Thursday, December 19, 2013

A question for zero hour contracts

By zero hour contract, I don't mean you turn up for work, do basically nothing and go home with £300 as disgraced Lord Hanningfield claims is quite common in the House of Lords.  Zero hour contracts are ones where the employer has no obligation to give you work i.e. you can turn up for work, get told there and then whether you're needed and can be sent home without pay for the day.

Some groups such as the Chartered Institute of Personnel and Development have claimed that Zero hour contract workers are happy whilst others such as the TUC demand tougher regulation of these contracts. To quote the TUC

"The growth of zero hours contracts is one of the reasons why so many hard-working people are fearful for their jobs and struggling to make ends meet, in spite of the recovery"

Into this mix Vince Cable acknowledge the clear signs of abuse but decided not to ban them on the grounds they offer 'welcome flexibility'.

So are they good or bad? Well, in my view it's a form of exploitation which should not be welcomed but that's not the point of this post. Whenever you hear someone talking positively about zero hour contracts or providing surveys and market research proclaiming the benefit then the one question you should ask is :-

"If zero hour contracts are so good, I assume you're on one?"

If they're not, then show them the hand and walk away.

I, would of course welcome the use of zero hour contracts in the House of Lords. We could take a quick online vote as to whether we needed them today or not. Clearly, if Hanningfield's claims turn out to be true then many are just turning up to claim expenses.

I'll finally note that Hanningfield's defence of his actions were "I have to live, don't I?"  - well, there's always income support and housing benefit which is what the rest of the country uses. If you need the money Hanningfield then get on your bike and get a job. From the sounds of it there's a lot of highly flexible and 'happy worker' zero hour contracts out there.

Cloud Standards and Governments

As any activity evolves from product (+rental) to more commodity (+ utility services) due to the forces of competition then a number of factors come into play. There is a period of re-organisation as a new market emerges, a co-evolution of practice due to the changing characteristics of the evolving activity along with inertia caused by past practices and past means of operation. We see this with abundance in the cloud market today. 

Under normal circumstances a set of defacto standards would appear in the market and a period of consolidation then occurs.  Ideally, one common standard emerges but in some cases we get a limited set of alternatives. As the market continues to develop then multiple implementations of the defacto standards tend to form and a competitive market is created. 

It doesn't always happen like this, sometimes competition is restricted by the development of a monopoly / oligopoly. In such circumstances when user needs are not met or competition is restricted in a mature market then it may become necessary for Governments to force a set of standards onto that market in order to create a level playing field e.g. document formats.

There are also variations in the type of market that is formed. For a free competitive market (as opposed to a constrained or captured market) then the means of implementing the standards must be open to all. This is why the ability to reverse engineer an implementation is often critical to forming a competitive market. In the most ideal form the standard would be implemented in an open source reference model.

IaaS however is not a mature market but an emerging market for which we already have a defacto standard being formed in the AWS APIs. The rapid growth of AWS is testament to its meeting of user needs but nevertheless for reasons of buyer / supplier relationship and second sourcing options then a competitive market of many providers is desirable.  Fortunately, for the core parts of the AWS APIs (there is a long tail of functionality) there are already multiple open source implementations from Eucalyptus to CloudStack to Open Nebula and even parts of OpenStack. APIs themselves are also principles in US / EU law despite the best efforts of Oracle. Hence we have the basic pieces that are necessary to form a competitive market of AWS clones and possibly also GCE clones.

For an emerging market this is a reasonable position to be in at this stage. If anything Governments should push forward to consolidating the market by reaffirming APIs as principles and even limiting any disguised software patents that may limit re-use. This would all be positive.

However, vendors have inertia to change particularly when that change threatens their business.  So today, I increasingly hear calls from vendors for 'Open Standards' in IaaS on the grounds of meeting user needs for interoperability.  However,  what I normally hear from users is they'd like a number of alternative providers of the AWS APIs i.e. homogeneity in the interface but heterogeneous providers. There is a conflict here because the 'Open Standards' proposed are not AWS APIs whilst users rarely say they need another set of APIs other than AWS.  However, this difference in needs isn't going to stop vendors trying to pull a fast one by trying to ban competitor's interfaces through Government imposed 'Open Standards'. 

