Alan Pelz-Sharpe, Sharon Richardson and James Lappin discuss Office 365,  Microsoft's cloud offering.

Office 365 is a 'bundle' of cloud services, including e-mail, the cloud version of SharePoint, and Lync instant messanging.

We discuss the challenges that Microsoft have in providing both cloud and an on-premise versions of SharePoint.  

Sharon said Microsoft will need to improve the way it moves cloud customers from one version of SharePoint to the next.  

The cloud version of SharePoint 2007 had been part of BPOS (the predecessor to Office 365). The cloud version of SharePoint 2010 was not released until nearly a year after the on-premise version. There was no simple upgrade path- instead cloud customers had to migrate over from BPOS to Office 365.

The plus side was that the online version of SharePoint 2010 contains almost the whole range of SharePoint functionality (it doesn't have the SharePoint records centre though).  It includes for example most of the service applications (such as the user profile service and the search service).

The main competitor to Office 365 was Google Apps, which doesn't have an on premise version. Google Apps is nowhere near as powerful and as configurable as SharePoint online, and it does not have an on-premise version.  Its features have developed very little over the past few years, and such changes as have been made have been deployed by Google without disruption to customers.

SharePoint Online also faces competition from two very different sources:

- The relatively simple filesharing services such as Huddle, Box and Dropbox, which like Google apps do not have on-premise versions

- The powerful, complex and configurable document management products from ECM vendors such as Open text, IBM, Oracle and Documentum, which have on-premise versions as well as cloud versions and hybrid versions.

Alan said that the document management products from the ECM vendors are overwelmingly deployed on-premise.  If an organisation is deploying a serious document management system, with a degree of customisation and of integration with other applications, then it is almost certainly going to want that application to be on-premise rather than in the cloud.

Alan Pelz-Sharpe is Research Director of the 451 Group.  Sharon Richardson is an independent consultant and founder of Joining Dots Limited.  James Lappin is an independent records management consultant and founder of Thinking Records Ltd.

Between 2004 and 2006 Ben Plouviez (@benplouviez) oversaw the roll out of an EDRM (electronic documents and records management) system across what was then the Scottish Executive (but is now the Scottish Government).

Six years later and the system contains 14 million documents and is used by around 4,000 staff.

In this podcast Ben reflects even-handedly on both the benefits that having an organisation wide records repository has brought to the Scottish Government, and on the promises that the system has not fulfilled.

The roll out of the EDRM was driven partly by the Scottish Executive's desire to breakdown silos between the various different parts of the administration. They  made the decision that wherever possible files would be open and accessible to the whole of the Scottish Government. There have been times when colleagues have found documentation that they would never have known existed were it not for the EDRMS.  

The EDRM's Scottish Government wide business classification scheme has not been an unqualified success, but nor could it be called a failure.  It is not terribly popular with users, who rarely use it to navigate to material that they wish to find.  However on the plus side the scheme has provided a stable and enduring  structure for the system. 

Ben has found that the electronic files on the EDRM system do not tell a narrative in anything like as clear or as useable way as a typical paper file used to do.  Ben questioned whether it was feasible  for records managers to expect their organisations to keep a full electronic file of every piece of work they carry out.  Ben said that the concept of the file is predicated on the concept of the document and we are now seeing alternatives to the document in the form of blogs, wikis, discussion forums, etc.  None of these new formats fit naturally into the file.  James Lappin found it significant that MoReq2010 specification used the word 'aggregation' instead of the word 'file'.  This implies that in the electronic world there are many different ways in which business communications can be collected (e-mails in in-boxes, tweets in tweet streams, etc..).

