Drupal

We selected Drupal as the platform for these sites because of its reputation of having a solid foundation, allowing large flexibility through modules, and having a huge developer and user community. There were also members of our team who had worked with Drupal on other site projects and could bring their experience with the platform to bear here. The fact that it’s free certainly didn’t hurt either!

But in looking at the features and functionality of Drupal, we encountered a problem. There are several ways to build out a site hierarchy and assign user permissions at each level of the hierarchy, but they don’t scale particularly well. Perhaps our biggest frustration with our current CMS, as site administrators, is how much time we devote to setting editing permissions. We wanted to simplify this task and put it in the hands of our content managers.

Monster Menus

Amherst College noted this problem as well when they moved their website into Drupal several years ago. Fortunately for us, they developed a solution named Monster Menus which adds both a way to build a site hierarchy and to assign permissions at each level. We’re hugely in debt to their work on this feature, and to the assistance of Dan Wilga who helped get us set up with a version of it.

The screencast linked to above does a much better job of explaining what Monster Menus is and how it works than I could hope to do in this space. In addition to the permissions features demonstrated there, Monster Menus and its related modules offer several other key features when dealing with a website where you know you’ll have thousands of pages and potentially hundreds of content editors:

• Lets you define the node types (think: page types like FAQ, Profile, News Article) that are available at any level of the site. This helps guide the content editors so that News stories aren’t showing up in the Course Catalog.
• Defines a content treatment for “sidebars” that automatically cascade down your site hierarchy. This way the Giving site can have a sidebar button to the Give Now form, create it on their site homepage and have it display on the right side of any giving page.
• A way to do specific scoped searches of content across your site.
• A content type for displaying an automatic list of sub-pages of any page on your site.
• Through the Media module, a file browser that lets you easily upload and include attachments, organizing them within your site hierarchy.
• A content type to display RSS feeds from any source, including feeds from pages or tags within the site.

Amherst also developed a full profile management system, named Amherst Profile, that is connected to their administrative database system and offers a sort of portal for members of their community. At this time, we decided not to implement this feature for our sites both due to time constraints and because we’re answering different questions with our sites than they answered using Amherst Profile.

Again, I want to thank Amherst for sharing this module with us. I don’t know how we would have been able to implement all the features it offers within our project timeline and have the interface come out as clean and easy to use as it is in Monster Menus. We’ll be working with the development staff at Amherst to do continuous improvements to this module, though much of that work will need to wait until after the Middlebury site launch.

The MIIS Home Page

I want to take you through each of the page designs for the MIIS site and explain how they work. This begins with the Home Page. This page has three main content areas: Be the Solution, News & Events, and the MIIS Spotlight. Be the Solution is a list of terms from a custom defined taxonomy. I used a function to fetch the taxonomy dynamically so that any additional terms will be immediately reflected on the home page. The links to the term pages are then based on the term names. MIIS can choose to grow or reduce the themes their institution highlights on the main page. During site development, their Web Strategy Team came up with an initial list of 17 themes (which they considered incomplete), then reduce the number to 5, and currently display 12.

The News & Events and MIIS Spotlight features are two places where we used the Views module for Drupal to display site content. MIIS Spotlight shows a block view that selects randomly from the story archive any news item specified as an MIIS Spotlight. I used a separate field on the News Story content type to allow editors to select whether their node was a Spotlight or not. This could also have been done as a separate content type, but that would have been a duplicate of the News Story content type and less efficient to maintain. It could also have been handled using a custom taxonomy, though the MIIS site already uses several of these and the editing form for News Stories is quite long as a result.

The other areas of the page, such as the list of languages, links to social networks, address in the footer, links below each of the features (such as “Expore academics as MIIS”), and quick links in the footer were all set using Advanced theme settings. These allow site administrators, or others with the Administer Themes permission, to make updates to these areas through the site administration interface without having to contact a developer to update the page template. The user interface for doing this isn’t particularly user friendly, but the settings need to be changed rarely enough, and by people familiar enough with the site to figure this out, that it’s much less painful than the alternatives. Some of these, such as the Quick Links, could have been done as a custom block in much the same way, but for others like the News Feed URL it was more easier done with theme settings and so I decided to keep both settings in the same interface.

Theme, or “Be the Solution” Pages

Examples: Business, Environment, Translations & Interpretation

We originally called this the “Theme Template”, but as you can imagine this got really confusing when we starting throwing around the word Theme in its normal web development context. The thinking on each of these pages was that they would be an aggregator of content related to each theme. Since we used a custom taxonomy to build the list of themes, we could tag various nodes in the site with each of the terms and they would appear on these pages in some sort of custom defined columns.

