Noun qualifiers of other nouns are ambiguous, especially when there are two or more. Newspaper editors know this – they use them extensively in their headings, as research shows that people interpret them in different ways. It helps persuade the headline skimmer to read, or at least start to read, the article.
The one that bugs me most is ‘content management system’ (CMS). Two different noun qualifiers, but which qualifies which of the three? From my experience they are content systems and management systems but rarely a system for managing the aggregated collection of content. Let me explain.
Many of the larger CMS come from an enterprise background where the task was to get in control of the many documents wandering around. They have a lot of functionality including nifty conversions from one format to another. And good tracking of where these documents are. They also often contain many tools for managing the workflow in their production, authoring, approving and publishing. The last is usually done by moving to a pre-publication area with a publisher role making the final act of releasing to the intended audience. Then Web came along and a functionality was added to send to a website.
There are also many content systems that store words, pictures and other media, in a structured way in a database and then publish to the Web by adding in (X)HTML coding and style sheets and, if the developer is smart, rules for displaying one or another type of content depending on the situation in the system or type of viewer. These come from a background of Web publishing and although they do that well, they usually don’t offer tools that work across the database.
What we need in government though is the third – content management:
- ‘Give me all the content that does not contain an entry in a particular metadata field and list with the contact emails of the authors.’
- ‘List all the pages that have been viewed only by internal staff.’
- ‘Find all the content containing expression X and change to Y and add to a new metadata field’
And, most of all,
- ‘Extract out all the content according to these rules, structure them in this particular way and send to go off to another CMS’.
It’s all do-able but so much hard work at present. We should have easy interfaces to manage content. I care, because we have many talented digital media staff in government (maybe not enough, but they are certainly there), who waste much of their time and effort in struggling with getting content in and out and between CMS’s. Usually because there is so little functionality to manage content across the whole collection and because exchange between systems is so difficult. There must be a much better way of doing this and so releasing government and public sector expertise to contribute to the public agenda. I’m carrying a banner to all those providers of CMS and web publishing services to say ‘Free our digital media experts!’
By doing more information structuring and using common structures, as we’ve been introducing for jobs and consultations, and have been done for a long while in press releases and descriptions of documents, we should be able to build systems that easily exchange information when there are machinery of government changes and that make it easy to manage the content. (And, of course, add in semantic web coding in a supported way!) Then we might be able to start describing systems as those that enable and facilitate content management.
Posted in: General
Tags: CMS, content management system
Amen to that, David. And good luck in the struggle for interoperability and ontological thinking in the face of vendor lock-in and quick fixes at the system specification stage.
How could this be made better?
- Common schemas for key content types: only a few content types have fairly well-defined conventions (e.g. News stories) which greatly help in their definition and communication. Your team has done good work to expand this into new areas, and it feels like there’s more fertile territory for easy-to-understand schemas defining the entities and relationships around content such as speeches, publications, biographies and even policy updates. Supported by some easy-to-understand documentation, these could help CMS integrators and government webbies jump-start the specification process, and provide content at least loosely structured in the right ways for future interchange and migration; at best, ready-made for machine-readable access via RDFa, RSS or other APIs.
- Specialised services, with open APIs: stop trying to keep it all in one system. Deploy specialised, centralised systems (which generate efficiencies in themselves) but which provide content out in formats which can be consumed and republished easily by other government (and indeed, non-government, sites). Civil Service Jobs and Public Appointments spring to mind as great examples.
- Less content: again, there’s good work from your team in promoting archiving and web continuity away from the corporate CMS. With the pace of change in government and technology, let’s make the task easier by making our sites smaller. Keep content snappy, up-to-date and discoverable, but don’t necessarily house it in the CMS indefinitely.
Steph Gray says:
February 13, 2010 at 1:26 amYou could encourage vendors to move towards a Content Management Interoperability Standard (CMIS – http://en.wikipedia.org/wiki/Content_Management_Interoperability_Services). Several big-name enterprise CMS vendors are already participating in the process.
In Australia, government agencies are encouraged to adopt the Australian Government Locator Service (AGLS) metadata standard. Takeup levels tend to vary widely between agencies, though.