So when your local vendor comes and starts talking to you about the importance of 'Open Standards' ask yourself some basic questions ...

1) Is this an emerging market?
2) Is there a defacto standard developing in the market?
3) Is there one or more open source reference implementations of the defacto?

If the answer to all three is yes (as is the case with IaaS) then there is no need to force a standard onto the market and everything seems to be healthy. A good response to the vendor would be to tell them to come back with a service offering that provides the defacto.

If the vendor insists that you need their 'open standard', then ask yourself ...

1) Is the vendor 'open standard' different from the defacto?
2) Does the vendor's products / services offer the 'open standard'?
3) Is the vendor trying to create an advantage for themselves regardless of actual user needs because they cannot compete effectively in the market without this?

In all likelihood, the call for 'open standards' in an emerging market is just an old vendor play of 'if you can't beat them, try and ban them'. 

Caveat Emptor.

Tuesday, December 17, 2013

On Politics, IT and Strawmen

Shadow Minister Chinyelu Susan 'Chi' Onwurah describes how project management in Gov IT has been politicised. The article states a coalition portrayal of 'Waterfall as monolithic Labour' and 'Agile as dynamic, entrepreneurial Tory' and how to 'think that the solution to effective ICT deployment is to simply change your project management methodology is arrogant and naive'. It then goes on to describes how Labour will undertake this 'transformation with more humility' requiring 'diverse methodologies' and learning from 'the process itself'. 

This all sounds very reasonable but there is a problem with this portrayal. The UK GOV IT approach is already on the path to becoming more balanced. 

With any course correction, an over steer is a natural part of shifting an organisation away from one all encompassing method to a more balanced approach. I already see signs of the shift to a balanced approach happening which is remarkable in such a short period of time.  For example, whilst HS2 might be seen as controversial in some circles, it is showing signs of strategic understanding and IT leadership which I've rarely seen rivalled in the private sector. Rather than being 'one size fits all' shop with little of poor understanding of why action is taken, the CIO has been using a mapping technique to break down large scale projects and determine what methods are appropriate. Its future appears neither as an Agile / XP / Scrum Development nor a Six Sigma / Prince / Waterfall shop. Its future appears more balanced, more the use of the right methods where appropriate.

Unfortunately this doesn't help the argument that Labour will make a difference because it's difficult to claim a difference (a more balanced view) when this is already happening. The portrayal proposed in the article is unfortunately a necessary instrument of what we know as a strawman argument. If you were going to have two 'extremes' of past and present then in my opinion they would be :-

Under previous administrations, UK Gov IT could be characterised by an over reliance on large outsourcing arrangements, inefficiency in contracts, poor data and focus on user needs, poor development on internal engineering skill, extensive use of single and highly structured methods of project management, lots of discussion on open source / transparency / open standards but little execution.

Under the current administration, UK Gov IT can be characterised by a movement towards both insourcing where appropriate and use of SMEs, a greater challenge in contracts, a focus on user needs and collection of transaction data, co-operation with other governments, a focus on recruitment of engineering talent, use of more appropriate management methods from scheduling systems such as Kanban to a plurality of development techniques, a focus and action on open source, open data, transparency and open standards.

The change from these two states is not due to some political ideology but a desire of many people (both internal and external) to make things better and to reduce some of the excesses. Onwurah claims pointedly that Mark Thompson who responded to the Shadow Minister's speech also wrote the Conservative Technology Manifesto - something which I'm not aware of and a bit surprised about. I happened to work with Mark on a 'Better for Less' paper which had some input into the changes and that paper was driven solely by a focus on this principle of making things better.

For the record, until the worst excesses of spin of the Blair years, I had always voted Labour. Due to my disillusionment with Labour, I have subsequently voted Liberal Democrats. I have never once in my life voted Conservative. I don't personally agree with many of their policies. But then I wasn't asked to be involved in writing the paper because of my political beliefs but because the group felt I had something to contribute.  Politics never came into it.

But neither manifestos nor politics really made this change possible. As Mark pointed out 'the important point is that one would be hard pressed to find a single outcome in Labour’s Digital Britain to which those in the Government Digital Service (GDS), or their digital leaders network, are not already deeply committed'. The real change, something familiar to all of us that have worked in the open source world, was strong leadership and the UK GOV IT appears to have had this with Francis Maude.

