UK Web Focus NOTE: Sidebars edited to demonstrate capabilities

“UK Government Promises To Go Open – Again”

In a post entitled UK Government Promises to Go Open – Yet Again Glyn Moody provides a rather cynical view based on his experiences of Government promises regarding ICT and openness: “after years of empty promises, the UK government assures us that this time is will really open up, embracing open source and openness in all its forms”. However there is also some optimism in the column:

“… there is a ray of hope. For as I reported a month ago, the Cabinet Office has settled on a rather good definition of open standards that includes the key phrase “have intellectual property made irrevocably available on a royalty free basis”, which does create a truly level playing-field that allows open source to compete fairly.”

The column concludes:

“Let’s hope it really marks the beginning of a new era of openness in UK government IT – and that I won’t have to write this article ever again.”

Publication by the Cabinet Office of the “Government ICT Strategy”

I have previously commented on the Government’s attempts at agreeing on a definition of open standards in a post entitled UK Government Survey on Open Standards: But What is an ‘Open Standard’? and pointed out some of the difficulties (is RSS an open standard, for example). But although it may be difficult to provide agreement on such definitions, I welcome the fact that the Government is asking such questions.

This is particularly important in light of the recent release of the Cabinet Office’s recent publication of the  Government ICT Strategy (PDF format). In the introduction the Right Honourable Frances Maude, Minister for the Cabinet Office lists the following challenges central government is facing:

• Departments, agencies and public bodies too rarely reuse and adapt systems which are available ‘off the shelf’ or have already been commissioned by another part of government, leading to wasteful duplication:
• systems are too rarely interoperable;
• the infrastructure is insufficiently integrated, leading to inefficiency and separation;

The first bullet point could be interpretted as a signal that the government is looking to procure off-the-shelf proprietary systems.  However the other two points seem to challenge that perception, as it is precisely such monolithic proprietary systems which fail to provide the interoperability and the integrated infrastructure which is needed.   Instead in order to address these challenges the strategy announces that it intends to:

impose compulsory open standards, starting with interoperability and security;

We know that the government is prepared to take ‘bold’ decisions – but is this a perhaps unusual decision in being one that those involved in IT development activities within the high education sector would welcome?

What are the Open Standards Which Will Be Made Compulsory?

It is also pleasing to see that the Government has invited feedback on the open standards which it feels are relevant.  A SurveyMonkey form on Open Standards in the Public Sector invites feedback on its proposed set of conditions for an open standards (discussed previously) as well as listing open standards in 23 technical areas for which respondents can specify whether they think the standards should be a PRIORITY STANDARD, MANDATORY (must be used), RECOMMENDED (should be used), OPTIONAL or SHOULD NOT USE.

The 23 areas are Accessibility and usability; Biometric data interchange; Business object documents; Computer workstations; Conferencing systems over Internet Protocol (IP); Content management, syndication and synchronization; Data integration between known parties; Data publishing; e-Commerce, purchasing and logistics; e-Health and social care; e-Learning; e-News; e-Voting; Finance; Geospatial data; Identifiers; Interconnectivity; Service registry/repository; Smart cards; Smart travel documents; Voice over Internet Protocol (VOIP); Web services and Workflow and web services.

Rather than attempting to comment on all of these areas I’ll explore some of the issues with the approaches which are being taken in the survey by addressing just two areas: “Accessibility and usability” and “Computer workstations”.

“Accessibility and Usability”

The first section covers “Accessibility and usability” and addresses Human Computer Interface standards (e.g. ISO/TS 16071:2003);  Web Content standards (WCAG 1.0) and Usabilty (sic) standards (e.g. ISO 13407:1999).

This is an area of particular interest to me, so how should I respond to the survey (which is illustrated).

The first question, on WCAG 1.0, is easy – this has been superceded by WCAG 2.0 and should no longer be used.  So that is clearly be in the “Should Not Use” category.

