Drupalcon: Drupal.org Redesign Panel Discussion

Notes from redesigning Drupal.org

The Drupal Association decided on Wednesday night that their number one priority for the next year was to focus on planning, coordinating and raising funds for a redesign of Drupal.org.

In this panel, Nedjo Rogers talked about the aims and lead-up to the redesign. Tiffany Farriss discussed the scoping into an RDF, Angie Byron gave an update about the progress so far, Daniel Zhou talked about how his pivot module will aid in module evaluation, Derek Wright talked about improving the developer toolsets. Finally, Kieran Lal talked about the fundraising efforts, and then it ended with an open discussion.

NEDJO ROGERS -- Aims and lead up to the redesign.

Some of the aims for the redesign was that Drupal.org should continue to engage the community, getting through the RDF process, identify the remaining work at the implementation level and help promote a fundraising campaign.

Nedjo is a relatively new member to the Drupal Association (DA). The DA needs to support the Drupal community, and that is represented if we take a look at the founding statutes of the DA.

Last night, there was a meeting of the permanent members of the DA where the redesign of Drupal.org was decided to be the number one priority for the next year. This has to be DA's number one priority.

What are some specific goals for the Drupal.org websites? There are two major goal areas: 1.) Increase Drupal adoption and market share and to get the word out to various groups, and to showcase what's possible with Drupal. 2.) To increase the quality and capacity of Drupal development. Drupal.org is the primary contact for learning and carrying out development.

What is the Status of Drupal.org currently? It's evolved organically over the years. It's got huge strengths, extremely active user base and a unique setup with skilled and dedicated community managers, but at the same time there are growth pains and needs. The current site predates the exponential growth of Drupal, and there are clear pains and needs. So Drupal.org needs to build on and support the Drupal community, and so it needs to strengthen the current community.

TIFFANY FARRISS – Scoping of the Redesign

Tiffany is a strategist, and has a lot of experience with dealing with RFPs. So what needs to be done to get ready for an RFP for Drupal.org? The work that we need to do now is to look at the strengths, challenges and opportunities for Drupal.org and create a targeted RFP for Drupal.org. The RFP will be pointed for an IA firm with wireframes and to have a single design. Design by committee does not work, and they need to just pick a firm that has a aesthetic and IA vision along with a firm deadline to work towards. We already have some insights and user profiles, and we need to harness this knowledge and condense it into the RFP. We also need to be clear with who is going to help manage the project and who will be involved with the process. We need to get thoughtful and coherent responses from the firm for how they will participate with the Drupal community, and to make our home a better place for us and for other new people coming into the community.

ANGELA BYRON -- Progress so far

We've done some initial research, and we don't need a marketing firm to tell us who the community is. So who uses Drupal.org? What are the different personas?

1.) Mary the Manager -- She's told by the dev team that we're using Drupal. She needs to get into Drupal.org and figure out who's using it, how is it being used in various different sectors -- government, non-profit, commercial. How can she get Drupal to do what she wants.

2.) Wendy the Webmaster -- She's used to HTML developing sites by hand or even developed her own proprietary CMS system and got tired of maintaining it. Needs to know what modules are available, and where to get site recipes for how to build a wiki, etc.

3.) Danielle the Designer -- She's concerned about how to get Drupal not look like Drupal, and how to convert PSD file to theme.

4.) Dan the Developer -- He's got a CVS account and developing modules, and needs all of the latest developer news.

So we have to be sure that we don't have a redesign where the end result conflicts with the goals for different personas, and for example, make it impossible for devs to get the information they need.

So what about splitting up sites into subsites -- Right now, we have Drupal.org and groups.durpal.org, and so we're discussing splitting it up into more sub-domains.

1.) drupal.org 2.) downloads.drupal.org 3.) docs.drupal.org 4.) developers.drupal.org

This would allow us to have more modules on the sub-domain sites that are more specialized -- like having a wiki module on docs.drupal.org or the diff module from developers.drupal.org. It will also allow us to have a person or team maintain the themes area.