I do fear that the worst excesses of sharp suited management consultants are itching to come back with the panacea of  “big IT” to achieve joined-up government. In my view, we would be wise to keep off that path and continue to create the more balanced future that seems to have already started. 

I have had concerns in the recent past that Labour has yet to show the example of strong IT leadership that will be needed going forward.  If it was up to me then I'd keep Francis Maude moving the changes forward regardless of political persuasion.  Maybe Shadow Minister Chi Onwurah can provide this future leadership - I simply do not know.  What I do know, from a competition viewpoint, is IT is more important to our future industry than petty debates and if this is what the future leadership might mean - politicising the debate and creating strawman arguments - then my concern grows. 

I hope I'm misunderstanding what is happening here but the politicisation of project management methods helps no-one.

Friday, December 06, 2013

Once more unto the breach, dear friends, once more ...

If you've been paying attention then you'll have noticed that the UK Government has had a fairly torrid time of trying to implement a policy for open standards.

It eventually won that fight despite efforts from numerous groups such as the BSA, large US companies apparently attempting to influence the processquestions raised over the role of one of the chairman  and software patent heavyweights piling into public meetings. At the end of several rounds of shenanigans and lobbying, the UK Gov was finally able to announce a comprehensive policy towards open standards. It was finally able to do what the only respected peer reviewed academic paper on the subject said was a good thing to do - adopt the policy.

However, this fight was just to get the policy approved. We should of course expect many more fights when it comes to specific open standards. Why? Well, it has all to do with money or more importantly the ability of certain companies to help themselves to large amounts of your and my money. 

To explain why, let us first understand why we need standards. One of the fundamental purposes of standards is to provide you with choice. If two systems use the same standards then you can switch between both and this switching encourages competition. The reason why UK Gov has a preference for open standards is that it also has a preference for a fair and competitive marketplaces i.e. if the standard is owned by someone or encumbered with patents then that person can limit competition in the market. 

Shouldn't we just therefore make everything into a standard? No. There is a cost with standardisation in that it limits innovation at the interface of the standard, hence the implementation of standards is only suitable for activities that have become commonplace and mature. 

But surely the marketplace as it matures will define its own standards? Yes. These are called the defacto standards and in most cases they should be adopted. However, in a mature market where the defacto doesn't meet some user need or encourage competition then Governments might need to introduce standardisation to create a competitive market.

Can you give me examples? Certainly. Two examples spring to mind.

1) In the case of cloud computing you have an evolving market where defacto standards are developing. As an evolving market it is too soon for any Government to attempt to implement standards. 

2) In the case of documents and word processing systems (like Open Office, Libre Office and Microsoft word) then you have a highly mature and established market. However, in this case despite there being a defacto there is not easy switching between a competitive market of products and this is most clearly shown when people talk of the cost of transfer from the defacto to another. 

So why should I care as a vendor? If you are the defacto provider for a mature marketplace and there is a significant cost for people to transfer off your system then you've in effect created a captured market and this can be very lucrative. You won't experience the normal competitive pressures of a free market, you can charge high prices knowing that most won't wish to incur a cost of transfer. This actually creates an incentive for you to lock people further into your system and you'll try and stop anything which threatens this such as open standards. 

So why should I care as a customer? Well, if it's a mature market with a defacto and there are limited alternatives unless you incur a high cost of transfer then you're likely to be paying well over the odds. What's operating is not a free market but a captured one and your exit cost from the existing defacto is simply a long term liability which will keep on increasing. In such cases, you should be in favour of open standards because that's the only way you'll get a fair and competitive (known as 'free') market and limit future liabilities.

So that gives some basic background on open standards. The real question is why do I bother mentioning this now? Well the UK Government has just launched a process to select open standards for government documents. I fully expect the lobbying machine to be out in full on this one as what's at stake is oodles of cash and control of a market.

Microsoft Office is obviously the defacto in a well established and mature market which has little differential between product offerings but also high costs of transition. The market has all the appearance of a captured market with weak competition. 

However, doesn't Microsoft Office 2010 uses .docx which is Microsoft Office Open XML Format and therefore an open standard? Alas no. 

