Multilingual SharePoint Sites–Part 2

Some months back, I wrote a post about the basics of multilingual sites in SharePoint.  The post was a good primer for anyone that needs to understand SharePoint-centric concepts regarding multilingual web sites.  Unfortunately, the post didn’t really describe important details outside of the SharePoint sphere.  In particular, the post excluded all the ASP.NET-centric details.  In this post, I want to share at least some of those additional details.

Globalization

One of the early design goals that Microsoft had for SharePoint was that ASP.NET developers would be comfortable creating SharePoint-based solutions.  The theory is that if you’re a competent ASP.NET developer, you can simply pick up the additional SharePoint API universe; SharePoint is a good .NET citizen, so this idea shouldn’t be a stretch. 

Whether or not you believe a good ASP.NET developer could easily pick up SharePoint, SharePoint does borrow very heavily from many .NET facilities.  With regard to multilingual sites, this includes the Globalization namespace. 

The Globalization namespace is a group of classes that are responsible for allowing ASP.NET applications to understand the numerous languages and cultures that applications can target.  It includes everything from calendar differences and languages to date/time formats and string comparisons (and a whole lot more).  It also, importantly, provides a facility to allow developers to create a resource pool of commonly referred to assets (e.g. element labels, images).  These assets are all referenced using standard labels, creating an index of asset variants for each culture.  At runtime, based on the culture of the current user (usually indicated by a browser setting), the .NET framework will dynamically select the appropriate asset/resource based on a generic label describing that asset or resource. 

Resources and Resource Files (RESX)

Resources or multilingual assets are defined in a Resource File (RESX). There’s a resource file for each culture represented in the application. All resource files use the same labels to describe the asset, but with a culture-specific value.  For example, the text shown next to the text box where a user would enter their user ID to authenticate with the application would be an example of a reference resource. 

Figure 1 – Resource example for Login Page

In the RESX file, which is just XML, you’ll find the following entry

Figure 2 – Login_UserID label in the EN-US resource file

To add, edit or delete values, Visual Studio provides a “designer” view of the file.  In the designer, you have the ability to quickly and easily define the various labels and the corresponding values.  Figure 3 shows the Visual Studio interface for editing a RESX file.

Figure 3 – Visual Studio designer interface for RESX file

For every culture your application needs to support, you create a specific RESX file.  Each file would be named for the culture it supports and every file would contain the same labels, with values corresponding to the specific culture.  The example presented here, the RESX file is for the culture EN-US (US English). For more information on resource files, naming conventions and details on creating the files, take a look at this MSDN article on resource files. 

In SharePoint terms, the RESX would correspond to a specific variation of the same culture.  In effect, you will have at least one RESX file per variation.  These files define elements of the user interface that end users do not supply.  Whereas content on any given page is created and managed by content contributors, there are also elements, like the label for the User ID field on the login page, that are “baked” into the code of the application.  RESX files provide a mechanism to define what that label will say in the context of a specific variation or culture selection (usually set by the browser displaying the page).

Information Architecture and Visual Design

Beyond the somewhat mechanical processes for inserting culture specific content into a page, a more critical aspect of multilingual sites is Information Architecture (IA).  The IA defines the navigation paths (global navigation and its relationship to other sections and pages within the site) and the overall interface layout.  This means the decisions about where various interface elements are placed, what nomenclature is used and what sort of content is shown are all made through and by the IA (both the person – Architect – and their output – Architecture).

When developing a multilingual web site, consider that the interface will be at least slightly and potentially radically different based on the language being displayed.  The simplest example is word length.  If we compare an interface in German and one in English, it’s very likely that there will be different space needs for labels in the navigation, as well as content.  As a result, the IA must anticipate interface movement and allow for enough white space to accommodate an interface that will grow and shrink based on the language’s need.  This too is a challenge for constructing HTML and JavaScript, since both components of the web page may need to “react” to language differences.  However, beyond this relatively easy challenge, presenting content is matter of having the appropriate language-specific content.

A more complicated scenario is one involving differences in how a language is read.  For Hebrew or Arabic (as two examples), the languages are read right to left.  As a result, the whole orientation of the interface needs to shift.  The main navigation will need to start from the right, global navigation elements will be positioned in the upper left and text will flow from right to left within the content sections.  As such, you may require a unique master page and page layouts for these languages to sufficient accommodate the display differences.  The same is true for languages that are read vertically instead of horizontally as in Manchu.

