting open source; localization to the Chinese market; and kick-starting a strong domestic software industry.

But then things changed in 2005, despite CIO Magazine’s analysis above: in trade talks, China laid aside the SGPR in favour of concessions on industries such as textile and colour television.

So now, according to Lou Shouqun of the China OSS Promotion Unit (a non-government organisation) in a recent presentation, the Linux revenues in all China last year (2006) were about 218M RMB (20MЄ), with a market share (by revenue) of just 3%. Other UNIX systems were 52% - the financial services and telecommunications industries in China have heavily used Solaris, AIX and HP/UX, amongst others. Windows was 42%. While the overall market is growing about 10% per annum, Shouqun believes Linux in China is growing faster than the market, albeit from a low base.

Things have changed even more significantly since the 2005 abandonment of the SGPR directive: Microsoft seem to have successfully found favour with the federal authorities. Fortune magazine documented Bill Gates recent summer visit to China; the history of Microsoft in China, and how Microsoft has very successfully wooed the Chinese policy makers creating an apparent “win-win” situation. It is a fascinating article, and I recommend it to you if you haven’t already read it.

So, why are there few committers in China, and apparently meek participation in the global open source, particularly when the number of software professionals in China is rising so rapidly ? Why is it that Windows is far more successful in China than Linux, and is Microsoft’s new strategy truly a “win-win” ?

The urgency to make money is IMHO a national obsession in China. Consumerism, and chasing Western fashions and brands are all-consuming. You must remember that, within the life times of those of us in our 40s or more, very many Chinese were incredibly impoverished to appallingly abysmal standards of life. It really is only in the last twenty years or so that national living standards have consistently dramatically improved – while admitting of course that there remain many challenges and disparities today across the huge country. In my own experience, making money is far more important to many Chinese than political discourse.

If your parents and grandparents have supported you, a single child, through your professional education as a software developer, their expectations (and needs in their senior age) will be that you will support them. Your partner’s parents and grandparents will have similar expectations. An engineer, including a software engineer, is considered a respected professional: many policy makers, senior business managers, and senior party members also have engineering backgrounds. As a software engineer, your own, your family’s, and society’s expectations are all that you will be financially successful.

Can you be truly financially successful in China if you are an open source developer ?

In the West, much of the open source activity in fact is carried out within companies, including Intel, Novell, IBM, Sun and Oracle, amongst others. Foreign companies in China with recognized global brands – such as Microsoft, with its very wealthy founder – are highly attractive as employers, since they not only in general pay well by local standards but also potentially open the possibility of international travel.

But, which foreign companies have so far established software development laboratories in China which contribute to open source development ? IBM have a Linux technology centre in Beijing. Intel announced in 2004 development centres in Beijing, Xi’an, and Guandong to help Chinese companies develop desktop applications for Linux. Oracle promotes its products on Linux in China, including via the Oracle technology centre in Beijing. But these centers appear, on the surface, to be solution centers which promote Linux based application solutions perhaps as a response to the apparent promotion of Linux by the Chinese authorities and prior to the SGPR retraction; rather than development centers actively contributing to globally available open source. I of course am very open to correction, but it would appear that to date only Novell has opened an R&D centre in China specifically for Linux system development. It is also noteworthy that IONA (of which I’m Vice-Chairman) has Chinese committers from its Beijing R&D centre on the Eclipse STP and Apache CXF open source projects.

For domestic Chinese activity, the largest player in open source Linux is Red Flag. It is ambitious to develop into an international player in Asian Linux, and has recently announced a specific initiative. However its current sales revenue is still relatively small, even by Chinese national standards, at just 40M RMB (3.6MЄ).

To the extent that open source development is being conducted in China by foreign companies, with Chinese committers, then arguably these initiatives need to be more openly promoted and publicized to the Chinese software development community. The open source movement in China needs major foreign brand name companies to visibly invest and recruit in China for development of open source.

Well, then how about Chinese open source start-ups ? Are there any budding MySQL ABs, SugarCRMs, or xTuples ? Yes, and I mentioned Huihoo and Red Flag earlier, as examples. But IMHO the open source industry in China is currently fragile and considered so both by potential employees (ie software developers) and, as importantly, Chinese corporate customers. It is frankly easier – and perhaps more socially acceptable with one’s parents – to work for an established organization, particularly if it is a foreign brand.


