SharePoint 2013 and WCXM

A lot has been written about and presented on SharePoint regarding it’s use as a WCM/WCXM tool.  My position, generally, is that unless you’re using SharePoint across the enterprise, it probably won’t be a good fit for the majority of firms as a WCM solution.   I reiterated this guidance in a SharePoint 2013 WCXM advisory paper and subsequent “Just how good is Web CMS in SharePoint 2013” web cast for the Real Story Group.

After I tweeted that little had changed between 2010 and 2013 WCM capabilities, Waldek Mastykarz suggested that I “didn’t get the memo.”  Waldek and I had a short twitter discussion on whether the latest release significantly improved SharePoint’s WCM capabilities.

Admittedly, there was some merit to his response, though I still maintain that for the majority of cases, SharePoint’s WCM capabilities are not significantly changed and the platform isn’t any better a fit for loads of WCXM scenarios than 2010.  However, I thought it would be interesting to post a set of questions I asked Waldek directly and let you hear what he has to say.   At Waldek’s suggestion, I’ve also included some additional commentary that he and I exchanged through e-mail.

1. In a blog post I published on the Real Story Group blog, I contend that not a whole lot has changed for WCM in the new SharePoint 2013.  You tweeted that I “hadn’t read the memo.”  Have I missed something really materially different?
   
   In SharePoint 2007 and 2010 the publishing model was based on Publishing Pages. One of the consequences was tight coupling between the physical location of the page and how it would appear in the site’s structure and in URLs. Additionally it was challenging in previous versions of SharePoint to reuse content across different sites. Finally it was maintenance-intensive to have an intelligent website which content and presentation would adapt to who the visitor is and which device he uses to browse the website. All of those can be now solved using the standard capabilities of SharePoint 2013.
   
   Next to the new search-based publishing SharePoint 2013 contains a number of capabilities that help you optimize your website for public search engines. In previous versions of SharePoint those capabilities had to be covered by third party solutions. Those are just two of the many WCM capabilities that SharePoint 2013 offers and that change how we think about and build websites on the SharePoint platform.
   
   [SHAWN’S NOTE] The search-based publishing features are not available in the Office365 version of SharePoint and only available on-premise.
   
   
2. You contend that search driven navigation is a major change in the WCM capabilities in SharePoint 2013.  How so?
   
   In previous versions of SharePoint the navigation was based on the physical structure of the website. If you wanted to publish a press release, you would navigate to that particular branch of your site and create a new press release page. With SharePoint 2013 search-based publishing and Managed Navigation this is no longer necessary. First of all we can leverage Managed Metadata and taxonomies to build more flexible navigation structures. Secondly using search-based publishing we can tie those taxonomies to the content and have everything published dynamically on the website. This not only increases web content management flexibility but also lowers the effort required to build and maintain websites.
   
   [SHAWN’S NOTE] SharePoint’s managed metadata service was available in 2010, though basing web site navigation on the term store is new to 2013; in 2010 metadata-based navigation was available for document libraries and presenting documents ordered by terms applied to those list items.
   
   
3. What are some of the other changes in SharePoint 2013, with regard to WCM, you think deserve highlighting?
   
   SharePoint 2013 contains many new and improved features for building public-facing websites. First of all it provides us with Search Engine Optimization features that we can leverage to optimize the website for public search engines. Next there are the Rich Text Editor improvements that make it easier for content managers to work with content. One of the common challenges in previous versions of SharePoint was working with content from Microsoft Word where all of the internal markup would be copied into the web page. In SharePoint 2013 this is no longer an issue as it automatically cleans the HTML of any Word markup. With regard to search-based publishing we can now leverage information about the visitor, such as information from social networks or his click-behavior, to personalize the content displayed on the website. With this we can build truly intelligent websites where the presented content is automatically tuned based on who the visitor is and what he did on the website.
   
   SharePoint 2013 also simplifies building websites optimized for mobile devices. Using the new Device Channels and Image Renditions capabilities we can now easier build websites optimized for different audiences using SharePoint 2013.
   
