What Structured Content Can Do For You: Article Model


excited you’re joining us today. We have seen
three panelists who can talk about it. But before we get started I wanted to sit back
and acknowledge our previous webinar we did about this topic way back in November. The
National Cancer Institute came and spoke to us about open content and structured data
and how National Cancer Institute was starting to incorporate this model into their agency
system. What Leslie also mentioned was open content structured data and open content model
between working group and the Federal Government that she had been facilitating for a few months.
One of the gold of that working group was to release two content models that can be
used for government agencies to start to share basic information across federal websites.
That’s happening some. But how do we start to really start to share common models? So
what we are doing today is we’re talking about the first of the models that we alluded to
last week the article model. And I have three folks from various agencies that are talking
about how they are attacking this in their agency but also to talk about the article
model. The three folks I have with me today are Holly Irving, who is a taxonomist for
the Office of Communications and government relations at the NIH, National Institute of
Allergy and Infectious Diseases. She has been there for seven years. And she focuses on
the institute’s internet site and special projects for the public site. So her background
is [indiscernible] Sciences and she worked for 16 years in the — at the national agricultural
library and has educational background in biological sciences with a specialty in human
nutrition. I also have William Powell from the US census — Logan Powell from the US
census bureau for media and promotion working on content modeling there for the past year
in preparation for migration into Web CMS. His role is strategy develop — [indiscernible]
And social media. He has an NBA various degrees. Last but not least we have Russell O’Neill
from usa.gov. He is the IT manager and working with content model since the first CMS back
in 2004. So we have a great group today participated in our working group meetings and Holly is
now the facilitator of the working group. She will be maybe talking about that but what
we want the panelists to do today is talk about how the content model can work for other
agencies and how [indiscernible] It can work in your agency. I’ll pass it over to Holly.
I’d like to thank kirsten who is recapping this webinar for us. We’ll be providing slides
with the recap and sending links during the webinar. [ Poor audio ] Don’t worry about
catching the links. We’ll provide all that information to you directly after the webinar.
So with that in mind, thanks so much Holly. Take it away.
>>Thank you, Jacob and thank you all for joining us this afternoon. I hope we can give
you some ideas about how structuring content can be important to your agencies and how
content models can fit into that process.>>So I’ll describe or attempt to describe
what structured content is. I’ll talk about some of the uses and benefits using my own
organization as an example and provide a very brief overview of the article content model,
again using my office as an example.>>So this definition is taken from the digital
government strategy objectives. The four principles. First of all, structured content is information-centric.
So it requires a shift away from thinking about presentation and thinking about the
content as modular discreet pieces of either data or Web content. Now, these pieces are
also referred to as chunks. You may see them referred to as elements or metadata. and it
can be tagged up, so that the chunks are recognized across the Web, mashed up and shared in multiple
ways both within and outside government, and presented in a way that’s most useful to the
consumer. Related to that structured content puts the customer first and allows them to
shape, share and consume information whenever and however they want. It allows us to innovate
with less through collaboration. It allows to us streamline development and reduce costs,
and be consistent in our standards and delivery. And last it, allows us to be transparent but
also accountable and provide digital services in a safe and secure manner. And incidentally,
I think the digital government is a really good read and I would encourage you to take
a look at it if you haven’t already. I work in the communications office of NIAID as Jacob
mentioned and the missing. Our public website is to make all our audiences aware of the
research going on in our institute and for us, as I’m sure there are for all of you,
be there many demands on our content. Most frequently our Web content is generated first
with the intention of publishing as a public — Web page on our public site. But we also
have this good information that we can offer to other sites not through links but through
syndication, which pulls our content into their own user interface. And if nobody can
find our content on the Web, we basically wasted our time. So we need to do to increase
what we can through findability through search engine optimization.
>>We also want to be able to leverage our own content to our advantage internally whether
it’s auditing our assets or creating dynamically generated Web pages, for example. In order
to reach many of our customers, we must — must push our messages out on social media which
is where they are getting a lot of their information. And last but not at least is the requirement
that our information be usable on mobile devices which is how many of our customers get their
information. In fact, it’s a responsive design project that’s really fueling the movement
NIAID’s structured content. So writing and formatting the content is efficient or effective.
We need to create one and publish it everywhere and in order to do that NIAID needs structured
content that’s open and usable pretty much as described by the Digital Government strategy.