Should, therefore, the answer to the use of WCAG 2.0 be to select it as a Priority Standard, a Mandatory Standard or a Recommended Standard, Optional or, perhaps, Should Not Use?  These terms have been defined in the survey system:

PRIORITY STANDARD – a standard that you think is is important and a priority

MANDATORY – a standard that you judge MUST be used by the UK public sector

RECOMMEND – a standard that you judge should be used by the UK public sector but recognising that there may be exceptions/caveats that mean it is sometimes not appropriate

OPTIONAL – a standard that you judge may be used by the UK public sector

SHOULD NOT USE – a standard that you judge should not be used by the UK public sector

I have previously suggested that public sector organisations in the UK should be using the BS 8878 Code of Practice for Web Accessibility as this provides a policy framework for developing accessible Web sites and provides the flexibility in the selection of accessibility guidelines, such as WCAG 2.0 which may not be applicable for use in some circumstances.  However BS 8878 isn’t included in the list of standards.  I think that WCAG 2.0 is important, but not applicable in all cases, so I guess I should select the Priority Standard option.  In addition, since it is possible to select multiple responses, I would also choose the Recommend option.

From this first two standards I have already found reasons why the Mandatory response may be be appropriate and noticed some logical flaws in the design of the survey form – it seems it is possible to select multiple responses, including ones which may be contradictory.

The third ‘standard’ is also confusing as it covers the ‘Central Office of Information Standards and Guidelines‘.  However this isn’t a standard but a set of UK Government recommendations and policies. The guidance document contains a section on Delivering inclusive websites which appears to have been published in 2009 and which requires Government Web sites to conform with WCAG 1.0 to a AA level. This ‘standard’ is not compatible with the first two areas and so therefore the Should Not Use recommendation should be given – not because the recommendations are necessarily wrong but because it is not a standard. However it is not possible to annotate the responses submitted using the survey system.

“Computer Workstations”

The misleading “Computer workstations” section is of particular interest to me since it covers various Web standards, document, standards and standards for office applications. In the list of Web standards the choices are HTML 4.01, HTML 5 or XHTML. Here the choices are between a W3C HTML 4.01 standard which was ratified in December 1999, a W3C HTML5 working draft which has not yet been ratified and which is still evolving and a W3C standard for which a version number isn’t specified which could lead to confusions over the ratified XHTML (1.0) standard and the moribund (but recently updated) XHTML 2 working draft.

The list of document types are also interesting.  RDF RTF is listed as a standard – although this is a proprietary format which is owned by Microsoft. Similarly the inclusion of PDF from version 4 covers both the proprietary version owned by Adobe as well as the ISO standard which is based on PDF 1.7. The ODF and OOXML open standards are listed although the Microsoft Document format is also included as well as the Lotus Notes Web Access format.   There are similar confusion over the open standards for spreadsheets: HTML is suggested which, although this is an open standard, will not provide the interoperability which open standards are meant to deliver.  As with the document formats ODF and OOXML are included but the proprietary MS Excel format is also listed. This pattern is repeated for presentation formats, although this time MS PowerPoint is listed.

Other Areas

The section on “Biometric data interchange” is interesting, although I know nothing of the standards used in this area. But what are the implications of responding to the question on. for example, “ISO/IEC 19794-5 Information Technology – Biometric data interchange formats – Part 5: Face image data”. If this is a Mandatory Standard could this mean that it is used in situations which I feel infringe personal liberties? The initial response might be to suggest that the standard will only be used in appropriate areas – and yet we have seen that defining WCAG as a Mandatory standards has led to it being enforced when its use may be inappropriate. It does seem to me that there is a need to define a policy layer which helps to ensure that Mandatory clauses are not used in inappropriate areas.

I’ll not comment further here on areas which I know will be of interest to the JISC development community:

Conferencing system (six standards listed), Content management, syndication and synchronisation (which covers various standards such as XML Schemas, OAI-PMH, RSS, OpenURL and Z39.50), Data integration between known parties (which includes XML, XML Schemas, XSL, UML, RDF and OWL), Data publishing (which covers RDF, SKOS and OWL), Identifiers (which covers DOIs, ISBN, ISSN, XRIs, GUID, URIs, URLs and PURLs), Interconnectivity (which covers various Internet protocols), Service management (which only includes ISO/IEC 20000) or Service registry/repository (which includes UDDI, ebXML, ebRS and edRS), e-Learning (which covers IMS, IEEE LOM and SCORM), Geo-spatial, Web Services and Workflow and web services.

or areas which will be of less direct relevance to our development community:

Business object documents, Smart cards or Smart travel documents. e-Commerce, purchasing and logistics, e-Health and social care, e-News, e-Voting, Finance and VoIP.