@StephGray – several commonly-found government document types are enumerated by AGLS. Check http://www.naa.gov.au/records-management/create-capture-describe/describe/agls/encoding-scheme-document.aspx
Many of the usual suspects are there, but a few glaring omissions exist (”public consultations” come to mind).
Gordon Grace says:
February 13, 2010 at 12:48 pmAnd amen from me too. Uncanny timing, David – have been trying to find time to post along similar lines myself about my frustrations with CMSs and how they do almost everything but manage content; and what a fantasy CMS for government might need to do.
To add to the excellent ideas above, there were some great comments on this post of mine about quality assurance standards for content reviews: http://neilojwilliams.net/missioncreep/2010/the-perfect-page
In particular, that CMSs should learn which editors need monitoring and which can be let alone:
“User G has submitted another article, you have approved 90% of his past articles so unless you say otherwise this article will be automatically waived through and published in 59 minutes ..”
… with content editing, user accreditation and QA on the scale of Directgov and Businesslink, just think of the hours this kind of thing could save.
Neil Williams says:
February 13, 2010 at 5:47 pmYes, there are no perfect systems, but certainly many in the marketplace that fit your requirements, either out-of-the-box or with little customisation.
Perhaps you have felt it like tyranny because the completely wrong system was selected or the right system was implemented horribly by an incompetent agency. Either way I don’t think the market needs more systems. I realize there are many failed projects and unhappy CMS users.
What we need in government is informed buyers. This is what will make the big difference and a blog like this is certainly very helpful. Putting some pressure on vendors is also a good thing.
If you would like to focus on time savings, usability is a key area that many buyers underestimate, in particular as they move beyond the initial sales demos in the procurement process.
Janus Boye says:
February 13, 2010 at 6:32 pmDavid, there is a system that will do exactly that, and enforce security policies on documents on-the-fly too. Its not a CMS, but a content-based information assurance system. It’s targeted at data loss prevention and information audit and it will answer every one of the questions you asked – very quickly. It links very nicely to a CMS. The company also has a solution for Sharepoint information sprawl too.
Tom Foale says:
February 13, 2010 at 9:59 pmInteresting post David. In response to your requriements:
1. Give me all the content that does not contain an entry in a particular metadata field and list with the contact emails of the authors.
- Why do you have metadata fields that don’t have entries? Any decent CMS will be able to define certain fields as mandatory, it’s just got to be implemented correctly. A far bigger issue is that people put the wrong metadata in. That’s not really a problem with the software but with educating users.
2. List all the pages that have been viewed only by internal staff.
- Given the way that most CMS are deployed in order to be robust and scalable, identifying internal user views directly from CMS reporting is nigh on impossible. You’ll need to have closer interaction with whichever analytics package you have, and be able to trust it. What about staff working from home? Or people accessing from common IP addresses (particularly behind the GSI firewall) who aren’t actually staff at all? This is a wider problem with the internet, not content management.
3. Find all the content containing expression X and change to Y and add to a new metadata field
- Not sure why you can’t do this. Many CMS offer this option, although typically it’s to more technical users as it’s the last thing you want someone doing by accident.
4. Extract out all the content according to these rules, structure them in this particular way and send to go off to another CMS.
- Aren’t there people biting your hand off to do this kind of work? If there aren’t, contact me and I’ll get it sorted out for you!
By the way, content management systems are meant to be tyrannical: http://contentedmanagement.net/blog/the-mirror-stage-in-content-management/
Regards,
Philippe
Philippe Parker says:
February 15, 2010 at 10:52 amThanks all for comments. I’m convinced that there are good solutions, but for some reason we in government are just not selecting what we need to make digital media work effectively. Often it is because we are trying to use a single product to do deliver websites, intranets, extranets and online services. And also as Janus Boye says, we need informed practitioners really guiding the procurement process more strongly.
@stephgray – interesting thought. I was this morning starting to think along similar lines and you expressed it so well. There’s a centralised press release service NDS run by COI, a centralised jobs run by Cabinet Office, both with views/feeds back onto corporate departmental websites. Why not indeed develop the theme and provide specialised centralised services. What might be included in that? Ministers and their responsibilities, organisation descriptions (especially of all EAs and NDPBs), a document repository, Contact Us. In fact there are so many in common across central government bodies, that people also want aggregated. Maybe this is something we should quickly work up? The real question is – what would be left behind for which one couldn’t deliver through centralised APIs?