Content model defines a chunk or pieces of metadata that you will need to meet the requirement
and goal of your Web content. So as an example, let’s look at the apartment Hunting. By the
way, this example for no particular reason is for a studio apartment in in rockville
Town Center in Maryland. So what information is key in making a decision to look at an
apartment? If I were a building department owner and I know the needs of my clients,
I might say that these ten basic pieces of information are important for online apartment
shoppers. So logically, this is the information that I want to make available on the Web.
I have defined my chunks of information and now I have a content model. So this is the
way the studio apartment fits into the apartment content model. The building name and the location,
square footage, is there a floor plan available, yes, there is on the site. It has one bathroom
and no bedrooms because it is a studio apartment, has one closet, price in the 1300s, the contact
information is available on the site and the availability is provided, as well. In addition,
there could be related information about the builder’s reputation and background or other
studio models that are available in the building, and optional information you could give a
title to the model, a longer description, building amenities, other visuals, videos
are very, very popular for these kinds of situations, whether or not it’s pet-friendly
and deals are offered.>>Here’s an example of how that availability
chunk looks like on the Web. It’s part of the navigation for the mobile site along with
some of the other elements that we looked at, both key and related are optional and
on the full website having this metadata about the apartment allows customers to self-search.
They can select on some of the key as well as some of the optional information, and then
check for the availability of apartments that fit their criteria.
>>So in the same way that the apartment content model works, the article content model defines
the chunks that are common to most government websites. And it’s not just for journal articles
or news articles. It can be used for straight Web information like program or product descriptions,
et cetera.>>The working group identifies five fundamental
elements that are designated as requirements. These are core to the content model. There
are 12 optional fields, some of which can occur once like a title and there are eight
elements like topics or references that can be in multiples. And some of the elements
have subfields to provide more specificity. This model is not exhaustive. It’s a framework
for extension and modification as needed to suit your content. And this at this point
is just a very mile high quick overview. In a couple of minutes, Logan is going to speak
in more detail about the model and Russell is going to talk where the model is housed.
>>So the article content model and the required fields. As I mentioned earlier, NIAID’s project
has generated a need for us to work much more carefully at structuring our content. And
we are actively moving in that direction. So here’s how this NIAID research feature
one of our Web pages fits into the five required elements of the content model. The title description,
source organization, topics, and the article body which in this case could be broken even
further down into these five parts.>>As you can see, we don’t use topics at
this time. But the assignment is subject tag is a real priority for us.
>>And here are some of the optional fields for the article content model that are represented
in the research feature. The references we have a link to the “science” article, the
article in “Science” magazine where the research was originally published. And for related
URLs, we have a link on our own site to the lab information about — information about
the lab that is responsible for this research and to our malaria portal page.
>>So to summarize, structured content, discreet pieces that can — of contents that can be
manipulated, can be distributed where and house it’s most useful to our audiences. Because
we have multiple demands, we can let one Web product effectively and efficiently serve
many purposes using structured content. And content models define how you chunk it up.
— how you chunk up your Web content and the article content model provides a foundation
for agencies to adopt and adapt as needed. And now you’ll hear more about the content
model from Logan Powell.>>Thank you, Holly. That was an incredible
and wonderful overview and summary of content models and, unfortunately, you know, how they’re
composed if — and, uhm, you know, how they are composed between building and developing
and design them. I’m going to go through the process which we went through — through which
we went when designing this particular content model or the required identifying — identifying
the required and recommended fields. This is the screen from schema.org referring to
the article content model that’s displayed there. Schema.org is a relatively well established
— I guess you could say a repo of content models of number — of a variety of sorts.
You can see that article here is a child of creative work which is a child of sing and
that’s just one of a number of different types of content that are described in schema.org.
So one of the reasons we chose to use schema.org was because of this right here is a list of
the users of schema.org. To give you a sense of the — the importance of this list, one
of — as Holly had mentioned, one of the important aspects of creating a content model is making
your content interoperable with established platforms so for example, if you want your
content to show up on Google and you want that SEO to be in there, you’re going to want
to use a content model that Google can understand and actually Google uses schema.org. Facebook
uses it, eBay, a number of other high ranking — Expedia — are using schema.org to look
for content in a specific metadata standard. The way that we narrowed down this incredibly
long list of metadata that’s available in the content article on schema.org, this is
much longer than content model that we’ve created, we use the m1313 schema that is outlined
in the open data project or the open data project is sponsored by the GSA and it’s something
that’s basically a sort of best practice kind of core identifiers for the content across
agencies and what we did is essentially take this “Common Core” metadata and overlapped
or mapped it to the schema.org metadata for an article to identify the most salient metadata.