Microsoft Office 2010 provides a number of formats (docx, pptx, xlsx) that collectively are called Office Open XML format (OOXML or OpenXML). This file format was submitted to ISO after intensive lobbying including accusations of rigging and it known as the OpenXML standard (ISO/IEC 29500 adopted in 2007). 

Alas, this standard was broken into two parts. There is the standard itself, which is known as strict OpenXML and was accepted by ISO (international standards organisation) and then there is transitional OpenXML which was supposed to be a transitional file format for Microsoft to ease the removal of some of the past closed source legacy from their file format. Of course, Microsoft Office 2010 implemented transitional OpenXML but only reading of strict OpenXML.

So when you use Microsoft Office 2010 to send a document in an OpenXML format such as .docx to someone who uses Libre Office which reads and writes strict OpenXML (as defined by ISO) and they edit the document and send you back a .docx file, then what you see is often a corrupted file or something which can only be read and has to be converted.

You probably think that something is wrong with their software. Well, this problem isn't because of their word processor but instead Microsoft Office 2010 didn't implement strict OpenXML except for reading.  It certainly claims that .docx is OpenXML, and that OpenXML is an open standard approved by ISO but the version of OpenXML that ISO approved (strict) is not the same version of OpenXML that Microsoft Office 2010 .docx uses (transitional). This means every other system has to reverse engineer the transitional OpenXML format that the default .docx uses and that's never perfect.

So let me be crystal clear, Microsoft Office 2010 didn't even fully implement the open standard it created and was approved by ISO after its own intensive lobbying. It did promise to provide reading / writing of this format in the future.

Hang on that's old news .... doesn't Microsoft Office 2013 now implement reading / writing of strict OpenXML? Well, the latest release of Microsoft Office does now implement strict OpenXML, unfortunately the default .docx is still transitional OpenXML, you have to specifically select strict OpenXML when you save the file (which is buried in the options) and of course, in order to use Microsoft Office 2013 you need to be running Windows 7 or Windows 8.

So basically, yes you can have the open ISO standard strict OpenXML if you upgrade your operating system, buy the latest version of Microsoft Office 2013 and remember to save in the right format. It must be noted that whilst Microsoft Office 2013 does now provide support for strict OpenXML and even includes group policies for this, the default is .docx which is using Microsoft's own version of an open standard (transitional OpenXML) but without actually being that open standard (strict OpenXML). It's all a bit messy and confusing.

So what does this mean in practice? Well, let us assume that the UK Gov chooses the ISO approved strict OpenXML as an open document standard. First, that means Gov departments would need to upgrade to Microsoft Office 2013 which means upgrading to Windows 7 or 8 and all the changes needed to applications (NB. Microsoft XP is out of support in April 2014). They'd also have to set group policies so that the default was strict OpenXML.

However, let us suppose you create a document, a .docx file and sent it to someone else who happens to be using Microsoft Office 2010. Unfortunately though they could read the .docx file (Microsoft Office 2010 provides read support for strict OpenXML) they couldn't write / edit such a document without changing it to another type of .docx (the default transitional OpenXML). Hence you might send out strict OpenXML but you could easily get back transitional OpenXML though both files are called .docx.

Now given that some companies / organisation are still just in the midst of rolling out Microsoft Office 2010  then you're going to have all sorts of problems trying to introduce use of strict OpenXML as the .docx format and in practice the default transitional OpenXML format will rule for many many years to come. At best, you're going to end up with a messy estate of .docx files some of which are strict format and some of which are transitional format but all are called .docx.

Fortunately if you're using Microsoft Office 2013 this is ok and the more messy the environment, the more incentive there is to all upgrade (including the Operating System). Alas, there are only a limited number of alternative companies providing word processing capable of reading / writing Microsoft Office .docx (transitional format) as it currently stands because of the effort needed to reverse engineer transitional OpenXML (remember strict is the standard).

For example Libre Office does provide the ability to save files as two different .docx formats - one for  "Microsoft Word 2007/2010 XML (.docx)" which is the reverse engineered transitional OpenXML format and one for  "Office Open XML (.docx)" which is the strict OpenXML format that is the ISO standard.

The upshot of this, for many it'll just seem easier to stick with Microsoft, upgrade the OS and Office package, accept a messy estate and in all probability stick with transitional OpenXML as the .docx format. So much for a competitive market and open standards then.