Discussion

Despite the rhetoric in the introduction to the Government ICT Strategy document it seems that the survey is simply revisiting work which has been published previously in the e-GIF guidelines. Looking at the Technical Standards Catalogue, for example, there is a section on Specifications for computer workstations which lists PDF, MS Office formats and Lotus notes which I mentioned previously.

Looking in more detail at the survey form I find that the form is full of typos. For example (with the typos given in bold):

• There are many different defintions of the term ‘open standard’. We’d like your feedback on our proposed definition.
• Usabilty  (there are multiple occurrences of this typo)
• coding of continous-tone still images (there are multiple occurrences of this typo)
• Data defintion – Government Data Standards Catalogue (there are multiple occurrences of this typo)
• Ontology-based inforamtion exchange (e.g. OWL)
• Persistient identifier (e.g. XRI) (there are multiple occurrences of this typo)
• Digital Object Indentifier (DOI)    (there are multiple occurrences of this typo)
• HyperText Tranfer Protocol (HTTP)  (there are multiple occurrences of this typo)
• Authetication (there are multiple occurrences of this typo)
• Elecrtical standards (e.g. ISO/IEC 7816-10)
• Terminal infrastrucure standards (there are multiple occurrences of this typo)

Does this matter if the meaning is obvious?  For a conversational email message or blog post perhaps not but for a formal process for gathering information it is of some concern. This is particularly true when there may be particular standards which could be mis-identified be typographical errors. So although I spotted the errors listed above (initially when reading the document and subsequently by putting the document through a spell-checker) I have no idea if the following examples could contain errors:

• ISO/IEC 7816-15: 2004/Cor 1: 2004
• Contact cards – Tactile identifiers BS EN 1332-2 Identification card systems – Man-machine interface Part 2: Dimensions and location of a tactile identifier for ID-1 cards

It should also be noted that the survey form itself contain flaws. As illustrated below although the form repeatedly invites respondents to “suggest other standards within this category that are not listed. Start a new line for each in reality it is not possible to enter more than a single line.

Glyn Moody felt that there was a “ray of hope” in the Governments’s apparently enlightened approach to open standards. I fear he is mistaken – sadly I see nothing to indicate that the government has an understanding of the implications of any decisions that may be taken as a result of this flawed information-gathering exercise.

UK Government’s Open Standards Survey

I was alerted to the UK Government’s Open Standards Survey by Adam Cooper of JISC CETIS, who has already encouraged readers of his blog to participate in the survey. I’ve skimmed through the questions but haven’t yet completed the survey. What stuck me, though, was the draft definition of the term “open standard” as proposed by the UK Government.

Respondents are invited to give comments to the following five conditions:

1. Open standards are standards which result from and are maintained through an open, independent process
2. Open standards are standards which are approved by a recognised specification or standardisation organisation, for example W3C or ISO or equivalent. (N.B. The specification/standardisation must be compliant with Regulation 9 of the Public Contracts Regulations 2006. This regulation makes it clear that technical specifications/standards cannot simply be national standards but must also include/recognise European standards)
3. Open standards are standards which are thoroughly documented and publicly available at zero or low cost
4. Open standards are standards which have intellectual property made irrevocably available on a royalty free basis
5. Open standards are standards which as a whole can be implemented and shared under different development approaches and on a number of platforms

I think the survey was wise to begin by being honest about the difficulties in defining an ‘open standard’ and inviting feedback on its proposed set of conditions. The survey follows on from work which has been carried out by UKOLN, JISC CETIS and JISC OSS Watch with our shared interests in helping the sector to exploit the potential of open standards. I thought it would be useful to revisit our work before I completed the survey.