>>Now, the way that this can be applied in your government context, in your agency context,
is going do vary between agencies. Some agencies are going do be more general sort of transactional
agencies, for example, I don’t know the IRS might be a good one where we’re going on to
the site to execute something, to go there, to transact to take a certain action and once
you’re done with that, you’re out. You’re done. You’re happy. Some like our agency the
census bureau, we’re much more interactive in terms of sort of a query response kind
of thing where people are coming to our agency’s website looking for data. So the way that
we are thinking about content models at our agency is every — all bits of content are
about data so at the core of our metadata is a blend of not only schema.org schema but
as well information referring to the nature of the data that the content is talking about
or the narrative or visualization, whatever is describing that content both in terms of
how it is connected to the larger ecosystem of search engines and other ways of sorting
or formatting the content with — including with that the, uhm, the data, the internal
metadata that — that our agency, uhm, is disseminating, the — the nature of the dissemination
work that we do.>>Uhm, that’s really I guess on that note,
I wanted to mention just a couple of things tying this into the larger — larger picture,
kind of as just recapping on what Holly had mentioned. The Digital Government strategy
information centric and customer-centric, mandates as well as Holly mentioned the cope
strategy which was published everywhere. The thing that those two things have in common,
this information-centric approach, to enable inter operability. So we want two types of
interoperability. The interoperability of your metadata within your organization as
well as the inter operation ability of the organizational content within the larger ecosystem
of the Web. So the schema.org metadata essentially allows us to plug into the larger ecosystem.
It allows us to plug into the Google, the Facebook, the twitter and so on and enabling
that content to be understood by a larger ecosystem and then this more custom tailored
data is the smaller internal interoperability the metadata that establishes a standard so
their internal organization, your internal systems, can kind of speak to each other and
understand, format, consume and exchange that information internally in an intelligent way.
>>So that’s kind of a sip none us is of how we came up with the kind of science behind
the process of developing a content model. I wanted to toss it over to Russell, who is
going to talk a little bit about if you guys have — if you guys have questions or comments,
suggestions, feature requests, how you might do that and how the get hub account for our
content model and our community is structured. [ inaudible ]
>>Hopefully thanks to that great introduction by Logan and Holly, everybody right now is
super, super excited about the article content model and what the structured content working
group is working and wants to help us improve, make suggestions and get involved. So that’s
our goal. So my goal in the next 10, 15 minutes is to show you what we have developed to facilitate
that collaboration and communication so you can get in there and help us make it better.
>>to started I wanted to take five minutes to show you the website we have put together
that kind of will house all of the work we have been talking about. And it’s a — I’m
not going to say the url, it’s long. But it will be sent out at the end after the event.
So basically this is the homepage. Not going to talk too much about it. There it is. I
did want to take a few minutes to show you a few things about the website so you can
look around and see for yourself when you have time explore it and get involved. The
content model, this is where the two content models that we talked about. Today is focused
on the article content model, all the content models are presented in a similar way. They
start off with a brief description and then they show the structure of the model. So Holly
went through it at a broad level. This is the specific. So this is all the fields that
Holly was mentioning all the subfields and so the indentation is your — is a subelement.
So for example, a title is going to have potentially a short title and definitely a full title.
And it’s only going to have one of those or both but it will be only one short title and
one full title or one full title only but they will all go under the title field. Likewise,
the article body consists of multiple article sections and then each article section has
a title and a body. So that’s how you can kind of read the summary where it kind of
just tries to outline all the field and shows you how the sub fields relate. Okay? But you
know, maybe the name isn’t that descriptive. You want a little bit more. If you scroll
down this will have a definition and describe in more detail each of the fields you see
above so for example, it talks about formats where it’s appropriate. So for example, the
language it talks about what language you might want to specify for language like [indiscernible]
Or things like that. So this will give you more information on each of the fields. Additionally
in cases where there are subfields, there are links, so if you click on down, it will
just take you right to the definition for that link. So the article body consists of
just one or more sections. And then the section with the title in the body. So that’s kind
of how you view. The event model has the same setup. Take a look on your own and provide
feedback. That’s how you read the content model and it’s set up with where any models
we add will be the same way. The key thing to get out of this is how you can contribute
to this conversation and so we have a page that describes it so you can very easily find
it but I’m going to talk to it right now. Like Logan mentioned everything you see here
on this site is open source and on get hub. Get hub is basically a sort of repository
for open source projects. You can do what you want from GitHub and bring it back to
the community. We are using it for the websites and models themselves. We are trying to make
it simple on GitHub or click on the links on this page you will be taken into the GitHub
site. When you first go to the page you will see something that looks like this. It will
resem this be basically but to actually contribute to the conversation you do need to have a
free GitHub account. I’m sure most people know how to create an account on a website.