   Also, when it comes to scalability, because SharePoint 2013 uses enterprise-class search engine (previously known as FAST) as a fundamental piece of the content delivery mechanics, it is much more scalable than traditional database-driven approaches – especially if you take into account content personalization capabilities such as Recommendations and User Segments that are now a part of the standard functionality of SharePoint 2013.
   
   Another thing worth mentioning is that using the new and improved REST APIs it is so much easier to reuse content from SharePoint 2013 outside of SharePoint. With the new API you can very easily create native apps for Windows Phone, iPhone or any other device/OS that will interact with the content store in SharePoint. With the new APIs you not only get access to the static content but can also benefit of the powerful capabilities based on search-based-publishing such as recommendations and user segments which allow you to easily build apps with rich and dynamic experiences.
   Even though SharePoint 2013 has been released just recently, there are already some customers having their production websites live built using SharePoint 2013. One of the cases is available at http://technet.microsoft.com/en-us/library/jj822912.aspx and presents the implementation of the mavention.nl website which was the first website worldwide live on SharePoint 2013.
   
   [SHAWN’S NOTE] Search engine optimization features include improve URL readability (e.g. no longer necessary to have the “pages” library name in the URL in all cases) and the ability to inject arbitrary metadata into a page through SharePoint site settings.  I had noted editor improvements (specifically improvements in markup produced), though SharePoint still lags other editors like TinyMCE.  Device Channels and Image Renditions are improvements,  the capabilities represented by these two features were available 2010 (automated thumbnailing was a feature of an image library and the mobile capabilities of 2013 simply make configuring device channel easier, though there’s little improvement over 2010 and you’re still required to build an entirely new site to support specific mobile devices).  Finally, the FAST engine was available in 2010 and the capabilities represented, beyond search-based publishing, are little improved (if at all) in 2013.
   
   

   [Waldek’s Response] Image Renditions are more than the thumbnail capability available in the previous versions of SharePoint. In the past SharePoint automatically generated thumbnails for images. Those were however fixed and always sized (as opposite to cropped). With SharePoint 2013 you can specify yourself which kind of Image Renditions you want to use on your website and how every single image should look like for a particular rendition. This makes Image Renditions way more useful than the thumbnails from the past.
   
   [Shawn’s Response] I did point out that Image Renditions were updates to thumbnailing.  I understand that there are some additional capabilities with redition; as a developer, it does open up some possibilities that weren’t available OOTB in 2010.  However, I question, as the average IT buyer, how much value the feature adds over thumbnailing in the context of WCM.  For me, it’s not hugely consequential across the range of sites potentially “powered by” SharePoint (i.e. for product-oriented sites, image renditions could be very valuable; for the average marketing site, not as much – or at all).

   [Waldek’s Response] Device Channels in SharePoint 2013 are nothing like the mobile capabilities of SharePoint 2010. In the previous versions of SharePoint, whenever it detected you were browsing from a mobile device (detection was done based on User Agent-substrings and properties from the .browser files on the server) it always served you with a predefined experience that was challenging to customize. The only customization capability here was to use Control Adapters (custom code) to modify the rendering of various pieces of the website. In SharePoint 2013, with the introduction of Device Channels, this whole situation has changed. Not only we can distinct between multiple channels, but we can also use different Master Pages and content panels to optimize how the website is rendered on mobile devices. I also don’t agree with you when you say you need a whole separate website for mobile. In my opinion, when optimizing content for mobile you have three options: you can use responsive web design, build a mobile website or build a mobile app. With SharePoint 2013 you can make use of all those options (or a combination) but neither of them is a must.
   