The problem was that we didn’t have a good way to get those columns. Some of them, like Degrees & Programs, indicated that the content nodes would be in the Academics portion of the site, probably nodes with the Page content type, which we would need to distinguish from nodes in the Careers section of the site using the same content type, tagged with the same term. At the time, Monster Menus didn’t have a way to select nodes tagged with a term in a particular portion of the site tree, so we worked with developers at Amherst to add that feature. I’m hoping later to add functionality like this to the Monster Menus Views integration, so that a view can be passed an argument to select nodes within a portion of the site tree.

This worked well for a time; MIIS editors could enter paths to areas of the site, we would fetch that page’s title and use it as the column title, then select a certain number of nodes from that area of the site tagged with the term for that theme. This works really well if you’re selecting news items, or a photo stream, or other content this is refreshed periodically because the theme page will always show the latest nodes tagged with that term. It works really poorly when used to select a list of degree programs where the institution needs to place the items in a defined order. To solve this, I split out the Degrees & Programs column into its own field using the Link module for CCK. We can now enter the links to degrees in whatever order is required and then specify any number of additional dynamic columns.

More to Come

This is getting to be quite lengthy and there’s a lot to discuss about the MIIS site. I’ll end the post here for now and make this a series of posts on the topic. Check back shortly (or subscribe to this blog’s RSS feed!) for a new post on the section pages, course and faculty views, and our struggle to come up with the perfect sub-page list bucketing algorithm.

In the mean time, if you have any questions about the MIIS site and its technical implementation, leave a comment.

And here are the blogs ranked by number of posts:

As there is not currently a place in the information architecture for the Middlebury web site for Curricular Technology, per se, it will be up to the LIS Website team to determine an appropriate place and confirm the location with the Curricular Technology team.

Existing Infrastructure

We’ll take the current LISt blog, rename it to just “LIS” and grant all staff in LIS Editor access to the blog. Student workers and LIS-related people like department advisory groups will have Author access granted as needed.

Porting Content

Users and groups with existing blogs may move their content into the LIS blog by going to Tools -> Export -> Download Export file in their blog, then Tools -> Import -> WordPress -> Upload file and Import. After doing so, they should go to Posts -> Categories and use the Category to Tag Converter to change any custom categories their blog had into tags on the LIS blog.

Categories

The LIS blog will start with this set of categories:

• Audience
  • External
  • Internal
• Areas
  • Academic Consulting Services
  • Collection Management
  • Enterprise Technology & Infrastructure
  • LIS Administration
  • User Services
• Institutions
  • Middlebury
  • MIIS
  • Language Schools
  • Schools Abroad
  • Bread Loaf
  • MMLA
• Teams
  • Area Directors
  • Curricular Technology
  • Digitization
  • LIS Website

All other existing categories in the LIS blog will be changed to tags. In general, we’ll encourage the use to tags to mark things like posts that have photos in them, posts about particular workgroups, posts about particular projects, or other things that tend to have a more temporally mutable quality to them. Categories will be used chiefly for broad categories that change infrequently.

Private Blogs

The LIS blog will not be used to store private content. Groups, such as ACS, who wish to have a private blog should continue to use and maintain their blog in its present form or create new private blogs if one has not yet been set up.

Internal vs. External

Posts that are likely to be interesting only to LIS staff should be marked with the Internal category. Examples of this content might include a notice about LIS goal setting. Note that these posts would still be readable by the public, but that those subscribed to the public feed wouldn’t see them in the feed. Posts that are likely to have broader appeal, like a cookie party for students, should be marked with the External category, or both the Internal and External categories.

There will be links to both the External and Internal feeds on the blog homepage with a description of the content featured in each so that site visitors can choose how they will read about LIS.

At least, it’s been my hope since the beginning of this project that people cared a little about this.

Still, sometimes it takes a conference room of 30+ people to convince me that I’m wrong about something. This afternoon we met to discuss a possible solution for implementing a personal portal experience on our new website. There was plenty of criticism for the proposed software and I didn’t take it very well. During the meeting and then later in the afternoon I advocated a compromised position: we should implement the parts of this software that dealt with creating personal web spaces and profiles and ignore the rest. After a few hours of reflection, and a few beers, and a brandy, and a glass of port, I’ve come to a different conclusion: we should scrap the whole idea, look back at the features people either requested or were prematurely promised and develop a method to get that stuff working on our new site.