GitHub is no different than any other website. Just create a free account to get your log-in
but you need that to the conversation. Everything else is hopefully really easy as you’ll see.
>>When you’re here, you know, just to go through a few basics at a high level what
you see here are the actual source code for the website itself. I’m not going to talk
too much about that. I’ll show you a couple of things but you can pretty much ignore it
if you just join the basics of the conversation. The real thing to join the conversation and
give ideas, contribute feedback, is through issues. If I just Clish on issues, you will
see I get into essentially a forum and it really is just like a message forum like you
see on any other website. Just call it issues on GitHub because of its source code nature
so you can see we already have a conversation going on schema.org hub relationship to the
content models after this went up on DigitalGov.gov. You can see there’s a conversation going.
We hope that everybody on the webinar and everyone else that’s interested joins this
conversation through this model at least. And you can see it’s just a conversation like
a comment post on any normal website out there. So you can see Logan started with the conversation
and other people responded and it’s been a good conversation. You see that to contribute
just at the bottom, just like anything else, write your comment. You can just like in Twitter
you can add people. Logan was responding specifically to lucify and also to everyone. You can contribute
very easily but just typing what you want and hit comment. You can create a new issue.
We would love your feedback. If you are using this article content model we would love you
to create a new issue and say I’m using it for this and this is how it’s working out.
We want to hear how you’re using this stuff so please contribute. It’s easy and we would
love your feedback and ideas.>>So that’s issues. But let’s say you want
to do something a little more than just contribute to the conversation. You can also go back
and contribute to the code itself if you want. I’m going to talk very briefly about this.
I’m not going to go through the whole process but it’s called a pull request. In your GitHub
account you create your own copy of everything that you see here of all this source code
and that’sforking the project. You fork the project, you make changes, suggestions, hey,
I’m thinking you should add this field here, oh, you need to change this language, whatever
you want to do and then when you’re ready and you think okay this is great I should
send it back so that we can potentially add it to the master site, you do what’s called
a pull request which will basically say here’s my suggested changes. You can then now accept
them and they will automatically get merged in so they show up. Don’t worry too much about
that. I just wanted to go through the process at a super high level so you can find out
more if you’re interested in doing that. Anyone cab contribute. But I did want to go through
if you are interested pull request, just a couple of files you should know. So everything
you see on the website is stored in simple text files. We wanted to make it as easy as
possible to get feedback no matter how you wanted to do it. If I look at the FAQ page,
this file is basically what generates this page here. And if I look at the actual source
code, that you see it’s very simple text. It’s not complicated. It’s very much — it’s
a format called markdown which is a format that GitHub uses as well as other places.
If you never contributed to wiki, they often have a special format to make things easy,
the heading, bold, underlined, markdown is just like that. You can see the structure
but it’s not that complicated. There’s no funky HTML. It’s pretty straightforward. So
everything you see here is what gets turned into this HTML automatically for you based
on the website so it’s very easy just this text file that you have to deal with. You
don’t have to deal with any HTML colors or anything like that. That’s for the content
pages. For the actual models themselves it’s similar. If I go into the underscore data
file, you will see that every content model has a YML file. It’s just another text format
that’s very popular but if I click on it sticking with the article content we’re looking at
you will see it’s basically just — it is a text file that describes everything in the
article content type. So you can see here’s where it says the title field, it describes
the title field, whether or not it’s required, whether or not multiple are allowed, sub elements
of the title, does it have a short title and full title and likewise a description required,
multiple allowed, as well as the parent element. It’s easy. It’s just text. And this text file
is what is used to generate — back — this entire page. So again, the goal is to try
to make this as simple as possible for people who want to contribute in all sorts of different
ways because we want the community, we want you and everyone else who wants to get involved
to get involved. We really want your feedback and ideas for improvement. So hopefully this
gives you a broad outline of how you’ll do it. So I hope you’ll get on GitHub and join
the conversation and tell us what to do to make it better and also to have ideas for
the future like the content model ideas that you think would be good across the government
standardized, let us know. So with that I’m going to turn it back to Jacob for some Q
& A.>>Thank you, Russell, Holly and Logan. Very
well done. I have actually done a number of these DigitalGov University panels and I have
actually never had three panelists be so concise and succinct in their facts. Generally it’s
like five minutes left and we’re doing Q & A. So we have plenty of time for Q & A today.
