Scotland Forum: Do we need to get Stalinist about standards?

The latest Home Nations Forum in Edinburgh focused on Standards. Alastair Dunning from JISC Digitisation gave a talk which illustrated how standards need to be thought of as existing in an organic, shifting environment and why the ramifications of choosing any standards need to be thought through. A lively discussion followed which continued through into the hour-long round table session, facilitiated by Alastair and Brian Kelly of UKOLN, which tackled some of the issues in more depth.

Read on for Alastair’s presentation, coverage of the further discussion, and Brian Kelly’s final roundup, plus details of the other presentations from the day.


The Standards Dilemma: Alastair Dunning, JISC Digitisation

See Alastair’s presentation on Slideshare

PlugPlugs are like standards – they are boring and dull but you need them. There are differences between UK and European plugs and if you have the wrong one you get unoperability.

Standards are everywhere but in context of SCA we’re talking about standards for digital content – file formats and metadata. There are also software languages, operating systems etc – not just content – but that’s where I’ll focus today.

When I started I though it would be simple to implement standards – just tell everyone to use the same thing eg tiffs for images. In a way it still is that simple but what we’ve learnt over the last 10 years is that it’s no longer feasible just to dictate standards.

So what have we learnt in the last 10 years?

Alastair Dunning at SCA Scotland ForumIt’s impossible to have a Stalinist view on standards. They cannot just be dictated from the top down. Skills are required to implement and exploit standards. Plenty of people find acronyms intimidating and if you tell staff that you want to use that particular standard it may present problems at the start eg METS (metadata and transmission and encoding scheme) – great package but takes time to learn. If you decide to adopt it then its an investment not a one off process. Although even basic skills can be exploited eg Tiff but some prefer eg JPG 2000 as can do more with them so again, requires learning skills.

End users require different types of standards. TEI is great for scholars but may not work for other types of user who want more simple types of mark-up.
But if you go too far and don’t have any kind of markup at all then it might lack the information you need.

Cost of standards fluctuate wildly – MJPEG2000 is hugely expensive for digital video as the file sizes are huge though great quality. But if you go to the other end of the scale, say with Flash, they are much poorer quality (example of YouTube).

Open standards might not be the most useful. It is just not feasible for some kinds of format eg OGG for music rather than WAV or MP3 as the latter are much more embedded in people’s consciousness.

How do we describe digital content and the formats we use for that?
The problem of digital silos – collections of digital content live in different places as they have been described in different ways. It makes it difficult to share collections. Basic Dublin Core can be used to describe some images but VRA4 may also be used and it is hard to bring them together. There are crosswalks and translation services to do this but are we fighting a losing battle?

Different people interpret things in different ways even if they are in the same community but developing communities of practice can help avoid the digital silo effect. Although standards may be fixed they inhabit an organic environment so any standards policy that is implemented interacts with other issues throughout an institution and that needs to be considered eg staff skills and training, costs, operating systems, relationships with end users.

Standards need to be agile. Content may be used by different users but also by the same users in different ways and it will be reused many times. Migration is not just about digital preservation. So it is important to at least use tools which can get you to XML and let your data to be freed up so others can use it in unexpected ways in mashups etc.

What is JISC doing about this?

Conversation with experts to provide guidance
UKOLN – wide-reaching concerns for all kinds of standards
Netskills workshops
Cetis
TASI (soon to expand remit)

JISC standards catalogue: experts from JISC services and beyond provide information with new editions created over time. Provides entries on key standards, including risk assessment.

SCA involvement: Planning for next iteration of standards catalogue which will be a catalogue tailored for different audiences, providing informed advice on cost, skills required and suitability for audiences. There will be an advocacy campaign to promote use and gain better understanding of audience requirements along with events like today’s – a discussion about what it means to implement standards.