   [Shawn’s Response] Again, my point was that mobile channels are an update and not a radical departure from 2010 capabilities.  I would argue that the variations feature, combined with the agent detection provided roughly the same capabilities as the 2010 mobile channels functionality (including control over master page and layouts).  Is it the same, no.  Are device channels easier to manage and control, perhaps.  However, for whom are these features relevant? I would argue they’re mostly for developers, not business users or even “power users.”  In the same way, I’m evaluating value to the end buyer.  This functionality will be useful to only a narrow range of customers and site within the context of WCM tool buyers.   Lastly, mobile channels are not available in all versions of SharePoint, most notably in Office365.

4. Would you say that improvements in WCM are evolutionary or revolutionary?
   
   Some of the new WCM capabilities provided with SharePoint 2013, such as SEO-related features or Rich Text Editor improvements, evolved with the product. As SharePoint matures more capabilities are added and the existing capabilities are being improved. On the other hand there are also some capabilities like search-based publishing that are new and that offer us new ways of thinking about how we manage and publish web content. Using search in SharePoint 2013 we can now build more dynamic websites where the content is displayed not only based on relationships known to content managers but also based on information that we have about the visitors ensuring that they see the most relevant content that the website has to offer.
   
   
5. Given SharePoint’s fairly heavy infrastructure requirements (e.g. disk space, RAM, number of servers to run the various services), do you believe that SharePoint can effectively compete with other, more pure-play WCM tools (e.g. SiteCore)?
   
   The exact hardware requirements depend on what you are exactly trying to achieve so even when you see some high numbers, it doesn’t mean that this is exactly what you need just be able to publish a website on SharePoint 2013. The only way to really determine what you need is to get a clear understanding of your requirements, test your website against them and make an educated choice about what kind of infrastructure you need to fulfill those requirements. Given the changes in how SharePoint 2013 is licensed for public-facing websites as well as what the rich Web Content Management capabilities it offers, SharePoint 2013 is a very interesting candidate to consider for choosing a web platform – even for organizations that don’t have SharePoint implemented yet. Compared to the past, licensing of SharePoint 2013 is very attractive now. Not only the price per server is lower but you also get best-in-class search capabilities as a part of that license.
   
   [SHAWN’S NOTE]  I’m not sure I agree that pricing is improved over 2010 and certainly the “it depends” answer on hardware is an appropriate response (I use that answer frequently).  However, “best in class” search is something you, as the buyer, should evaluate in the context of other available search technologies.  The Real Story Group has very deep research on Enterprise Search Technologies that will help you understand your choices and the trade offs.
   

I want to thank Waldek for participating in the discussion and hope that it has helped you, the buyer.

About Waldek

Waldek is Microsoft SharePoint Server MVP and works as SharePoint consultant at Mavention. Waldek participates frequently as a speaker and expert in community events such as SharePoint conference in London, SharePoint Connections, SharePoint Saturday and DIWUG. Recently Waldek became a Virtual Technology Solutions Professional for Microsoft Netherlands. In this role he answers customer questions around SharePoint Web Content Management (WCM).  Blog: http://blog.mastykarz.nl  Twitter: http://twitter.com/waldekm

Fixing Azure Challenges - Part 1

Over the last year, we’ve built a few Azure-hosted applications (including our own web site).  Throughout that time, we’ve hit a few walls, suffered setbacks and learned a ton about how to work with both the service and the Azure SDK tools (including the sometime challenge of dealing with version differences with both the SDK and the Azure management interface).

Since we’ve been helped by the many folks that post to twitter and their own blogs, we’re hoping to return the favor to the larger community.  Here are a few of the initial challenges we ran into and how we solved them.  In all cases, your mileage may vary.  Certainly if you have feedback, a better way to do things or something else (related), please do comment.  We’ll be publishing a follow-up post with additional solutions in the coming months.

NOTE: This post was written with the assumption of Visual Studio 2012 and Windows 8, along with the v1.8 release of the Azure tools.  We’ve seen posts referencing similar problems with older tools, but these tips may be less helpful unless you’re using the same versions.

Error when publishing applications

This particular issue is fairly non-descript, since it encompasses a whole range of causes and subsequent actions on your part.  However, problems with publishing happened frequently enough that it caused some serious delays in getting new builds of applications out to the cloud.  The following are the ones that caught us most frequently.