But there are also a lot of challenges with this approach -- like it's annoying that you have to log into to g.d.o and d.o. And no way to aggregate all posts across both sites, and not being able to search across all sites or view issues and docs in same search

There are still some issues and open questions: then we could have a panel page on the developer.drupal.org site that gives relevant information. So at the moment everything is shown in d.o., but this would allow more navigation and specific landing pages with more targeted info for each persona.

[Showing Dries' survey chart results] Dries also asked the community what should be improved on d.o. Finding modules sucks, search sucks and finding documentation sucks. And so these are the areas that we're going to need to improve upon.

We're planning on having a top-down process with the IA and design of the site because design by committee doesn't work in this case. We need to choose a firm that we trust. We're also going to have a bottom-up approach to come with our own solutions for module-finding and other pain points that we have.

So again, we're going to have two approaches: 1.) IA and design of the site by a small firm. 2.) And the community will handle the implementation of custom Drupal coding where a lot of them can be handled by DROP tasks where people can help get involved. It will help empower people to actually get things into their own hands and to help get stuff done.

KIERAN LAL -- This is actually a fairly complicated process, and can be a landmine issues when a lot of the work is done by volunteers. Specifically, what happens if you start paying people? Navigating this is a difficult issue, because you don't want to upset people. The first step was to make it the number one priority for the Drupal Association, which took 8-10 months of work. And we haven't found the DA lead for this project yet.

In about 3 months, there will be the d.o. doubling and our servers are predicted to crash again. But we're trying to avoid that with a lot of help from Narayan. If you do a redesign and add more features, then you have more load. So by May 15th, Drupal.org should crash again, and so they've added more servers to be prepared for that. But we've now built the political will, and also built up with what is going to be needed on the infrastructure backend.

DANIEL ZHOU -- Aiding module evaluation w/ pivots I'm going to present about pivots, which is an approach to make it easier to find and evaluate pages. Wow, there are a lot of modules, and it's hard to know whether or not a module is good when you're on a page, and it's also hard to find similar modules. You could use ratings or votes, and you could also use quality metrics like number of downloads. These haven't been implemented yet. The approach that I've been able to has been to mine information from the Drupal.org forums and to make recommendations.

Try out the pivots module on the project pages like on the scratch drupal site at: http://scratch.drupal.org/project/tinymce -- which is for the TinyMCE WYSIWYG Editor.

Notice that there's a block on the right-hand side that show related conversations -- ranked by the how recent it is. This makes it easier to see if the module has been active. You can post a message in the forum, and it'll show up on the module page. Another nice thing is that if you click through to the conversation, then it'll list the other modules that were mentioned in the conversation. So it'll be easier for you to get to other modules. The module that they're using is to find matches, for example the devel and image module. It will display the related modules on the module project page, and will rank other posts because TinyMCE and FCKeditor are mentioned in conversations over and over again. It'll be going onto Drupal.org real soon. This will able you to access a module, and find complement and substitute modules -- this will be available as a contributed module as well.

NEDJO ROGERS: This will not be the complete solution, but through the redesign process we're able to already create some improvements right away.

DEREK WRIGHT -- Improving developer toolsets Helps maintain the project.module. There are a lot of different problems with finding modules on Drupal.org -- there's 3000 project nodes right now, which was great for 2-3 years ago. But the explosive growth means that the current system can't deal with that. Now it's time to make my characteristic Drupalcon plea for help.

We need more tools evaluating modules -- [showing endless scrolling on the module page]

There are also a lot of other things that are very useful, and things that are in the works -- quality and popularity metrics, which are very valuable because they can be calculated automatically. Things like: * Does this module have any official releases? * Does it have any documentation? * How many open bugs are there? * How often are their CVS commits? * How many people have downloaded it? * How many sites are pinging back with the update status?

And we can display this information on the side, and it gets a score of say "15" and different people will interpret this info differently. Like just because there's not a lot of CVS activity doesn't necessarily mean that it's a bad module.

So it's not the end of the story -- like we probably need humans to give feedback as well. Like modules bundled into installation profiles and other site recipes.

But we need to help stop the bleeding first, and so take a look at this two issues -- http://groups.drupal.org/node/6186 -- which is an aggregation of the conversations about the project node redesign.