@TomFoale, @JanusBoye – thanks and not only usability but also accessibility. Many CMS provide accessibility in their output to endusers of websites, but not to the users putting information in or managing the content. We found this when selecting a CMS at ONS when there wasn’t at that time a single product on the market that met government requirements on accessibility for users rather than endusers.
@PhillipeParker – LOL Yes often ‘Document0′ is entered – or a such a generic subject that it is of no subsequent use in selection of the content! That would be a better example. It’s interesting that everything I describe as little basic tasks are possible – but just not easy at the manager interface level. You’ve set me thinking that maybe we need a review team to look at how some of our services are set up. Of course many of them are still digital paper (publishing individual pages with hand-coded links everywhere), rather than web publishing of content and those won’t, by and large, allow any of this kind of functionality.
David
David Pullinger says:
February 15, 2010 at 11:38 amI work in Steph and Neil’s team at BIS and we’ve done a lot of work on essential fields for speeches, ministers, policies… We’re happy to make this info publicly available.
Will Callaghan says:
February 15, 2010 at 12:09 pm@Janus Boye: Are there *many* CMS that fit all the requirements? Customisation required? Or OoB functionality?
John Goode says:
February 15, 2010 at 12:11 pmI’d back your campaign that wants to ‘Free our digital media experts. I’d also commend you on defining some “standards” around standard content types, more on that later.
Your post highlights what many web and digital staff are frustrated with! I do agree with @JanusBoye some of these are due to the wrong systems or bad implementations, and there’s too many of those!
I think some of this is due to not having a coherent content strategy already in place when specifying and procuring these systems. CMS systems are just tools, and before getting anywhere near one you should have all the building blocks in place. e.g. metadata strategy, editorial strategy etc.
CMIS promises much but there are so few vendors who support it, which is disappointing! See http://en.wikipedia.org/wiki/Content_Management_Interoperability_Services
The future is about Open Standards for Content and Content types, which have clear benefits for accessibility, interoperability, re-usability etc.
With the high profile launch of http://www.data.gov.uk. I think we need to apply the same principles that we apply to data to web Content. The Guardian’s Open Platform is one example of how content could be accessed via their Content API http://www.guardian.co.uk/open-platform
We also need Open standards for content types, seems like you and some of the brilliant minds on here have started that process already!
Once we’ve defined what the requirementrs, then implementing them using a CMS shouldn’t be a problem!
Zahoor Hussain says:
February 16, 2010 at 12:51 am@WillCallaghan – thanks for the offer. I feel a working group coming on that defines some of these content types. Ideally they would have as much in common as possible and only differ in certain parts. I think DCSF are probably following a similar route in seeking to extend the metadata-driven approach adopted by sites such as BBC and Guardian Unlimited. National Strategies was the first major metadriven front end in government, although various aspects of it are also in ONS and businesslink.gov.uk
@ZahoorHussain Thanks for encouragement. The standards for data began in the Web content and information world with John Sheridan’s work on The London Gazette and mine on Public Sector jobs and consultations in RDFa. We’ve got another three in the pipeline and probably need to push these through. The one that would help most people most of the time is a consistent Web content standard describing official documents – whether the RDfa metadata (which would essentially be Dublin Core) is held within or external (or both) to a PDF and other document formats including HTML. You’ve encouraged me to push on with this.
As well as wondering if we make it a requirement for government CMS that they must be CMIS-compliant. it would make machinery of government changes so much easier!
Lots of food for thought.
David Pullinger says:
February 16, 2010 at 11:10 amDavid: with regards to “Many CMS provide accessibility in their output to endusers of websites, but not to the users putting information in or managing the content. We found this when selecting a CMS at ONS when there wasn’t at that time a single product on the market that met government requirements on accessibility for users rather than endusers.”
I’m not sure when you were looking, but back in 2004, Plone was demonstrated with a visually impaired user authoring content in Plone via a braille keyboard and text to speech software (also Plone’s i18n system was providing the whole site in German for her):
http://www.netsight.co.uk/news/articles/SnowSprint
As for the rest of your comments, I think the standardization of certain documents would be useful. I don’t hold out much hope for it to be honest though. CMIS may help a bit with interoperability, but on its own doesn’t give you anything more than what you generally already have with CMSs (ie. CMIS does not solve any of your issues around standardising content types… just as WebDav, FTP, etc didn’t before it).