Continuing with the above example, you also have the challenge of fonts.  Languages that utilize radically different character sets will require the IA and the designer to consider both font face and size choices. For example, for any font size choice in the cascading style sheet, will the text be readable across all languages represented by the site.  Most European languages have characters with relatively little detail compared with Asian languages.  Font sizes that are too small or font faces that carry too much embellishment may make detailed characters muddled or simply unreadable.  As such, these choices represent both a visual design and information architecture challenge, since font size differences will also present spacing issues to resolve.

Bringing it all together

With all of the details provided in the two posts of this series, here’s a quick review of the important parts:

• When developing your Information Architecture for a multilingual site, include the various cultures included.  Each culture, in SharePoint terms, will be a “variation.”  A culture, remember, is a combination of a language and a country, represented like EN-UK (English – United Kingdom) or PT-PT (Portuguese – Portugal).  This culture approach makes it easy to distinguish between two countries that share a broad language (e.g. Spanish), but differ in usage (e.g. Spain vs. Mexico).
• Developing an IA for a multilingual sites involves many more decisions and test cases to resolve than a single language site.  It’s important to explore the implications for your specific IA based on the languages that need to be supported and, when the visual design is complete, any challenges a specific design might pose based on the supported languages.
• Within SharePoint, decide what variation will act as a the “primary” or source variation.  This is the variation that will syndicate content to all other variations.  For example, if the source variation is German (from Germany), your content will start in German; once a page is approved, it will be copied to the other language variations in your site collection (e.g. EN-UK, PT-PT).  From there, each non-source variation will be responsible for translating, approving and publishing a language specific version of the German content.
• You will have at least one RESX file per variation.  If you have lots of different cultures, you will have as many RESX files and they must all contain the same labels.  Because labels are not evaluated during compile-time in Visual Studio, you’ll only discover missing (or conflicted) labels in a resource file at run time.  This is not, obviously, a good user experience.  As a result, you should thoroughly test and control the modification of RESX files.  Take a look at this blog series from Carel Lotz regarding one approach to effective RESX management: http://fromthedevtrenches.blogspot.com/2011/04/managing-net-resx-duplication-part-1.html
• Think carefully about the taxonomy (aka organization) of variations and labels.  As much as the IA process should define navigation, the overall taxonomy will drive label names and how labels are used in the application.  For the project example in this post, we used labels tied to interfaces (interface name prepended on label name).  This is one approach.  However, this approach neglects opportunities to leverage labels across interfaces.  Conversely, label use across interfaces can make maintenance more challenging as label changes will necessarily have different impacts across the application.  Here, experimentation and testing are key.
• A SharePoint multilingual site is really a combination of SharePoint variations and .NET globalization.  You must necessarily implement both; end users will leverage the variations component and your developers will have to provide matching RESX files for application-specific labels and static text.

As you may have surmised, there are a lot of details to consider when developing a multilingual web application.  SharePoint does provide decent facilities to enable basic multilingual sites and the .NET framework provides loads of flexibility in implementation.   Just be sure to consider the whole picture – it’s a combination of SharePoint centric constructs (aka variations), good information architecture/design and the technical “infrastructure” to make the whole solution work for end users.

Creating a Multilingual Sites in SharePoint 2010

As we start 2011 and reflect on 2010, one of the major initiatives for many clients was creating globalized, multilingual sites.   The message was especially clear when we spoke to attendees at events like the Enterprise 2.0 conference, KMWorld and J. Boye.  In particular, I was surprised at the number of attendees  planning or that have implemented “localized” or multilingual SharePoint applications.  Unfortunately, as I learned directly, the information available from Microsoft on all multilingual scenarios is uneven.
To their credit, Microsoft has done a relatively good job of making multilingual interfaces within SharePoint available.  Microsoft has released 40 language packs for SharePoint.  These language packs allow organizations to create sites and site collections in any of the languages supported by the language packs.    If your work involves intranet or extranet scenarios (externally facing collaborative sites for trust non-employees), the language packs will serve you well.  If you want a French or Arabic collaboration site, just create a new site and stipulate the language.  Individual users are free to contribute list items in the site’s language and see localized SharePoint interfaces.