We have a chat box if folks want to ask questions. Please submit your questions. I know this
is a new concept for a lot of people. One thing I wanted to touch on we had a number
of agencies help us find these basic elements of these models. And so you know, it’s a finite
group of people and a group of agencies that all agreed to these elements. So we really
would like to hear what other elements you would like to. And then — and we’ll talk
about this in a future webinar but we’ll be asking you in the post webinar survey what
should the next governmentwide content model be? We’ll talk about that later. But I already
have one question for the panel. And I think that, you know, we talk a lot about the benefit
of structured content model but someone has asked me, you know, what happens if you don’t
follow a content model? What happens? So they are having difficulty imagining an effective
example. So kind of give some feedback of what might happen if you weren’t using a content
model. I have some ideas but I’d like to hear from the group on what they think of this
concept.>>[ Pause ]
>>I –>>[Inaudible — Multiple Speakers]
>>>>This is Logan, I can chime in.
>>Sorry, go ahead, Holly.>>I think, uhm, — I know that, uhm, by not
using a content, uhm, by not using a content model, you may be missing out on some opportunities
for your content. As Logan mentioned, we base ours on largely on schema.org and this is
a project that is — this is a model or through schema.org there are a lot of people who are
participating in using these elements and improving search engine optimization by adopting
this content model. I think you might be possibly losing out on some opportunities. The other
that I can think of is utilization of resources. As I mentioned, reworking content for a variety
of vehicles and platforms may not be the best way to use your row sources. So if you can
use that model get your writers and your content creators into that content model framework
and you can take advantage of that product and leverage it a lot of ways.
>>So yeah, I just want to he had phi that a little bit. I would say if you don’t create
a content — it depends how sophisticated your information is in general. If we’re talking
about a very small amount of information and you’re using like a word press content management
system or if you are using a small sort — if your site has a very small footprint, then
it may not be necessary just because people could find their way through all of your content
without having to search it, sort it or anything like that. But as soon as you have a significant
amount of content information or data, as soon as you have an amount that’s just too
much to chew on, too, big to — to bite, uhm, all at once, you are going to have to start
thinking about the structure of that information. The impetus behind our foray into content
modeling and one of the original kind of conversation starters for this group was, you know, okay
well you’re starting to — you’re going to build a CMS, you’re going to migrate your
content into a content management system. One of the first things that you’ll be doing
when you’re going through that process is figuring out, how do I structure this content
so that I can put it into this content management system? That’s a prerequisite to figure out
— what are the pieces of this type of content that I stick into the content management system
to show up in the format that we have our HTML in over whatever. So right off the bat
if you want to manage your content in a CMS you need to talk about content modeling.
>>You can — there’s two ways to approach t start with the design and figure out you
want it to look and feel how it designs first sort of approach and figure out all the different
pieces of the information that’s going to fit within that design, or you can take an
approach that kind of, uhm, alternative opposite end of the speck tum, like what is the nature
of our content and these vier yaws types like if we publications or we have tables, we have
data visualizations, this is just we have all these different types of content, how
do we structure them? It serves as the backbone between the content so they are related among
the dimensions of the models. So you have every piece of that content model is basically
serving as a way to connect it to other pieces that have similar meaning. Right? So I have
a topic tag or a topic metadata field and one piece, one type of con dent. Like I have
a data visualization that shows an age-sex pyramid for the last 50 years so I could have
a date range in there, a topic which is like sex and age or something like that, population,
uhm, and then I’ll have that thereby attached to a report that could also be with age-sex
change — not sex change, but age and population pyramid changes and evolution over time like
a report. Those two things would thereby be connected through the metadata. So the metadata
is really kind of just abstracting out the meaning of your content and without really
going through this process of understanding or creating the content model — I should
say that that requires that process in order for you to organize it so that it can be not
only used by a content management system internally, it’s also a best practice for leveraging it
externally so search engines can find it so that you can if you do have lots and lots
and lots of content, the search space can be reduced by the dimensions on that model.