My main beef has been the way public sector goes about procuring systems and how from my experience really don’t know what they are actually doing. We are involved in a public sector procurement process at the moment, which has now been delayed as the client wants to go back out and consult more with its stakeholders. Only it has specifically excluded anyone tendering for the work from the consultation. This basically means none of the vast sea of local expertise can be used to help them, as anyone who has any experience and is qualified for the work is part of the tender process.
-Matt
Matt Hamilton says:
February 16, 2010 at 2:38 pmI’m surprised the term ‘Open Source’ hasn’t come into the conversation yet. Allow me to rectify that.
The government’s Open Source policy only really tackled the subject from the procurement perspective. But the real power of Open Source comes when you roll up your sleeves, and build on the off-the-shelf product – be it via plugins, contributing code to core, or forking the original product. Choose the product which comes closest to what you need; and make it exactly what you need, as the French military did with Mozilla Thunderbird:
http://puffbox.com/?p=1351
Ask not what Open Source can do for you; ask what you can do for Open Source.
Simon Dickson says:
February 16, 2010 at 11:53 pm@MattHamilton – you’re right, we were looking a little earlier than 2004; nevertheless was grateful for the link. Sounds worth following up.
@SimonDickson – I wondered that too! But thinking through it, there have been (which may no longer be the case now) two perceived stumbling blocks: first when a Drupal microsite went live in government for MoJ last year, it was believed to have the highest requirement on the number of simultaneous users ever, and the manager (who may wish to contribute – hint) was not quite sure what was going to happen. It can be difficult for government procurers to accept a context that is untried and the stakes in poor publicity can be very high. Generally these problems are solvable with virtual servers and balanced loading etc.
The second is the location of where the expertise is – we need more coding capability in the way you describe that understands what public sector needs. It’s easier for the established procedures to procure and know what you are getting through the traditional SI route – we need to find a way to break away from that and into a dynamic digital comms world with sufficient intelligent user functions inside public sector and enough external SME suppliers with the expertise. We’ve got some good case studies (some with your help) and perhaps the time is soon coming when we can really start trialling this approach in the mainstream.
David Pullinger says:
February 17, 2010 at 5:55 pmI see the issues which have effected us are also issues that effect our lords and masters.
I always get the feeling we have taken a bigger leap than we are prepared for. When we ask industry to come up with a solution for our problem they always give us the most expensive solution with the most constrained contract.
Their in business to make money so they can’t be blamed the blame lies firmly with our inability to articulate what it is we want, we take the “jumpers for goalposts” analogy to the next level and are often still wearing our jumpers, leaving anyone trying to decide what to sell us a very easy choice. Give them everything we offer and lets try and help them decide what it is they actually want.
The in house CMS we use gets better every release because its backed by ideas and user requirements. As an example I was involved in a recent semantic web workshop and after a very short time and a quick chat with our developer we came up with a plan of how we could implement some basic semantic web enhancements. The CMS is being refined through real requirements but more importantly its solving an issue that we have identified ourselves and not an issue we have been told we have.
Andy Paddock says:
March 1, 2010 at 9:58 amIt took us nearly two year to purchase a CMS. The question of what we needed the CMS for was never really asked.
In my experience buying a new CMS has been a way of avoiding asking questions about our organisation and how it needs to adapt to the web. A new CMS will not improve the quality of content if there is not culture of editing or knowledge of web copy-writing.
Got to get the people right first.
Len says:
May 25, 2010 at 3:59 pmWe built our own, with a long trusted IT partner. Result it a flexible system that can be added to, and works good as gold.
Ben Smith says:
May 28, 2010 at 10:22 amConsistent and highly descriptive metadata is the key to what you’re describing.
Of course, often the client doesn’t know what queries are needed until they start to use it in anger, and you need the queries to know what metadata to add to the content. Thus the content lacks the metadata required to make the queries work properly.
A classic chicken and egg…
David Hamilton says:
May 28, 2010 at 12:02 pmBTW: Really think that ‘dot gov’ blogs should support integrated commenting systems, such as Disqus or OpenID. How about it?