However, there is an alternative - ODF.

So what is ODF? It stands for Open Document Format, it covers a number of office formats (.odt, .ods, .odp) and is an ISO/IEC standard 26300 adopted in 2006. It is supported by multiple technology providers including AbiWord, Adobe Buzzword, OpenOffice, Atlantis Word Processor, Calligra Suite, Corel WordPerfect Office X6, Evince, Google Drive, Gnumeric, IBM Lotus Symphony, Inkscape exports, KOffice, LibreOffice, Microsoft Office 2010 onwards, NeoOffice, Okular, OpenOffice, SoftMaker Office and Zoho Office Suite.

Microsoft also provides support for ODF however as it says "Formatting might be lost when users save and open .odt files". Brilliant.

One of the big advantages of word processing / spreadsheet and presentations systems like Libre Office is they have close to feature parity with Microsoft Office but they also aren't tied to a specific operating system i.e. you can get Libre Office for Ubuntu, MacOSx and Windows.

So why doesn't Microsoft just adopt ODF? In a nutshell, control of a market. Why simply give up a captured market if you don't have to and especially if you can persuade people that the exit costs that your product has created aren't actually an ongoing liability. If you can use those exit costs to persuade people against moving, explain that your product really is 'open' with OpenXML, add in a complex mess of strict and transitional formats then you can hopefully can get people to stay put and just upgrade both Office and the operating system. You get to keep the captured market intact!

Oh, and by the way, those exit costs are significant. Microsoft's own people estimated that adoption by UK Government of ODF as the standard would cost in excess of £500 million. What that means is we're currently locked into an environment which it will cost £500 million to escape from (assuming the figure is correct) but what they fail to mention is that 'liability' is unlikely to decrease over time. It's that 'liability' which keeps us paying for new versions of Office, new versions of OS, application upgrades  and also prevents a truly competitive market for what is fundamentally a common and well defined activity - writing documents etc. 

This is where Government really has to step in because it has power and influence in the market.

This is clearly a mature but captured market which the introduction of an open standard will encourage greater competition, reduce long term liabilities such as exit costs and therefore benefit users. There is no reason why Microsoft can't upgrade previous versions of Office to write strict OpenXML (they've had six years) and change the default .docx to this, however it's probably just not in their interests to do so. The pragmatic choice for Government would therefore seem to be ODF.

If you select ODF then there is also no reason why Microsoft cannot sell office products based upon ODF (which it supports) and if not Microsoft then there's lots of other potential vendors (see list above) including open source solutions who will. Since Portugal and other Governments have already gone down a route of establishing ODF as the standard then it also doesn't make much sense for the UK Gov to create its own island of technology standards. 

But then again, standards are rarely about pragmatism and more about vested interests. Now this is the reason why there will be a fight because you're talking of an established defacto with a captured market potentially being forced into giving up that control and competing in a free market. I expect to see lots of intense lobbying over why OpenXML should be the standard despite the default .docx of Microsoft Office 2010 and 2013 being transitional OpenXML and not the approved ISO open standard strict OpenXML.

At the end of the day this is about word processing, spreadsheets and presentations. We don't need umpteen file formats for this and certainly not two different formats for .docx such as strict and transitional. This will be claimed to be a necessity for reasons of transition but we've had six years and no end in sight to this mess. In reality this is all about prolonging control of a captured market and almost nothing about end user needs.

This is why Governments must act but of course, it'll need you to help counter the well funded  lobbyists that are likely to appear - just like last time.

Before I go, I thought I'd better clear one last thing up. I happen to love Microsoft Office. I use it all the time on MacOSX in particular Excel. However, I'd rather see Microsoft compete on providing me with a better experience / product which was based on ODF (an open standard) in a competitive market than what is happening today. I do understand the temptation to claim something is an open standard when it is not, I do understand the gameplay and how standards can be used to control the market.

I hope, maybe with a change of leadership that Microsoft learns it can compete purely on having better products and not by attempting to create and exploit lock-in for what is a common activity. I fear that this standards debate will just become another example of Microsoft rallying troops into the fray with the usual cries of being "open", concerns over exit costs that ignore long term liabilities and funded "white papers" masquerading as academic work etc.

Hence in the next post, I'll talk about some of the common tactics we're likely to see.