Dennis Nicholson (Director, Centre for Digital Library Research):
You say that we can no longer dictate standards and I don’t think it’s ever been possible but organisations like JISC could do more to ensure that standards are taken up and implemented. We always get the same message that cannot dictate them and that is true but surely with funding organisations they can attempt to say that this has to interoperate and you must make sure they do. Even if people adopt the same standards it’s not useful if they operate in different ways. I think the organisations responsible for funding could push the community more about why they need to use standards.

AD: You can’t just say “you must use this standard” and then go away and leave it. You have to develop communities between those on the coalface and those handing money out to make sure they are taken up, and we do that with the digitisation programme. But I think there was a naivety around standards a few years ago and it’s more complex than that.

In access management we do say that people MUST use this and we are probably the only people in JISC that do. It is very expensive and you do face a lot of resistance.

DN: We need to focus less on the standards and more on what we’re trying to achieve with them. We need to take a more proactive attitude. It does not need to be prescriptive but does need to be proactive.

Stuart Dempster, Director, SCA: JISC is just one organisation and the issues are really associated with a couple of things – there’s the implementation, what you do want to achieve with a Common Information Environment, but also dialogue needs to take place within the other funding organisations too such as the research councils – all of these are probably suggesting various things. The SCA is about putting some context around the advocacy or standards. It is also about the economics. The Library of Congress devised METS and yet it’s not implementing it. The British Library has faced problems with it, not least finding people who are skilled in METS to implement it.

Charles Duncan, CEO, Intrallect: We need to differentiate between two different groups of users – there are those who work with them every day and then there are the users who benefit them and they couldn’t care less about standards. Those users do not need to know the details of the standards – they are simply tools. The most effective way to implement standards is to provide tools that mean that people do not need to know about standards but can just get on with working with them. Reload is a great thing – the tool of choice for many people worldwide now.

DN: We should think about the possible creation of interactive tools that people at the coalface could use in a user-friendly way. It does not have to be prescriptive advice, can be sensible advice. It’s not easy but it’s a good point that tools and making it easy is important.

Brian Kelly, UKOLN: I was initially going to say that I disagree with my Stalinist friend but then I realised that Dennis talking a lot of sense when he went on to say that we need to focus less on the standards and more on the sustainability of the standards, the take up of the standards, the ways in which there might be different approaches. I’ve got a problem with this emphasis on the standards in isolation. We have a problem if we’re trying to give the message out that it’s simple – its actually not simple at all and I think we need to get a different message across. The slogan of “interoperability through open standards” is just too simple.

AD: I think the argument of why you should use open standards is simple and the core of it should be simple, especially if you are putting the argument to people.

DN: But if it’s complex it’s complex and people have to learn to deal with it. We shouldn’t apologise for things being complex. The current message doesn’t work and we need to get a bit [laughs] Stalinist!

Rachel Bhandari (ICT Development Manager, Museums Libraries and Galleries Scotland): While it is complex, our membership ranges from whole departments who deal with this to one man shows and we need to support the range of people in doing this. The toolkits we are building should help.

AD: It’s about having a different message for different people. My dad runs his local museum. It has about 2000 objects and all the volunteers who run it are over 60. They are interested but if you tell them standards are complex they will be scared off. But for larger organisations it’s a different message.

Lynn Corrigan (Information Systems Manager, Napier University): For too long the elephant in the room has been that standards cost and that message has to be got across to funders and investors. They cost, in time and development as well as in other ways and that is rarely considered.

AD: Also, finding data on that kind of thing just to find out how much is costs is hard. It’s not very clear. The SCA wants to look at this eg how much it would cost to implement METS.

DN: It will cost differently in different contexts – the peripheral costs and benefits will be different. Nobody wants to know about how much they cost – they are not sexy – people want pretty pictures and videos, not to know about the metadata. People don’t want to hear it and we have to change the message. Funders want to see the upfront stuff, not the background stuff. We need to get more to grip with these kind of issues. It’s a question of skills – nobody knows those standards well enough and shouldn’t some funding go into that?