So if we have 17 topics, I’m just interested in one, I can type that into a search bar
and it’s going to eliminate 16 out of 17 topics from my search space. So really the content
model helps people get down to the specifics of what they’re interested in, in that sort
of landscape of your content.>>One other thing. So also, the way to think
about the content model is also how it can be used to go back with the way to pull things
together. Like a basic example of some of the advantages of having common content models
is if you ever use Google and searching for an iPad or some product, now how sometimes
it pulls up like the ratings from review sites and the stars are right there? It’s able to
do that because of common metadata, a common content model that Google says, hey, if you
publish this particular content in this particular content model as part of your Web page, we’re
able to take that and understand it and build out a special visual presentation because
you did that. But if it wasn’t for that content model, it would be much harder for Google
to be able to try to do that so that’s another big advantage is being able to having common
content models being able to really aggregate and display things in very useful ways.
>>Thank you for answering the question. We have a lot of good questions, please send
them in. So what I’m hearing from you guys in the answer to that is that content models
are almost before the CMS piece before you start to put it into a file form to present
it. Is that a fair assumption?>>Yeah. I would say so. You don’t want to
go into a CMS, you don’t want to start migrating your content into a content management system
and have to change the metadata later because all that migration work that you put into
taking your stuff, sticking into that CMS, if it’s not intelligently labeled with this
metadata, this content model, temperature it’s not labeled all that work that goes in
there is not able to be related to other work. It’s not able to be sorted so well. Now, there
are things like Apache solar that might do a text-based faceted search. That helps people
who don’t have structure to their content but that’s an ad hoc solution, not the cleanest
and definitely not the way to present it. People are forced to go to your website to
get that content. If you ever want to avail your content to people who are going to aggregate
it into, I don’t know, like newsfeeds or their own content like if they want to create like,
uhm, a flipboard or something of your content, if them to leverage your content externally,
if you want to enable that to happen or even if you want internally to be able to in a
very controlled way link — relate content together, you’re going to need a model. And
I would suggest doing that work way up front. Spend the time on the content modeling as
much as you can before you start migrating your stuff into — because putting it in before
is a whole lot easier than changing it ex-post.>>I would agree with that. I think it’s a
good exercise actually to go and look at what you have as far as content tow kind of take
stock of what you have and what your goals are and what your needs are as far as utilizing
that content to its best effectiveness or best use. So before you even start with the
content model I would take a look at your content and figure out what you want to do
with it and where you want it to be. And then start looking at how you can model it so that
you can take advantage of that so you fulfill that mission.
>>Great response. Thank you, guys, for pointing that out. One question that’s come up in a
couple of areas, so I’ll combine these questions a couple of users different users have asked
them. Thank you for — keep the questions coming in. We’ll answer questions until 2:00.
You know, how a content model will work for mobile? You have talked about designers, Holly,
[ inaudible ] can you talk about how responsive designer mobile might advise you to use certain
elements of a content model?>>So this is Logan again. I know I’m sharing
my screen so I think I can use this as a segue into what I’m showing here. I’m not always
the best communicator and I’ll be the first to admit that but one person who is an amazing
communicator for a content modeling in general is this woman named Karen McGrane. I saw her
speak at the how conference last year. She is an amazing public speaker and has a really
good slide deck. You can actually view it here if you just look up adaptive content
Karen McGrane on Google and her name. You will find this. This is a slide I jumped to
quick and it explains that. NPR is a poster child for content modeling. They are incredibly
well versed with it and they are really pioneers in terms of structuring their content so it
can be leveraged across different platforms especially mobile so what they have basically
done is they have structured the content through content modeling and metadata and availed
that content through an API. Much of our understanding in government of application programming interfaces
or APIs, has been around data. How do we make our government data open and how do we allow
other people to leverage that data in their applications? It can also apply to content.
And it will in the future more and more often apply to content. Facebook’s content is available
through an API, Twitter’s content is available through API. So the social content is attached
with medical data and then you can pull that stuff out of Facebook and present it in different
ways through an API. But what NPR has done is that API not only enables them to present
their content on the website or on the mobile device or on your mobile device, your mobile
app, it also allows them to present it on radios like their little descriptions show
up on the radio or on the Sirius or on iPad apps, so they can create front ends to that
same content, they create new front ends, they told us in person, I met these guys it
took them two weeks to create an iPad app when it came out on their existing API because
that content was modeled through metadata and this content model. It was available to
be — you just create a new JavaScript in HTML and serve it up. You no longer have to
think about the future of the content. You’re no longer thinking about the presentation
as much. You’re thinking about, what’s the meaning of this content? How do we organize
it? How do we structure it so in the future we can present it any way you want. I highly
recommend you watch had your presentation and it’s also entertaining and informative.