Let’s now look at that second question I posed above: why is Microsoft now being much more successful in China ? Microsoft seems to have overcome concerns by Chinese policy makers by pro-actively taking a number of steps. The “trojan horse” threat has been overcome by allowing access (and hence inspection) under appropriate conditions to Microsoft source code, and China now has a federal laboratory to do exactly so. Microsoft has been very actively investing in the education sector, including rural classrooms and software engineering universities, and so aligning its investments with the federal desire to strengthen software skills nationwide – Microsoft is training 1,000 instructors and 20,000 software engineers, and offering online courses to another 50,000 engineers. It has worked with the federal authorities in the context of WTO obligations to ensure that more Chinese PCs have legally licensed pre-installed copies of Windows. In turn this is overcoming piracy issues since the pre-installed versions are more current, have less bugs and more features than older copies of Windows. It has also dramatically dropped the price of Windows in China. In summary, President Hu Jintao on a visit to Microsoft said that Bill Gates is a friend to China and the Chinese people: in China, this is an incredible endorsement, and it is difficult to find any analogy in the West for such an important and powerful public ratification.

It would appear that Microsoft is on a roll in China. It is very interesting to reflect on the Microsoft strategy and their execution of it within China, and contrast that to the open source industry and initiatives in China. Of which open source project or company will the President of China publicly endorse as a friend of China and of the Chinese people ? If the open source industry globally is to benefit from China’s rapid development, it is clear that investment – perhaps akin to Microsoft’s commitment to China – will be needed.

It is also very interesting to ponder to what extent the prices Windows customers in the West are paying are being used to subsidise Windows customers in the East. Cross-subsidies are of course not at all a new idea in any industry: I find the Microsoft case interesting because it is part of a much broader initiative.


Let me finish by a hypothesis for policy from the Chinese Government perspective. The war is not about whether Windows or Linux will ultimately win. Both are sufficiently low cost here in China to be usable. Microsoft is generously up-skilling the national software engineering talent pool, and the open source industry is also helping by publishing its source code and inner workings. The war is rather about building a vibrant software industry in China, capable not only of satisfying national needs, but also exporting and becoming a world leader.

The main requirement is excellent application development, on whatever foundation systems and middleware technology are de facto in the global industry (it doesn’t matter which, as long as they are low cost in China). Open source by the Chinese industry – and for the Chinese industry - could play a very significant role indeed. By fostering a national repository of re-useable Chinese application components – with documentation and test suites – written by Chinese developers (in Chinese first, and then maybe English), with a framework put in place and re-enforced by Government policy and investment, then the national industry could be rapidly enhanced.


“Therefore the Master concerns himself with the depths and not the surface, with the fruit and not the flower….When his work is done, the people say ‘Amazing: we did it, all by ourselves!’’”

The Tao Te Ching, by Lao Tzu


This was the basis for an invited keynote I was to give at the OS Summit in Hong Kong at the end of this month: but the conference has now been postponed until sometime in 2008.

Monday 5 November 2007

Think Liquidity.

Professional investors understand liquidity. They understand asset backed securities, and they understand the risks when assets subsequently emerge to be poorer quality than they were represented to be. Sub-prime assets can be embarrassingly illiquid and career changing.

In the world of enterprise IT investments, customers have likewise yearned for liquidity. Lock-in to assets available solely from any single vendor implies significant risk to the purchaser. Bad investment decisions into IT assets which subsequently emerge to be poorer quality than they were represented to be, can be embarrassing and career changing – particularly if the assets are illiquid and difficult to replace.

Financial markets have been driven by liquidity. However by contrast, purchasers of IT assets have found it challenging to be able to subsequently replace and substitute alternatives when desirable.

Until now.

Today, I believe that the software and hardware industries are fundamentally changing in favour of liquidity. Software is increasingly componentized. Software has increasingly recognized industry standards which facilitates substitution of alternatives. Open source provides liquidity through lowering the cost of change.

Enterprise software vendors should be trusted partners in providing and maintaining tailored solutions. The ability to successfully integrate a variety of components from a variety of sources to an enterprise level of service is valued. The ability to scalably manage multiple configurations, and evolve them dynamically over time, is valued. Tailorable, personalized solutions for specific customers, partners and staff, but all as part as of the holistic enterprise, are valued. Dynamic systems enable liquidity amongst software assets – no matter from which particular vendors specific assets are obtained.

New, and sustainable, business models are emerging from software vendors who deeply understand technology liquidity.

Single, monolithic, vertically integrated silos of software stacks is thinking from the last century. Integrated stacks are illiquid if any specific components or layers cannot be readily substituted by better alternatives on the market today, or which may emerge tomorrow, from any vendor.

It does not take an oracle to foresee what will happen if BEAS are purchased by ORCL. Overlapping products – portal servers, application servers, service busses, Java development platforms, whatever – will be culled: “synergies” throughout the two organizations will be executed. ORCL will attempt to cross sell its own offerings into the BEAS client base and migrate them away to ORCL alternatives.

Professional investors understand liquidity. Even if they are new to investing in IT equities, they therefore should have little difficulty in understanding that enterprise IT customers likewise yearn for liquidity of technology assets. In the past, enterprise vendors have been slow to offer liquidity. Now, the IT industry is changing fast, and the potential upside for investors in IT is vendors who understand and are executing on technology liquidity: vertically integrated illiquid stacks are from a former and sub-prime era.

Think liquidity.