The discussion continued during the afternoon in a breakout group on standards, facilitated by Brian Kelly and Alastair Dunning.

Round table discussion on Standards

SCA Scotland Forum May 2008

BK: What are the pinch points – to describe the key standards? Dennis and I decided no, it’s not that simple. It’s not so much about looking at particular standards but at the framework.

DN: There is also the question of support for people and the need for local expertise. I also want to bring up need for organisations that are looking to promote standards to interoperate with other organisations that do that. In Scotland there is an existing framework of sorts. The standards commission I am part of is one of a number of groups that sits under CISAG and try to advise on standards and also engage with the user community through Digital Access Scotland (DAS). If the SCA are going to coordinate things across the UK then they need to be aware of existing organisations and their set up, and there may be similar things in Wales and Northern Ireland.

BK: We might also want to think about learning from the past and learning from mistakes – we are now living in the future. 10 years ago it was all going to be great and interoperable…what mistakes did we make? [brandishes historical document] In 1995 we were encouraging people to use SGM, VRML, and Whois++ for searching.

DN: A case in point that it is one thing to say here are some standards, follow them and another thing to say how to implement them.

BK: It also mentions URNs…

CD: We really were in our infancy then and we’re only toddling now…there will always be complexity and lots of standards and where we are now is that when we try to bring together lots of standards everyone picks the lowest denominator and then decides that it’s not good enough. We have to recognise the complexity and raise the lowest common denominator.

DN: But one of the important things is that when an attempt is made to apply standards it needs to be related to function and if the function is to make, say, lots of searching systems work as a whole then the lowest common denominator may be the way to go although it will lose functionality. It depends what you want to do. We need to be talking about what we are trying to do and start from there rather than starting from the standards.

BK: I think when we’re talking about standards it’s a verb not a noun. In the past the standards bodies have looked at take up rather than how it should be used. It goes back to the question of context.

Penny Robertson (Senior Information Officer, SLIC): There is also the practical application of those standards. Some of the search material you get back from some of the standards we are using brings back no useful results. We have to explain to the people doing the work why were using those standards through practical examples and case studies.

DN: You need the local expertise. It’s a process, its not about implementing particular standards but about the people who are about to do something having the support to use those standards. You have give people handles on which to build that expertise.

LC: Paul Ells described how they trained people in his project. You have to explain to people WHY we do it. Leading on from that is the end user. We still don’t do enough analysis of our end user. I’ve been analysing search strategies from the end user. You find that they can’t spell but if they misspell that’s one user who doesn’t get the result they want but if the person putting it in spells it wrong then there might be 8,000 students a day who are being disadvantaged by being unable to find the thing they are looking for.

BK: So we should ensure that interfaces can take into account spelling mistakes?

LC: That’s very difficult to do but it is one of the things about standards that needs to come in with the training and with the philosophy – to make sure that all the data is there and to explain that it matters and is why we add these extra fields and why we have to make sure we are accurate and code things correctly.

RB: And that standard terminology is appropriate to the level of your audience. American system is sidewalk and things like that. Goes back to knowing before you start who it is for and what they will be doing with it. Can’t be standards for standards sake – has to be appropriate to the audience

BK: Seems to me we are talking here about metadata content.

LC: It’s a very real and very international example.

PR: There used to be the idea that cataloguing didn’t matter because machines would come along and do it but we are now returning to idea that you do need people with those skills.

DN: You do still get cataloguing departments where the people don’t understand what they’re doing and why. What about the idea that JISC might help fund something to help with spelling etc?

AD: It seems to be a long term thing abut training and advocacy

DN: Training is important but there is also a general lack of understanding in organisations and you need someone to take that on and have responsibility for standards who knows what they are for. Not all organisations will be able to support that.

RB: And so you need an external resource to help those smaller organisations.

BK: It seems like we’re looking at standards to support resource discovery across disparate target audiences.

RB: Is it possible to have one set of standards that can do that?