That’s the external example. But also internally they are using these APIs themselves eating
their own dog food so to speak where it helps them build applications on top of their own
databases their own stuff much more quickly so they are just mashing their own content
up together and presenting it in different ways. So they don’t have to kind of go back
in figure out how to reorganize this content into new products. They just mash up their
existing products together and create new products with that because the structure is
already built in. This is an example of how NPR’s content management system presents them
with the options for that content model is he they add the date, the type, comments,
the title and teaser and so on. This is typical front end to a content management system that
puts it into this machine readable format so I would highly recommend you read this
article, watch this video. It’s a great primer to content modeling if we haven’t covered
the bases, this should probably drive it home.>>Russell, so we have kind of talked a little
bit about content [indiscernible] Why you should use them. The next question is, so
if my agency wants to adopt this model we can get the YML file [indiscernible] What’s
the next step for getting [indiscernible] Into or some other CMS? Russell, do you want
to take a hack at that?>>Yes, great question. So the YML files is
really just for the visual presentation on this website? So you don’t need those? If
you decided you want to take advantage of this model and I hope you do, all you really
need is this description — is the description that you can see in the HTML version and when
you’re building into the CMS most CMS have a fairly straightforward way to build your
content models? So you can say I’m going to use a simpler example than the article but
I’m creating a new widget or whatever and then it can say within a widget it has a field
for a title, a description, all this and maybe it has a date or maybe it has — most CMS
products have a pretty straightforward way of creating the model that you see here or
whatever model you need? Which will generate the type of field which Logan showed like
from — uhm, npr uses based on that definition. You don’t need the YML to build out the content
model in your own systems? That’s just more if you want to contribute directly back to
the working group and directly back to the website? But it does provide the descriptions
you can give your developers and say this is what we want to implement, this is the
baseline let’s talk about this so they can then take it and build into whatever CMS product
you’re using. .>>Thanks, Russell. Uhm, so lots of good questions,
uhm, here. So — so what — this might be something that’s a good takeaway for the working
group. What techniques have you used in order to deal with the issue of description or body
in the CMS being a large block [indiscernible] Markup and [indiscernible] Script [indiscernible]
Et cetera, the people are used to? Have any of the folks on the panel had experience with
that?>>That’s a great question. I can talk about
from our experiences on usa.gov at least.>>All right.
>>On usa.gov, we actually right now are using a content model that is very, very specific
to a link-based portal involving links and URLs. It was a large transition when we switched
to this model some years ago and it’s been quite a few years now where they, uhm, to
migrate the mindset of okay I’m writing a document in Word at the time before we implemented
the first CMS or even as a block of HTML, to — I’m filling out metadata fields in a
CMS that will then generate the page. It is a transition. I think one of the beauties
of the elegance of like the article con content model is the article content model is really
adding more descriptive but the fundamental is still that block of HTML but it has a little
more structure to it with sub sections. That’s one of the nice things about the article content
model when you get down to it outside of the metadata about the article, the title, date
and stuff like that, the content is still just a bunch of subsections, right, in that
element. So that’s still just HTML is the idea. But for more complicated content models
where you are giving it up it is a bit of a transition and, you know, and I think the
easiest to say is with usa.gov it just took time as content managers switched to that
mindset of filling out those metadata fields to filling out that more descriptive information
as opposed to just a block of content but they also saw the benefit once we went live
again in the case of a link-based pour at that time by splitting up the content model
gave us the ability to do if we’re using a URL on 15 pages, if it was a big block of
HTML we would have gone to go to 15 different pages to change the content potentially. We
only had to change it once and it was automatically fixed on 15 pages. It’s a training issue and
a different mindset depending on how structured you’re going? But I think once you can the
benefit of it, I think most people will see that the benefits do outweigh the little bit
of mental shift if you will that has to happen.>>I can add to that by saying, NIAID is in
the middle of figuring things out, how detailed we want to be, how we want to chunk up that
information logically and usefully for us and so it’s very much a shift as Russell was
talking about if you are — those issues and build that into a CMS. I think again with
the article model, we have the foundation there, but you can add the complexity and
the specificity that you need. It comes down to looking at the content itself, figuring
out what you want to do and at that point, you can start thinking about what chunks,
what elements are going to be meaningful to you and at what level of specificity.