Administrator Rights

In order to debug or publish, you need to start Visual Studio “as Administrator” or you’ll get an error stating that the current user has insufficient privileges.  The solution is to right click on the Visual Studio shortcut and pick “Run as Administrator” to fix this problem.  If you’ve pinned VS to your task bar (like I do), just right click once on the task bar item, right click again on the main Visual Studio link and then pick “Run as Administrator.”

Out of Memory

When either trying to publish your solution directly to Azure or simply trying to create a package, you may receive an error that is simply “Out of Memory Exception.”  Apparently, with only 4 Gb of RAM available, Visual Studio is unable to create the Azure deployment package.  It’s really unclear to me why so much memory is necessary, but often without a clean restart of Visual Studio, I get this message.  The solution is simple, close down all other applications, restart Visual Studio and try again.  I can occasionally get away without these steps, but if I’ve been debugging or opening lots of other applications, I can guarantee my publish/packaging operation won’t work.  I never had to restart Windows, just exit VS and start again.

Unknown Error

In the midst of publishing a package (after initiating the “update” on a cloud service), you get an error from the Azure web-based management console that states your update was unsuccessful.  For me, this usually happened after several minutes of waiting for the package to upload, then cycling the instances and, at the last minute, erroring out.  When pressed for additional details (by clicking that not-so-helpful “I” icon), there are no details.  This happened to me recently and it was the result of having both http and https endpoints defined, but without the SSL certificate referenced by the service configuration setting loaded into the target subscription (the client we were working with used multiple subscriptions to separate “production” from “QA”).

In this case, the production environment had the SSL certificate loaded, but the QA environment did not.  Because my Azure settings listed the thumbprint for the SSL certificate and both end points were defined, the update failed in QA, while successful in production.  Unfortunately, it was difficult to figure out the root cause since the management console didn’t give any details. 

Only after I looked in the log, click on the “failed” entry and then clicked the “I” details button at the bottom of the interface, did I realize the missing cert was the culprit.  Embarrassingly, it took a call to Microsoft support to figure out that the additional detail was indeed available, but double-clicking the log entry doesn’t work.  You have to click  on the details button at the bottom of the page to get additional XML-based information regarding the entry.

If the SSL certificate isn’t your issue, the logs should help you figure out what is and provide more insight into the failed update.

 

Unable to manage subscription

More often than not, being able to manage a subscription is due to the lack of a management certificate, the fact that your Microsoft ID (aka LiveID) hasn’t been granted co-administrator rights (assuming you didn’t create the subscription) OR you’ve got a filter set.

The last one is the easiest to fix.  You set your subscription filter at the top of the management interface.  Just click the little funnel and then check or uncheck the subscriptions you want to see.  Occasionally, you’ll forget that you’ve unchecked a subscription you’d like to see OR someone will add you as co-admin on a subscription you need to work with, but don’t see it because of a previous filter setting (don’t ask me how I know this).  So, if you can’t see the subscription, just check your filter settings.

If you’re not filtering your subscriptions, the owner/primary administrator may have not added your Microsoft (Live) ID to the subscription.  In this case, there’s nothing you can do, but ask them to add you.

Finally, Azure requires various certificates to be loaded into the subscription to allow you to perform various functions.  These certificates are shown in two places: 1) in the settings section under “Management Certificates” and within Cloud Services under “Certificates.”

In both cases, you can upload your certificates manually through the interface.  However, the certificates have to be .CER type, not the .PFX (read about the different certificate formats for Azure here – NOTE: the older management interface for Azure is shown, but the basics should still work).  For the certificates required for Remote Desktop, you can create them dynamically through the Azure tools in Visual Studio.  This is done by choosing the “Configure Remote Desktop” option after right-clicking on your Azure Deployment project.  Clicking on the dropdown list box will show you existing certificates or allow you to create a new one.