Also http://groups.drupal.org/node/7191 -- which is a page of project quality metrics -- assembled all of the discussions and gave his feedback for what needs to done and when.

Summer of Code -- Drewish worked on project module metrics and created a whole framework to plug into these metrics and wrote a plug-in that fits into update_status.module. It's all about ready to go, and they still have some discussions about how it should be displayed. Users should be able to click a node and go see a view of different metrics. Should also be able to a page and sort by how many people are using a specific module.

Last thing is that a lot of possible solutions, the main roadblock is that it is determined by gnarly code within project.module -- but now views exist, and there are 4-5 people who understand project queries and all of the insane project query code, but about 4000 people who understand views. So maybe we'll have a contest with all of the available fields from a database dump, and then let people create a number of different views from it.

So come to the code sprint on Friday to help finish Views-ifying the project.module.

KIERAN LAL -- Fundraising

Kieran first got involved in understanding Drupal through CivicSpace, and then OSCON in 2005, and laid out three things 1.) Need legal entity to take money and handle legal issues 2.) Need more usability testing 3.) Improve Drupal.org, and improve most important code of the project.module. 10,000 people cooperate via the project.module on drupal.org, and so it's vital to the community. Nedjo jumped in and maintained the project.module back in 2005.

Kieran then moved into fundraising role. Now let's take a look at the different ways of raising money.

Drupal Associations has a four-part funding strategy 1.) Memberships of the DA -- where you get two benefits: get listed on the website, and you're able to show a logo 2.) Started advertising on Drupal.org/hosting, and they're on Drupal.org/paidservices & Drupal.org/books 3.) The Ad on the features list of Wordpress.com rakes in about $1 million dollars at $7 a pop. Drupal should be able to make $2-3 million if we put a box ad there. For example, Drupal.org made more in 14 days with a box ad on hosting page than in 3 months of Google Ads. 4.) Drupal hosting -- According to drupal.org, there are 11 companies in the world that do Drupal hosting. But we have a list of 70+ companies and about 25,000 hosting companies in the world. Most of the hosting services in the world don't have a presence on Drupal.org. Matt Mullenweg has said that a real benefit with the ad relationship with hosting companies is that anytime there is a dispute with the company, then it gives them leverage to get it fixed.

Looking at Consulting companies on Drupal.org -- 30 are listed on Drupal.org, and the list is about 10x bigger, but it's weird and hard to get listed on Drupal.org. Have to put some time and money into making these lists more comprehensive.

He's with Acquia, and was able to sell that Drupal was an amazing community to investors. So there is a lot of talent on the DA who are able to have skills in fundraising. Michael Meyers & NowPublic.com went out and raised $13 million for their website. Laura Scott, CEO of PingVision just did Popular Science, and they know how to scope a large site. With a little bit of effort and focus, they can raise $1-2 million for the Drupal.org redesign, and the potential is for there. May not need all of that money, especially if Angie can get an army of High School students to come and help out.

So you'll be able to put it on your Resume that you were involved in the Drupal.org redesign, and I'll give you recommendations on LinkedIn. It'll be an amazing year for the Drupal.org redesign.

Some people try to evaluate how much the code of Drupal.org is worth -- and arrive at an approximate value of $27 million, and look at the community and size of the community and value the community at $10 million or so. So maybe we need to also open up our data.

The more you give, the more you get.

Focus on being able to give out the data, and see what people can do in terms of mashing it up, visualizing it and making it more accessible for everyone.

QUESTION: Four different personas, four different designs and cross-linking? Joomla.org has their own forum site and download sites that have different looks -- The different sub-domains should look as similar as possible, and just be different underneath the hood.

QUESTION: What SPAM controls in forum postings in order to game the pivot module. Not sure what to do about that yet. As an aside, there are two modules, the PHP implementation and it doesn't scale very well. They also have another implementation in Java, which is a little bit faster and thinking about building in a Lucene engine.

QUESTION: Is the pivots module searching lists.Drupal.org? No, they're not searching that data

QUESTION: The listserves are drying energy from the forums, and discovered lots of energy in the listserves. Shouldn't you tear down the listserves entirely? I mean, wouldn't it set a high goal for the redesign of the website that we'd want to be eating our own dogfood? But it's tearing down our community to have a separation between our forums and support listservs.