Here’s what the application we looked at this afternoon offered:

• A directed announcements page
• A static list of constituent-oriented links
• A RSS feed reader
• Personal web space for all users
• User profiles fed form a central database like Banner
• Alumni gateway features like a career network

Here’s what we’ve actually been tasked with delivering:

• Professional profile pages for Faculty and Staff
• A page of personal bookmark links
• Maybe some sort of iGoogle thing. That’d be really cool!

That’s it.

Let’s deal with each of these point-by-point.

Professional profile pages

We already offer several systems for building out your personal web space. Many users have web-accessible directories on our old UNIX server where they can upload HTML documents. I got one of these accounts when I was job shadowing one of the system administrators at the college in 8th grade and then used it again when I took an intro CS course at the college while in high school. The server’s been upgraded a couple of times since then and now runs some version of linux, but it’s basically the same set up. If people want a bit more, they can create their personal web space on Segue, our custom developed LMS, or here in Wordpress, or as a user page in one of our wikis, or as a profile posting on our current CMS, or in a couple other really specialized systems. We will very likely soon also offer Google Apps for Higher Education, which will allow people to build out their personal site using Google Sites.

But the idea with these professional pages was that people would want to present information about themselves in a highly structured format, integrated with our new website and content management system. Faculty might want to post their CV, academic history, and current courses on a page that as the same look-and-feel as the rest of their site, delegate permissions to allow the academic coordinator for their department to help them update the information, and provide some dynamic content from our campus directory or course catalog. Doing this requires more of a thought out, structured system than we’d be able to offer with basic HTML document web hosting or even Google Sites. The idea was, therefore, to create a basic web space for everyone within our new CMS and let them build out their site in this space.

This creates some rather obvious problems. If this is just basic web space, how to we integrate it efficiently with the course catalog or online directory? How do we automatically enable departmental colleagues, such as the academic coordinators, to assist in building out the content for these profiles, as they have done in the past? If we want to add a piece of functionality to every one of these professional profile sites, can we do it in an efficient manner? Can we aggregate this content easily, knowing that our central database system doesn’t tell us which people belong to which departments or workgroups? How do we reuse this content in other areas of our site?

If we were to just turn on really simple web site editing for all users, very few of these things would be possible.

Instead, we should do what we have already begun to do for Monterey: use a specialized template in our new site for professional profiles that includes fields for all of the information we want to know about the person and might choose to use when segmenting or aggregating the information. We should then have these profiles pages created within the departmental or office sites relative to these people. We can then use the power of our new CMS, Drupal, and its Views module to create pages that allow the user to quickly and efficiently find the people they want to know about. An example of this already in action can be found on the development version of the MIIS site (here’s a link that will only work on-campus and will break soon: http://miis2.middlebury.edu/academics/faculty).

We have never been successful demanding that anyone use a particular system and will, naturally, still offer all of the options for creating your personal web space that I mentioned above, but for creating the types of profile pages that we know need to be present on the official departmental pages, we’ll offer this option with encouragement and development support for integrating the “official” profile with the database systems that already contain much of the information we’ll be presenting.

A page of personal bookmark links

This is the most important feature of our new website for our internal audience. During meetings last winter, faculty, staff, and students were both promised that this would be available and told us that this feature was required. Accordingly, we need to find a way to make this work. The resulting system cannot feasibly be one where the user can simply click a button on any page of our site to add it to their personal bookmark list; such functionality requires an integrated portal product that we are not able to run due to pressure from faculty to run systems such as Wordpress and MediaWiki for their specialized uses rather than a single portal application like SharePoint, Luminis, or the custom systems developed at Amherst and Gettysburg. However, that does not mean that we cannot offer customizable gateway pages with the following features:

• A default list of links
• A prominent form, encouraging users to log in and edit their list
• Drag-and-drop reordering of links on the page
• A quick webform where the user can paste a URL and enter a simple title for the link page
• Integration with our single-sign-on system, so if the user is already signed in to one of our web applications they will be presented with their custom link list immediately upon selecting one of the constituent gateway options from the global site footer

Some type of iGoogle thing

We know that our advanced, cutting edge users want to personalize their web space with drag-and-drop widgets, design their own homepage experience and have deep, interactive ties with our core systems. We have also heard, from several of the design vendors who visited us during our RFP stage as well as intelligent folks on campus, that rebuilding a local version of a really cool application like iGoogle is a fool’s errand. As soon as we’re done with our version of it, Google will come out with a really cool new feature that ours doesn’t have and people will just keep using Google’s version of the application.