After you publish, the new RDP certificate will be uploaded automatically.

Deploying different configuration settings

The Visual Studio tools are pretty decent and give you a degree of flexibility for deploying various environment-specific settings to your Azure subscription.  However, getting the hang of setting the various values took a bit.  Hopefully, if you’re challenged as I was, the following will help.

Generally, when developing an ASP.NET application, you have the option to create various build configurations directly in Visual Studio.  Each configuration is accompanied by a web.config transformation.  Each transformation is a derivative of the core web.config file, with specific XML-transformation directives to update, replace or delete various nodes within the file.  With Azure, these build configurations are handled exactly the same way, except that when you build a deployment package, you pick the configuration you want at the time you build the package; you don’t have to change the Configuration Manager setting.

In addition to build configurations, there are Azure-specific “service” configuration settings.  Much like adding Web.Config transformations, you can create various Azure service configurations, which allow you to stipulate variables like a blob storage account, certificate thumbprints, number of instances or what SMTP server to use (very similar to the AppSettings node in web.config).  Azure service configurations and build configurations are complementary.  However, build configurations (and the associated web.config settings) can only be changed when you upload a new build to the Azure environment.  Conversely, some service configuration settings can be changed through the Azure management interface (after deployment or at run-time), making them more flexible.  This is particularly handy when you have to make somewhat trivial application setting changes that would require a re-publish if the setting were housed in the web.config file (yes, yes.  You could RDP to the box, but it’s still easier to access the Azure Management console).

The various service configurations are managed through the Settings interface for your Azure project.   Set the drop down to “All Configurations” to establish the range of settings across service configurations.  If you click on Settings, for example, you will then see the range settings for your Azure deployment.  Each setting, while in the All Settings, any configuration setting that is configuration-specific is shown with a “select configuration” value.  Any setting that is common, simply displays the value.  While in “All Configurations” you can add or remove settings across all configurations. 

To change a setting for a particular configuration, just change the drop down value to the configuration you want and then click in the “Value” field of the setting.  Enter the new value and you’re done (don’t forget to save).  NOTE: You can add or remove settings within a specific configuration, so you’ll need to plan your setting approach so that all settings make sense for all configurations.

When you package up your application, you’ll be prompted to select the build configuration, as well as the service configuration. This will combine the specific web.config transformations along with service settings for the deployment you want to publish.

More information can be found on Microsoft’s site: http://msdn.microsoft.com/en-us/library/windowsazure/ff683672.aspx 

Finding that cursed subscription ID

As Azure has evolved, Microsoft has updated the management interface.  Where finding the subscription ID was relatively simple in early versions, later versions “hid” the subscription ID (for certain workloads) until after you publish for the first time.  If you’re also having trouble publishing, this is a double-whammy.

In truth, the subscription ID is always visible, though I’d argue the interface makes it a challenge to find if you’re unfamiliar with the management console.  To find you subscription ID, just log onto http://windows.azure.com with your LiveID.  Then click on the Settings menu item on the left (it’s the little gear thing).  Now, click on Administrators on the top menu.  This is the list of all administrators for your subscription.  Next to each individual, you’ll see the subscription name and the ID.  Just copy the ID from this interface.  If the column isn’t wide enough, just click and drag the column separator on the right.

 

Hopefully these tips will help you avoid challenges with Azure.  However, I’m certainly interested in your feedback.  If you have some, please comment.

Other Resources

Visual Studio and Azure: http://msdn.microsoft.com/en-us/library/windowsazure/ee405484.aspx

Setting up Remote Desktop for Azure Cloud Service: http://msdn.microsoft.com/en-us/library/windowsazure/ee405484.aspx

Running multiple web sites in an Azure web role: http://msdn.microsoft.com/en-us/library/windowsazure/ee405484.aspx

Collecting and logging data by using Azure Diagnostics: http://msdn.microsoft.com/en-us/library/windowsazure/ee405484.aspx

Being a Professional

I was reminded about what I’d had been calling “a professional” by David Heinemeier Hansson at 37Signals.  He posted “Your Life’s Work” on their Signal vs. Noise blog. It was a terrific piece that, to me, really spoke to a “disposition” that professionals should possess.  I especially liked this quote: “[if] you’re not committed to your life’s work in a company and with people you could endure for decades, are you making progress …”  In essence, you have to be truly interested in the work that you do and care about the outcome – not just today, but tomorrow and beyond.  If not, as David puts it, you’d have to question whether you’d making any career progress. 