David Hamilton says:
May 28, 2010 at 12:05 pmHello David
Earlier today, a colleague thoughtfully pinged me with a pointer to your post – and he highlighted that there were several connections that could be drawn between the points you have raised and themes I have regularly hammered upon from conference podiums and from my blog (http://www.gollner.ca).
The core point that you are raising – or at least as I immediately interpret it – is that CMSs have historically taken a very narrow view of the “business of content”. In particular, CMSs have tended to fall into the trap of being “application oriented” and this directly, and completely, mitigates against processes (such as management, rightly considered) that need to be applied broadly to enterprise collections of content. In opposition to this inclination towards “application oriented architectures” (which is as commonplace as it is self-defeating), I have positioned what I call “Content Oriented Architectures (COA)”. Colleagues who have not yet recovered from wrestling with Service Oriented Architectures (SOA) have been less than pleased at the suggestion that there might be other variations on the theme.
I have a presentation on Slideshare that specifically addresses this topic. See http://www.slideshare.net/jgollner/content-oriented-architectures-coa
On my blog, I have a category dedicated to the theme of extensible and exchangeable content, which I abbreviate as xContent. Some of the posts in the last year might be interesting. See http://www.gollner.ca/xml
Specifically a recent post on “intelligent content management” recounts a presentation I gave recently in Palm Springs. In this talk, I start out with a full frontal assault on the “Enterprise Content Management (ECM)” – effectively declaring that ECM has been an abject failure in that it has nothing to do with Enteprrises, Content or Business Value. See http://www.gollner.ca/2010/03/intelligent-content-2010.html
Also recently, I have contributed a chapter to a forecoming book on Information Management Best Practices and this chapter recounts a case study of content management that actually worked and worked on a grand scale (across a large military community and one that included NATO multi-national programs). One of the core lessons was that the success that was realized derived directly from the fact that we designed a solution around the content first and this solution established how various CMSs and other tools could play roles in the enterprise process. This inverted the normal practice of embracing a tool and praying for integration to appear as if by magic (say through a CM integration standard that is universally adopted by product vendors – which by the way has been tried several times in the past and has always failed). The lesson that comes out of this case study (one that spans a 15 year period) is that content management at the enterprise level is something that must be designed, deployed and defended as being independent of any one CMS or supporting application.
Your points, and the associated frustrations, find strong corroboration in the history of CMS technology. Fortunately this same history also provides glimmers of hope.
Joe Gollner says:
May 28, 2010 at 5:05 pmThanks for interesting responses.
@BenSmith – I agree a bespoke service can work well, and indeed I’ve have been responsible for a number that have been recognised as good. Equally, alas, there is a record in government for commissioning new ones that have cost too much and never delivered. Probably for the same reason as @AndyPaddock gives – the need to have a really good Intelligent Customer Function inside government. For this kind of reason it is government good practice not to be funding code writing unless unavoidable and to use tools already developed by third parties (open source where possible).
@JoeGollner initiated some interesting thinking for me concerning what we call things and how we position what we need. There are increasing numbers of research papers on the way that language can completely change outcomes in IT tool introduction, and this might be a context in which to review some of the frustrations we experience. Also very encouraging to feel part of a wave of frustration, I’ll follow up your links – thanks!
David Pullinger says:
June 11, 2010 at 10:34 amI don’t think you’ll find a one-stop solution to these problems unfortunately; I agree with BenSmith, a bespoke solution may be your only route but judging by your comments that’s an answer that simply isn’t available.
maybe your answer is in acquiring the right people rather than the right technology
BF Matt says:
July 22, 2010 at 1:03 pmWe also built our own CMS based on open source together with a local IT company that provides excellent support. It absolutely fits our needs as we defined everything with our partner.
Mike Rakia says:
December 2, 2010 at 9:46 pmThere is a scheme that will do accurately that, and put into effect protection policies on documents on-the-fly too. It’s not a CMS, but a content-based information assertion system. It’s under attack at information loss impediment and information check and it will answer every one of the questions you asked – very quickly. It links very nicely to a CMS.
SEO says:
February 3, 2011 at 6:13 amTo be honest I do like CMS’s like wordpress and joomla, they make life easier for me as developer but 95% of times more difficult for the clients.
Ioan Popa says:
February 9, 2011 at 2:36 am