NOTE: Tony Byrne, Founder of the Real Story Group, pointed out recently that one of the advantages of using software from larger software companies like Microsoft, Oracle and IBM is that there will be far broader and better tested multilingual support.  Many smaller vendors either don’t provide multilingual support at all or provide a far smaller set of supported languages.  Both issues could be particularly problematic for SharePoint deployments as many rely on 3rd party add-ons from smaller vendors.  You should do your homework on potential solutions to ensure they support the languages you will deploy.

Where documentation for multilingual sites falls short is in the internet scenarios – specifically creating multilingual “publishing” sites (think about the typical corporate web site).  There is some limited content in blogs, technet and MSDN.  However, this content either discusses language packs and variations abstractly or, in the case of most blog content, applies to SharePoint 2007.  All of it falls short of the quick reference needed by many organizations deploying publishing SharePoint sites. 
In an effort to provide some small assistance to other folks implementing multilingual sites, here is a quick reference list of sorts that provide details on some of the basic setup.
Variations
Publishing sites (SharePoint sites created using the “Publishing” or “Publishing with Workflow” site definitions) use a concept called “Variations” to enable the presentation of site content in multiple languages.  This concept has little to do with language packs and more to do with a basic copy process.  When you setup a multilingual publishing site in SharePoint, you designate a root site, a “source variation” site and other “variation language” sites.  As you create content in the source language site, a SharePoint timer job copies content (e.g. pages) and site structures (e.g. sub sites) from the source variation site to each of the other variation sites. 
The root site is merely the default URL for your web site; when web users navigate to the root URL, they are automatically redirected to the variation site that matches their browsers language setting.  As a developer, you can even change the logic used to redirect users to a specific language site by modifying the variations landing page called VariationsRoot.ASPX.  In doing so, you can enable SharePoint to handle other redirection processes like redirecting based on a previously selected language preference stored in a cookie or redirecting based on profile data (e.g. you know the user is based in Belgium and prefers French over Dutch).
Finding the Variations Setting and Beginning the Setup
There are two settings that control variations at the site collection level.  Both settings are found in Site Settings under Site Collection Administration.  To begin the process, you click on the Variations menu option.

By clicking on Variations you begin setting up how variations will be configured for your site collection.  In the subsequent interface, the “Variation Home” is what is called “Root Site” in this post (and what Microsoft typically calls it as well).   The variations interface is shown below:

Variation Labels
Once you’ve setup the global variations feature, you need to setup each variation label.  Variation labels represent every variation site you want to include in your public web application, including the source site.  For example, if you want to have a French (France), German (Germany), English (UK) and Arabic (United Arab Emirates), you’d need a total of four variations.  Of those variations, you’d pick one to be your source variation; the rest would receive content created in the source. 
When you create the labels for the first time, you start by creating the source label.  In our example, we’ll call our UK English site the source.  There are a few fields to complete:

Field Name	Description	Example Value
Label and Description	This is essentially the portion of the URL that designates the site. If your web site URL was http://www.mywebsite.com/ the label portion would be inserted after the last slash like http://www.mywebsite.com/[variation label].  In most cases, the label is going to be the two-letter ISO code for the language, followed by the two-letter ISO code for the country.  Wikipedia has a complete list of the current ISO language codes  and ISO country codes.	en-gb
Display Name	This field is what will be used to not only describe the variation in the management interface, but it will also be the default name of the site when the variation site is created.  Typically the display name is the localized title of your web application.  If your web site was called “My Web Site” in English, you would enter that value for our UK English site.  For French, you might enter a display name like “Mon site Web”	My Web Site
Locale	Use this setting to specify the language represented by a specific variation.   There are numerous languages show in the drop down and potentially multiple versions of any given language (e.g. Spanish in Spain vs. Mexican Spanish).  You should choose the appropriate language that fits your needs.  This setting is used by SharePoint to match a user’s browser language setting to the appropriate variation.	English (United Kingdom)
Hierarchy Creation	This setting determines what portions of the source variation SharePoint will replicate to other variation sites.	Publishing Sites and All Pages
Source Variation	This sets whether the variation you’re currently defining is the source variation.  This setting is something that is set once on a specific variation and can’t be changed after you’ve created the hierarchy.  This option also sets the specific site definition used in creating all variations.  NOTE: only publishing site definitions (e.g. Publishing Site) are there by default.	Publishing Site with Workflow