For Consejo, this concept (of a professional) is at the heart of who we want to hire and I’m glad we’re not alone.

Intel and the Big Search Box

I was recently researching new ultrabooks and visited the Intel site for some insight on current processors.  When I arrived, I was  a bit shocked at what I saw…

Source: Intel.COM

Waiting for me  was not the typical corporate web site, with menu-based navigation (at least not immediately apparent).  Rather, Intel was presenting, dare I say, a very "Google" like approach – enter a keyword-based query to find what you want. 

I found the whole experience a bit off-putting.  I don't generally recommend our clients lead with a search-based content findability approach. In fact, I've argued against even making search a primary content findability technique.

If anyone has insight on this approach (e.g. real analytics supporting this experience strategy over more traditional IA), I'd love to hear about it.

Interesting cause for exception: System.InvalidOperationException: $metadata Web Service method name is not valid.

Interesting cause for exception: System.InvalidOperationException: $metadata Web Service method name is not valid.

While deploying a classic web service to a production server recently, we came across an issue when calling the service from a test web application.  We were able to add a web service reference to our test harness web application with no issues.  The web service would load with no issues in a standard web browser as well by url:

When calling the web method the response would come back with no errors.  We noticed the data was not being saved to the server as expected.  After checking the event viewer on the server we found this error that made no sense to us nor offered much detail:

 System.InvalidOperationException: $metadata Web Service method name is not valid.

After searching the web a bit, we learned this was more so a mask error that covered up the true issue.  After digging around for several days, I attempted to test the service directly in a debug state in the server's browser.

The service responded with:

Could not load file or assembly 'Telerik.OpenAccess, Version=2012.2.816.1, Culture=neutral, PublicKeyToken=7ce17eeaf1d59342' or one of its dependencies. The system cannot find the file specified.

We somehow published a version of the telerik.OpenAccess dll that was different then what the service was looking for.

After re-publishing the service with the correct dll version, the issue completely resolved and the web method started saving data to disk.

Lesson learned, the meta data error that originally prompted our research had no relationship to the real exception.  If you come across this error in the future, debug the service straight against the problematic server's web browser to get the true exception.

Interesting Code Exception–UnwillingToPerform

I have recently been busy putting the finishing touches on a new web application for a client in Indiana.  During the course of the project, we discovered that we needed to write a custom membership provider that would use secure LDAP to authenticate users (the ActiveDirectoryMembershipProvider does not, surprisingly, support the client's environment setup). 

While developing the basic provider was relatively easy to complete (and used an Oracle LDAP provider we built previously), we ran into a series of challenges.  One challenge is now a support incident at Microsoft (more on that in a later blog).

However, during one of the debugging sessions, we noticed an "unspecified operation error occurred" exception being thrown.  Digging down, we discovered the reason code: "unwilling to perform"

I have never thought about putting that kind of reason in my code, but if Microsoft can do it, perhaps it's O.K.

Write Better

When I started working as a consultant in the latter half of the 90's, our firm had writers that specialized in writing for the web.  These folks constantly had to help clients refine and organize traditional print copy for the "new" medium.

Recently, Jason Fried over at 37Signals published a short blog post that referenced a post by Maria Popova at BrainPickings.org highlighting writing advice from David Ogilvy, the original "Mad Man."  What I found fascinating is that the advice that Ogilvy delivered in 1982 is much the same advice our writers gave clients more than a decade later (in the context of writing for the web).

Clearly, good advice never goes out of style.