DN: There is a real need for material and maybe even meetings and training courses that explain why we’re asking people do this stuff and what the effects of not doing it will be. It needs to be from grassroots to chief executive level so that significant resource is put into it. People always underestimate how much money to put into metadata.

Philip Graham (Public Services Manager, The Royal Commission on the Ancient and Historical Monuments of Scotland): We want to open up our material, including to schools, and we would need huge resources to do that as would have to interpret each site and explain it in simple terms. We can divide our users into divers (want full, complex information), swimmers, and paddlers (just interested in local history).

BK: But do we do that description or allow others to do it? We have the example of public wikis that allow others to do that. It opens things up and there is the question of do we want to do that because of issues about spam and so on but we also can’t provide training to everyone.

AD: Descriptive metadata is just one part of it. Some kinds of standards, such as METS, require expertise.

DN: There is no skipping the fact that, if a project needs METS then it has to be funded in some way. Can’t pretend that we’re doing it right if we miss that out. It needs to be built into the funding requirements of the project. Public money needs to be spent in the best way possible.

BK: But could social networking to so some of this?

PR: The Library of Congress released an image archive through a deal with Flickr and released it with a soft launch and just the Library of Congress tag and it is amazing how many other people have gone in and tagged it.

BK: Flickr Commons, it’s called.

CD: A good example is if everyone’s talking about the same resource you can add an identifier. If you can identify an object you can bring information from different places eg the ISBN number of the book can bring you the front cover, index page, reviews etc. if you have the identifier then you have the potential to get away from the lowest common denominator. Do people see the identifier problem as something that’s solved already?

BK: Or is it insoluble? Suggestion is that the Wikipedia entry for Edinburgh would be its URI.

AD: A good example is Librarything which has a great database and if you were doing advocacy that would be a great case study.

LC: But Edinburgh already has a unique identifer in the Library of Congress.

DN: Wikipedia is strong at the moment but cannot reply on it. You need to be able to rely on a URI and we’re in a constant state of change.

AD: WorldCat?

LC: Still in the biblioverse.

BK: It seems to me that we’re focusing in on metadata, importance of identifiers, preservation aspects…

RB: Nobody wants to go there because it’s scary!

BK: We’ve got websites that disappear but also file formats that disappear. Do we have a responsibility to ensure that they are accessible in the future?

RB: Is there an agreed place to point people and say this is where you’ll find all your file format requirements?

AD: National Archives in the States?

BK: dcc.ac.uk also has some

RB: For those involved in pointing members to resources, there’s an awful lot of sites that profess to have what you’re going for.

DN: Yes, you need to be able to say this is the definitive place to go to for standards.

BK: And do we question every standard or say yes, this is a standard we’re safe with it? There will be arguments.

PG: What about open source standards?

BK: There’s open document format and OOXML (new isostandard for Microsoft Office) – two standards doing similar things. Like there are three for syndication.

DN: Competing standards are fine but there has to be somewhere that we go to find this out.

AD: There’s the standards catalogue…

DN: But how many people want to go there?

CD: Here’s an example: there’s Pronum, and while it’s fine to know it exists it is better to know that there is a service called Droid where you send it a file and it comes back and tells you what sort of file it is.

DN: So we need to build up real support

AD: How might that support be structured?

DN: Would be nice to be able to point people to particular tools on the internet that they could interact with and the end result would be some kind of guidance given their own situation. Not going to be easy but worth it.

AD: What variables?

DN: Might start off with “is the tool you’re about to build a search and retrieve tool” which might narrow it down to descriptive metadata. Then narrow it down to user needs. It would be like a database. Would need exploring and pilots.

AD: Or might be better as a human? as a service? Someone you ring up to consult?

CD: What about tools linked to those standards? That would be very helpful.

BK: But it needs to be a useable tool for the mass market or checklist. It might ask if the commercial world has deployed it, can you use it in a small museum, is it deployed in other contexts than at JISC, is it used elsewhere in a sustainable way?