There have been some unexpected benefits to having an organisation-wide records repository.  For example Scottish Government have taken information from the system's audit logs about who has read what on the EDRM and translated it into rdf triples (the non-proprietary format that underpins linked data and the semantic web).  They have then provided an interface to enable colleagues to query this data to find out what their colleagues have read on the system. This enables the serendipitous finding of documents of curerent interest, and provides a more human way of browsing and interrogating the system than that provided by either the business classification or by the search facility.  The Scottish Government have also used the same technique in relation to e-mail logs.  They have taken the records of who sent an e-mail to who and when, converted it to rdf, and provided a query and visualisation interface.  This means colleagues can find out who has been corresponding with particular colleagues or stakeholders.    Note that the content of the e-mail is not visible, only the fact that an e-mail has been sent, and only e-mails with at least one person in cc have been included to ensure that private correspendence between two people is excluded.

Ben talked about the plans for the future of electronic records management in Scottish Government, including their intention to replace their existing EDRM within the next three or four years.  He speculated on whether it would be possible for one product/system to fulfill both their collabortion and records management needs, or whether Scottish Government would have to implement several different tools to deliver that vision.

In this podcast James Lappin asks Matt Mullen to explain what Big Data is.  

This podcast was prompted by a blogpost Matt had written Big Data plus enterprise search = Big enterprise disappointment?

Matt contrasts the vendor driven,  enterprise-centric vision of Big Data (vendors selling tools to help organisations make use of the content they have accumulated over the years in different repositories) with the more transparent, idealistic and web-centric vision of linked data (organisations marking up their structured data with rdf and making it available for others to run queries on, or to use for data mash-ups).

Matt explains why it is easier for Google to make sense of the world wide web than it is for an enterprise search engine to make sense of documents and data from multiple different repositories within an organisation.  James and Matt discuss whether or not the distinction between structured and unstructured data is a meaningful one.   

The podcast was recorded at the Royal Festival Hall, London on 27 February 2012

Matt Mullen is an analyst for the Real Story Group, specialising in Search and in Web Content Management. He is on Twitter as @MattMullenUK

Richard Harbridge and James Lappin discuss information architecture issues within SharePoint.

Richard gives his rule of thumb for answering the following question - when a new area or function comes on board in a SharePoint implementation is it best to set up a SharePoint site collection or simply a site within an existing site collection?

We discuss the pros and cons of 'site collections' which are a feature unique to SharePoint.  Site collections are a hierarchical collection of SharePoint sites sharing common administrative settings and some common information archicture features such as content types.   Crucially a site collection cannot be split across seperate SQL server content databases, so there are storage as well as information architecture considerations to deciding how many site collections to set up and what for.  Microsoft recommends that each site collection does not exceed 100GB in size.

James asks about the relationship between site collections and search, and Richard describes some tips for configuring a SharePoint search centre with search 'scopes' set up to enable your users to target their searches at particular site collections or at particular types of content.  We discuss the strengths and weaknesses of refiners in SharePoint search.  Refiners are a set of links that are returned alongside SharePoint 2010 search results and which enable users to filter those results by defined parameters (for instance date modified, document type, project title).  James is disappointed firstly that the SharePoint 2010 refiners only filterthe first 500 results, but more importantly that they give no indication given to the user that only the first 500 results had been refined.

The discussion then touches on the managed metadata service in SharePoint 2010 as a way of getting controlled vocabularies out of the confines of a single site collection and into a place where they can be used by any site collection.  Richard outlined some of the ways in which the managed metadata service does not work as well as he would like (and mantions an article by Michal Pisarek in which these weaknesses are collected) but says he still recommends his clients make some use of it.

We finish by talking about 'business connectivity services' in SharePoint.  This enables data (in the form of database rows and columns) to be imported into SharePoint from another database within the organistion. Once the data is in SharePoint it can be used as a controlled vocabulary to improve the findability of content. Richard gives the examples of a law firm importing into SharePoint a list of its matter numbers from its customer database.   The one disappointment is that the business connectivity service does not work with the managed metadata service - it is not possible to import a list (for example a list of clients) into the managed metadata service from a line of business database and use that as controlled vocabulary within SharePoint.