Here’s a screen shot of the form in SharePoint:

Other Language Variants
After you’ve setup your source variant, you setup the other language variants you want to represent.  The setup process is similar to setting up the source variation, except that the check box indicating the variation is the “source” will be greyed out and disabled.
Once you’ve completed creating the other variants, click the Create Hierarchies button to begin the creation process.  NOTE: Creating the hierarchies is handled by the timer job.  As a result, you will either have to wait for the timer job to run or manually initiate the job to see the variations.  This is also true of other timer job-dependent activities like site structure and page propagation.   More information on timer jobs is provided below.
Timer Jobs
Technically there are multiple timer jobs that each handle a different aspect of the propagation process.  Each timer job runs on a different schedule.  Here is the list of variation timer jobs and the default interval:

Timer Job	Description	Default Execution Interval
Variations Create Hierarchies	This job creates the hierarchies defined in the variation labels section of the variations settings.  As you add new labels, this job takes care of creating the corresponding variation site	Every 24 hours
Variations Create Site	This job creates variation sites if the automatic creation setting is disabled in the Variations section of the site collection settings.	Every 5 minutes
Variations Propagate Site	This job creates variation sites if the automatic creation setting is enabled in the Variations section of the site collection settings.	Every 5 minutes
Variations Create Page	This job creates new pages in each of the variation sites once the page has been created in the source site.	Hourly
Variations Propagate Page	This job updates existing pages in the variation sites once an update has been made to a source page.	Hourly

Depending on your needs, you can modify the default interval to match your publishing schedule.  For example, if you publish pages frequently, you may want to increase the timing of your page propagation and creation to 30 minutes instead of an hour.
Keep in mind that there’s a version of these timer jobs for every SharePoint application (think URL).  This means that you’ll have 3x the number of application configured for variations to manage.  Thankfully, Microsoft provides a view of timer jobs per application.
A screen capture of the time job definitions for a sample site is shown below:

Additional Notes
Like much in SharePoint, the choices you make should be based on your specific requirements; there is usually more than one way to satisfy a need and multiple ways to develop a solution.  To that end, here are some notes about variations setup to help you make these decisions.

• Variations can be used for more than just language versions of content.  Microsoft documentation points out that the logic to redirect users to specific variations could be used just as effectively to redirect a client to a mobile device.  In this case your label could simply be “M” for a mobile version of your site.
• The timer jobs do not replicate other lists or libraries from one variation to another.  This means that you must create lists and libraries in each variation manually after the hierarchy is setup.  This is also true of content in those lists and libraries; the timer jobs focus on structure (e.g. sites and sub sites) and pages.  However, if secondary lists are important for your solution, consider creating a custom timer job.
• Variations are site collection bound.  This means that all variation sites are housed in the same site collection.  If, for some reason, you think you may need more than one site collection, you’ll have to recreate your variation setup for each site collection.  Each site collection would be managed independently.
• When possible, use features and feature stapling to modify the nature of any particular variation (if necessary).  For example, if you need a specific SharePoint list or library created in each variation, use a feature and staple it to each variation as it’s created to ensure the list exists.
• Content contribution is a bit trickier with variations, since each variation is essentially it’s own “web site.”  As such, you will need to spend more time training content contributors and more time supporting them.
• If you use the new managed metadata feature in SharePoint, you’ll need to provide translations for terms as well.  Since all variations are housed in a single site collection, you can even use managed metadata bound to a column vs. enterprise managed metadata to isolate terms that may be specific to your site.
• You are required to setup a root SharePoint site for the site collection.  However, once variations are setup, no one will likely see the root site (since the VariationsRoot.aspx page will be set to the home page of that site).  As such, you don’t have to do much with it.  However, you could use the site to store more global resources, like images and documents, for use across all variations.
• Especially during development, you should execute the timer jobs manually by going to Central Administration, go to the Monitoring section and click on Timer Job Definitions to find the right job.  When you click on a specific timer job, you’ll have the opportunity to run the job manually.  Manually running the timer jobs will allow you to see what’s actually happening within your sites without having to wait for the job to run automatically.

Please share your experiences with multilingual sites and what challenges you’ve had through comments on this post.  There is a lot more detail to setting up and managing variations, so the more feedback the better.

Continue to part 2 of this series