Much as I want to believe differently, I have it on good authority that Google has both more and better developers than Middlebury College. We’ll likely never be able to offer a better version of iGoogle than they have already created, so we might as well use the instance they’re giving us. We’ve heard that the iGoogle that comes with Google Apps for Higher Education is considered “deprecated” by Google. Still, I remember a phone conversation that Chris and I had with a Google representative in 2006 who assured us that Google would soon be charging institutions like ours over $10k a year to use the Google Maps API without advertising and that this change was only “months” away. By my count, it’s been at least 40 months away at this point. Of course, we can never tell when a 3rd party vendor will unilaterally kill support for an offering that they’ve told us they’re killing support for, but we’re talking about something that we already know will be used primarily by our power users.

They can deal with it if Google decides to move on.

Welp

To summarize: I think I was incorrect about the necessity of implementing a local profile portal to serve the features our requirements gathering effort found last winter. When this post was still an internal email, I was going to request that I be given the opportunity to publicly apologize for  my behavior this afternoon in attempting to push this through rather than listen to the honest feedback coming from our staff and students present at the meeting. This blog post should serve as that apology. I think I get it now and rather than implementing services that I think would be really cool, will advocate for things that accomplish the actual goals you set for us in this project.

I recently emailed my boss saying that my time working on our website after hours like this was at an end, but I find I can’t really operate that way. If I didn’t spend the time to type out this post, this would just be gnawing at me all night and that’d be a worse punishment that taking the hour or so required to get my thoughts down in writing. While I’ve said that I think I’m now more correct in my thinking about how this should all work, I have to acknowledge that I was incorrect previously which could very well mean that I’m still laboring under misconceptions. Please use the comments field here to let me and the rest of the development staff for Middlebury’s new website know if I’m still off my rocker on this issue.

Sure. Just create a new folder in the Web Redo Project Managers area for LIS.

2. When will the next round(s) of Drupal training be offered?

A second round of test training will be offered to the Project Managers group starting the week of 9/21. LIS Website Team members may request to attend as may the LIS Project Managers, though it would be a good idea to ensure that any who do attend are caught up on the current process, workbook, etc. Also, all involved should understand that they probably won’t be able to immediately go back and edit their site in Drupal after these sessions. There will be training work shops and sessions offered later in October for content editors at which point people should be doing content editing in the Drupal CMS.

3.  Can we schedule a meeting with White Whale during their visit?

Probably. Joe is, at the (not so) very least going to schedule a meeting with Tonya, who is responsible for the IA & Content Recommendations documents. I have a feeling that most of our questions will involve her. Joe will try to have the others from White Whale present for this meeting, if schedules allow. We should have a list of questions we want to ask going into the meeting. They will be on campus the week of 9/14.

What else?

Here is a list of the LIS related links I found in each section:

Academic
Library & Information Services

Campus Life
Digital Music

News & Information
Web Services

Technology
Banner Web
Help & Support
Library & Information Services
Segue
Telephone Services
WebMail

• Our blog and wiki
• Project scope
• Timeline / time pressures
• Surveys, google analytics, search analysis, HEAT stats, etc.
• LIS Project Managers
• Brief overview of recommendations

Questions from the ADs and my paraphrased responses:

What can we do about time pressures?

I mentioned that we weren’t requesting peoples tasks to be juggled or reassigned, but we did want the ADs to know that some members wanted to spend more time on this project. I noted that we understood that the amount of time we spend on project work will vary during the course of the project. I can explain more during our weekly meeting.

What resources do you need [the ADs] to provide?

We don’t have specific resource requests at this time, though we may have more after our meeting with Jeff on the 20th. We will also ask the LIS Project Managers what additional resources they will require to work on their areas of the LIS website and will send these requests to the AD team as appropriate.

What about forms that send information to people in LIS? [book order form, helpdesk request, etc.]

I said that this was something we had not considered previously, but providing access to these forms does seem within scope of our project and is something we should consider in our recommendations. I promised to bring up this topic during one of our next meetings.

What about an issue/request tracking system on the web?

Developing/purchasing such a system is not within the scope of our project. However, should LIS acquire or develop this system, we may have recommendations on how best to provide access to its information through the web. As LIS does not have such a system at this time, developing these recommendations is not a priority for our work.

Where is LIS in the site information architecture?

The Library website will be part of the Academics section of the site. LIS will not have a traditional site on the CMS, but the LIS landing page may be accessible in the Offices & Services section, through Search, through Quick Links, etc.

I’m a faculty member and I want to know about the curricular technology options that are available to me. How do I find this out?

This is specifically the question that we are looking to the curricular technology team to answer, which is why we have identified that team as a Project Manager for a portion of the LIS website.