DN: Doubtful that you could create one tool that could do everything.

BK: Need to clarify what we mean by tool…

DN: But only giving an example, need to make it easier for people to apply standards but difficult to pin down.

BK: What about if you don’t need standards – want to do machine to machine – might just need an architectural approach – something less complex – seems to be tendency to keep going at policy level to say that need a standard.

DN: I am appalled at number of communities who have put lots of effort into developing own standards but have not implemented them.

RB: We need good examples to show to people where the standards approach has yielded benefits for that organisation – that would bring more support.

BK: Also case studies that show the cost of failure of not following standards.

DN: If we stop focusing on standards and look instead at what it is we’re trying to achieve then the decision may not end up being one about standards.

CD: Can’t define the kinds of tools you want unless you know what the end users need to achieve and what the geeky people need to build the things and services that the end users will use.

PR: And those geeky people need to be recognised by their organisations.

BK: JISC could be too successful in promoting standards but it should be invisible to the end user.

AD: We have a problem in preaching to the converted – too many people higher up in institutions see something from JISC and just pass it on to the ‘technical people’ then won’t listen to the technical people when they make those decisions.

PR: Why isn’t this a government initiative?

RB: It’s not sexy.

PR: But we’ve just had an interesting discussion about it – why aren’t younger professionals interested?

AD: If you want to be a librarian, this is where it’s at.

BK: It’s not just standards, it’s standards for a purpose. Standards is a horizontal activity, not just for the geeks. At the top level is the users, and for the enterprise and preservation it is an enterprise issue, and there are business cases – so perhaps we need to get this model across and say that it permeates many different levels – not just interoperability through standards but all these things.

AD: You need advocacy and training at lots of different levels – need to make people aware high up that it’s a crucial strategic issue for their organisation.

RB: At the moment it’s JISC-led and quite academic in outlook so how can other sectors play a role and how will it work best for them too – not just about universities.

Summary of the session from Brian Kelly:

An interesting session even though Dennis and I didn’t have a big argument!

We seemed to be in agreement that we don’t want standards for their own sake. What we really want are the benefits that standards can provide and we need to focus on the user benefits – does it provide the fuctionality that end users want, the long term preservation to our resources that our organisations might need, does it provide the flexibility for data to be reused… The emphasis needs to be on the outputs of open standards rather than the open standards themselves.

So there is a context and there might be differences depending on the point of view – universities, museums, schools. There are different resource and expertise implications in the different sectors.

Our discussion focused on: resource discovery and the role of metadata to support that and the issue of preservation.

We felt that standards are very costly and really need to be embedded at a whole set of diferent levels. They are currently seen to be a techy thing and that raises the danger that CEOs might not pay attention to them and they might be argued about at the wrong level.

That’s a problem because standards should go across all those different layers – standards are not just for techies. The discussion needs to include users, enterprise issues, legal people, staff development, training… We felt that could gain a lot from case studies of successes but also case studies of failures – not at project level but at uses of standards. There is the example of 1995 standards – CGM, VRML, cross-searching were recommended – and we made mistakes by focusing in on the exciting new standards. Are the standards we are promoting today going to b tomorrow’s Whois++?

We need to learn from mistakes from the past and we need to have tools to support the use of these standards. At higher level they may not be an open source tool but embedded in commercial application. We need evidence of take up in mainstream activities, not just development activities, and also tools to use as a checklist for when starting a new project. We may then find that it is not a standard but an architectural approach that is required. We need a bigger and broader view of when and why we want the standards and how we will embed them in a sustainable way.

The point was also made that anything the SCA does in Scotland must interact with existing frameworks and set-ups. We were also very keen to develop the idea of supporting local expertise as much as possible and to support infrastructure within the institutions, such as peer support communities perhaps using social networking.

  • The next SCA Scotland Forum will be on 25 September. Find full details on our Events page nearer the date.