>>We have about five minutes left. So last call for questions. Are there — there are
a couple of questions around content modeling — there is an interesting question, what
sort of research contributes to developing content model? We talked about how we base
[indiscernible] Off schema.org and a couple of other plus this — a couple of other influences.
What do you think of that, a mental model?>>What was the end of the question.
>>Could a content model be formed by a mental model?
>>I would say definitely. That’s what a content model is a mental model. It’s a — it’s understanding
the meaning of — you have to get a really firm grasp what the meaning of that content
is and/or data is in order to structure it out. Because you’re basically abstracting
the meaning out of the content reducing it down to its sort of most essential meaning
by modeling it. The process that we have gone through internally is actually arduous because
we have so much data at the census bureau. We have just vast, uhm, amounts of data and
so many different directorates and so many different divisions within their directorates
that are producing data and that they are — they are producing similar types of data
but not calling it the same thing. So that deciphering process, that translation process,
and that mapping process is a lot of it for us internally when creating — we created
our met — when we created our metadata model so depending on the vastness of that landscape
within your agency it could be a much more time-consuming process. But it is just a matter
of talking to people who understand the content the best, understand the content themselves
are producing it themselves, know it by heart and could explain to you kind of in its most
essential parts — the way that I actually started and the way that I — a good way to
conceptualizing if you are going for a specific meaning, the mental model that you have or
that anyone has of the agency’s content of your agency’s content is start with the interrogatives.
The time, the audience, the user, the topic, the — so starting with those interrecognizetives
basically trying to narrow it down to what answer — what questions are this content
answering? That’s a good way to start the mental modeling process.
>>I would expand on what Logan said and pay a lot of attention to your users, to your
customers. If you have any kind of usability, any kind of analytics, find out how they think
about this information. And again, what them to get from it. You can even think about it
as top cops. What are people trying to accomplish when they come to the site? And then there’s
those interconnections, you know, when somebody goes to a page, what are the other things
that they want to see, what kinds of information can you bring to that page that is either
related or relative to that top inning so the mental model is a really good way to start.
>>We have three minutes left. We’ll be sending out an email after with a survey to talk to
us about what you thought of this webinar. We’ll have questions about the next content
model and if you are interesting know joining the content model working group. Link to the
GitHub page. Anyone can go on the page and talk about what we need to do or what you
would ask of these models, pull requests, anything. We will have a follow-up webinar
on the event model but with that being said I want to ask our panelists one more question.
So if there was one thing that you could tell — if there was one reason to do a content
model, and you only have this one shot, give me one shot to give people content model what
would that be? And you guys 10 seconds each. [ Laughter ]
>>I could be flip and just say because I said so. But I think what it does is it liberates
your content. It draws it out of silos and format requirements and all kinds of restrictions
and frees it to be — to move throughout the Web in any way you would like it to.
>>I could also say what Holly said. But, uhm, the, uhm, but really in general, I mean,
I think the thing is also I would add to what Holly said it also gives you a much better
insight into your content. It forces you to think about your content in a really nice
way that really helps you to Holly’s earlier point think about what you’re presenting to
your customer as well how you manage it on the inside of the organization.
>>Logan?>>Yeah. I would.
>>Don’t be shy.>>I’m not shy. [ Laughter ]
>>I would say it makes it more findable. It makes it more flexible, future proof, right?
You know, it makes it more usable because, you know, we, uhm, — I’m going to be humble
here and let’s just say that the government isn’t so great at, uhm, creating beautiful
and hot products. We are incredibly good at securing and, uhm, providing access to our
content and our data. Well, most of the time. And that’s what we can focus on. We can do
what we’re good at doing which is securing and availing that content. And allow people
to present — by creating these content models, we are enabling these third party like Google
and Facebook and so on to leverage that content. But also like in the future when we do get
better at that,it would take the con dent and modernize it. We’re future-proofing the
content by using it this way making it much more usable.
>>We’re out of time. Thanks, everyone, for joining in the call. I’d like to thank my
panelists. You can join us again on June 12 but you can always climb in on GitHub. Have
a great day. Talk to you again soon. Bye.>>[Event Concluded]

Leave a Reply

Your email address will not be published. Required fields are marked *