This is important because of the way Google Analytics works. The code to update the analytics data for the site is contained at the end of the body and is only run if the entire page loads. This means that when we do a redirect, the user’s click on that link is not recorded by our analytics tools. Because of this, you will see a 0 next to that link to MIDCAT on the LIS home page. It would be silly of us to assume that this link is never clicked on, but according to GA, it wasn’t. I’ve noted in the spreadsheet at the end of this post whenever a link on a page is a redirect. We could use standard log file analysis to determine the exact number of page visits to all of these, but as I explained earlier, we don’t have the log files because I deleted them all.

This is not to say that we should discount the results of the data here. There is still useful information in GA. However, when looking at the data, we need to take into account that there are links on each page for which we don’t have information.

Now on to the analysis! The following data is for the last year (July 26, 2008 – July 26, 2009). I figured that a full year would give us the most robust view of the site traffic.

I’ll start with the total page views for each major area of the LIS site. Unsurprisingly, the Library content is the most viewed, with the Contact, Quick Links, and Teach/Learn sections barely viewed at all in comparison. The Unique Pageviews bear out this pattern and the full table of data is available in the spreadsheet at the end of the post.

Two questions we agreed to ask of the Google Analytics tool were “What are the top five links on the LIS home page?” and “What are the five least clicked links on the LIS home page?” Here are the top five:

1. Library & Information Services (6539)
2. Articles, Indexes, Research Sources (6160)
3. Find Books, Articles & More (4289)
4. Middlebury (Home Page) (3175)
5. Course Reserves (3059)

Here are the five least clicked:

1. Content Provider (LIS website) (0)
2. P2P Music & Video Filesharing (0)
3. Telephone Services (0)
4. Digital Media Development (2)
5. Departmental & Workgroup Directories (3)

The full lists, as well as the lists for the landing page of each sub-site are contained in the spreadsheet at the end of this post.

One question that I was not able to answer was about click paths to the blogging and wiki sites. This question can’t really be answered for a couple reasons. First, we don’t have a full year of data for those other services and, second, links to those services would go through redirect posting which are flawed for the reasons described above. If there are sites within the LIS site that you would like to see click-path analysis for, please let me know specifically what resources we should examine.

Here are the broad recommendations I draw from this analysis:

1. The Quicklinks section gets little traffic. This should be removed from the site IA. A better designed site with the addition of a search landing page should not need this section.
2. Other than the link to the CTLR, the Teaching & Learning site is not used. Rather than try to have this information live in multiple places, any relevent content should be moved from here to CTLR and this sub-site should be replaced with a simple link to CTLR.
3. The Hours, Locations, Maps section is really the only think people click on About LIS to see. We should have this information on the LIS search landing page and move the remainder of the content elsewhere.
4. The Departmental Directory is not used, but the full Staff Directory is used. Eliminate both and replace with a link to the LIS search results in the central Directory, which is the same information as the Staff Directory. (Forthcoming improvements to the central Directory interface will improve this experience).
5. The Quick Phone and Email Contacts is the primary resource in Contact LIS. Eliminate the rest of the sub-site and add this information to the LIS search landing page.
6. The links in the top sections of the sub-site landing pages such as below “Need help?” on the Library site are rarely clicked. Avoid this interface in future site designs.
7. The “Did you know…?” sections of the LIS site are rarely visited. Consolidate this information into a central blog about LIS.

Here is the complete spreadsheet of results.

Here are the top 15 queries of the LIS search collection in the last year:

And here is the same for the LIS Wiki collection:

For comparison, and so you can see how infrequently the LIS search collection is used, here are the top 15 LIS-related queries from the “All” collection, which as you might expect indexes all the other collections:

(Note on this last one: midcat is the 16th term with 70 queries).

You can see the full reports for each, which include more information at these on-campus-only links:

Report for LIS Collection 07/24/2008-07/24/2009

Report for LIS Wiki Collection 07/24/2008-07/24/2009

Report for All Collection 07/24/2008-07/24/2009

I also sent an email to Chris Norris asking for assistance getting us some information from Google Analytics. He was out yesterday and today, but appears to be in his office next week, just booked straight through with meetings. I’ll keep you posted on this item. Here is the list of questions I sent him:

1. What are the top 5 search terms within the LIS? (I’ll get this from the GSA)

2. What are the top 5 pages on the LIS site?

3. What is the most common click path from /academics/lis to the LISt blog and the LIS Wiki?

4. What links on /academics/lis (the landing page) are clicked on the least?

5. Same as (4), but applied to /academics/lis/lib, /academics/lis/help, and /academics/lis/about.