Previous Work in Describing an ‘Open Standard’

The term “open standard” is somewhat ambiguous and open to different interpretations. In a paper entitled “Openness in Higher Education: Open Source, Open Standards, Open Access” (available in PDF, MS Word and HTML formats) Scott Wilson (CETIS), Randy Metcalfe (at the time at JISC OSS Watch) and myself pointed out that:

There are many complex issues involved when selecting and encouraging use of open standards. Firstly there are disagreements over the definition of open standards. For example Java, Flash and PDF are considered by some to be open standards, although they are, in fact, owned by Sun, Macromedia and Adobe, respectively, who, despite documenting the formats and perhaps having open processes for the evolution of the formats, still have the rights to change the licence conditions governing their use (perhaps due to changes in the business environment, company takeovers, etc.)

It should be added that this paper was written in 2007. Since then PDF has become an ISO standard so we could add the fact that proprietary formats can become standardised to the complexities.

In a UKOLN QA Focus briefing paper we tried to describe characteristics shared by open standards, which had similarities to the approaches proposed in the UK Government survey:

• An open standards-making process
• Documentation freely available on the Web
• Use of the standard is uninhibited by licencing or patenting issues
• Standard ratified by recognised standards body

It should be noted that we described these as ‘characteristics‘ of an open standard rather than mandatory requirements since we were aware that the second point, for example, would rule out standards produced by many standardisation bodies such as BSI and ISO.

Responding to the Survey

I’d like to share my thoughts prior to completing the survey.

1. Open standards are standards which result from and are maintained through an open, independent process

I would support this condition. It should be noted that this means that a standard which is owned by a vendor cannot be regarded as an open standard even if the standard is published. This means, for example that Microsoft’s RTF format is not an open standard and PDF was not an open standard until ownership was transferred to ISO in 2008. It should be noted that I believe that the US definition of ‘open standards’ does not include such a clause (there were disagreements on this blog over the status of PDF before it became an ISO standard).

2. Open standards are standards which are approved by a recognised specification or standardisation organisation, for example W3C or ISO or equivalent. (N.B. The specification/standardisation must be compliant with Regulation 9 of the Public Contracts Regulations 2006. This regulation makes it clear that technical specifications/standards cannot simply be national standards but must also include/recognise European standard).

I used to have this view. However I can recall an email discussion with Paul Miller and Andy Powell when they worked at UKOLN who argued (and convinced me) that this was an over-simplistic binary division of the world of standards. It should be noted that RSS (in any of its flavours) would not satisfy this condition. The question, then, is whether this is a concern? If the definition of an ‘open standard’ will be used to determine whether a standard should be used by the UK Government then there will be a need to avoid being too rigourous in the definition. My view would be to rule out this condition.

3. Open standards are standards which are thoroughly documented and publicly available at zero or low cost

I would agree on the importance for rigourous documentation for open standards, so that ambiguities and inconsistencies are avoided. This clause is, however, ambiguous itself – what is ‘low cost’ documentation? However I would be happy to see this condition included.

4. Open standards are standards which have intellectual property made irrevocably available on a royalty free basis

This is desirable, but what happens if it is not possible to negotiate royalty-free licences? This is particularly true for video formats. If the government uses this as a mandatory condition for open standards and subsequently requires services to make use of open standards might this result in a poorer quality environment for the end user? From an ideological position I would like to support this condition but in reality I feel that there needs to be more flexibility – there is a danger that if open standards are mandated this could mean that Government departments would be barred from making use of popular services – such as YouTube and iTunes – which many people fund helpful in gaining simple access to information of interest. I am therefore rather uncertain as to whether this should be a required condition for the definition of an open standard. It is worth noting, incidentally, that the W3C have similarly avoided grasping this particular nettle in the HTML5 standardisation work, with no specific video codex being mandated as part of the standard.

5. Open standards are standards which as a whole can be implemented and shared under different development approaches and on a number of platforms

This has always been a view I have held.

The contentious issues seems to be “Open standards are standards which have intellectual property made irrevocably available on a royalty free basis“. I suspect people will argue strongly that this condition is essential. For me, though, we are revisiting Martin Weller’s “Cato versus Cicero” argument. Should we be taking a hardline stance in order to achieve a desired goal or do we need to make compromises in order to accommodate complexities and the conflicting needs of various stakeholders?