There is an OG2List module, it's a great idea and it makes sense, but it's a lot of work. We do need to get the listserv discussions available via search. But it's a fairly large red herring to go down that path. It's on the nice-to-have list, but it's complicated. We spent $10,000 doing scalability and testing so that the OG2list so that it could handle 10,000 people, so would need to raise a lot of money to implement something larger than 10,000. And why do that when mailman works just fine?

QUSTION: Working on the newsletter, had a large discussion about creating a standalone news site on Drupal. Set up a working group and proof of concept -- http://Drupal-newsletter.org -- where you can sign up to become an admin. He didn't previously know about the redesign efforts, and just wanted to alert people to the newsletter efforts the possibility that it could be included within the redesign efforts -- possibly news.Drupal.org -- Personal vision is to create model of other news organizations.

The larger question here is -- I have an idea of foo.Drupal.org -- there a zillion of them, who are willing to do them. Things like security.Drupal.org, planet.Drupal.org, design.Drupal.org -- people have the skills and ability, but then there's a quality problem. For example, there's the bluebeach theme maintainer of Steven Wittens who got upset of the lack of quality on the front page posts and in the project in general. We have to have the quality and the resources. Keep forging ahead.

QUESTION: Any work on the Drupal slogan and other marketing and tackle other aspects?

Yeah, should we redesign the logo? And it was cute two years ago. And do we need to revive the brand and slogan? There are 300 people with very strong opinions about that.

QUESTION: What's unclear is how are we going to go about making decisions from a distributed group of people. There will be fragmented efforts on little sub-parts, and so how do you formalize decision-making in a distributed environment?

Formalizing the chaos is a bad idea. DA is the only formalized structure, but most of the work should continue to happen in the organic way in a way that it always has. The proposal of the Drupal Association has restricted the money spending to the IA and design, and the technical implementation will be left up to the community.

The branding and marketing, and overall navigation and structure of Drupal.org needs clear-decision making from the DA. So what's probably going to happen is that we'll get a team of about 10 people to get together an RFP for the IA/design firm. And then they'll be a decision-making process for which firm to choose, but not get into the details of choosing design colors. And we don't want to formalize the distributed implementation of custom coding tasks, and just stick with the normal scratch-your-own-itch way that this type of stuff has always been done.

So you should split off a working group onto groups.drupal.org where you can discuss the higher-level aspects and come to a final decision, and then come back to the community with some actionable issues that reference all of the previous g.d.o discussions.

QUESTION: What about listing current status of maintainership of the module?

Derek is in favor of announcing status of the project with an additional taxonomy like "abandoned," "maintained," "paid maintainership," etc. Derek is against adding an easy way for pinging the module maintainer directly, because he's already overwhelmed with people who haven't even read all of the release notes, documentation, etc.

QUESTION: What about aggregating information, and duplication or information across sites?

There's also a proposed my.drupal.org would be your profile page that aggregates all of your relevant information.

There is also the idea of having an id.drupal.org that has an OpenID server that will allow profile details to be shared across the various different sub-domain sites.

QUESTION: Like to see the forums disappear into different groups and to centralize it in the support forums. Especially since there are already a lot of legit bug reports in the forums

Need to make it easier for people to get help for the module. The pivot module does only search the forums, and not the issue queues.

QUESTION: Aggregating content from other sites as well?

This can't happen this unless with have single sign on, and we also need to have search work across multiple sites. Sony or Fast Company could ban together in order to get these higher-level things happen.

We could update Drupal.org to 6 if we didn't have all of these other things on the site, and we've got to solve these other issues first. And there may already be solutions to these problems. This will come from the RFP process where we get feedback from the IA people that this makes sense.

QUESTION: Could we ever get api.drupal.org except for contributed modules?


Here's how you can help -- Go sign up on the http://groups.Drupal.org/Drupal-org-redesign-analysis

Get in touch with us

Tell us about your project or drop us a line. We'd love to hear from you!