Good morning everyone.
Look at you, up for that first session.
You've had your first cup of coffee.
Thank you for joining us.
My name is Laura Wood and I'm going to kick off our presentation.
Before I do, a couple of words.
They are recording this session.
The Q and A, however, will not be recorded.
It's just the audio of the first part of the presentation that they're trying to
capture and the slides.
To my left, your right, Jody Combs, who is from the University of Vanderbilt, and also
working with ARL more intensively these days.
And, our invisible co-presenter, Beth Namachchivaya, from the University of Illinois.
Unfortunately, had to send her regrets.
She is ill and was not cleared for travel.
We are quite grateful she is not sharing with us at this time, although we wish she were
here for the presentation.
We do have notes from her so we will try our best to represent the content and the update
on her project.
So, jumping in, the issue of accessibility is not a new one at all.
Most of us are familiar with the Americans with Disabilities Act, ADA, from 1990.
People, I find that people may be less aware of the Rehabilitation Act that President Nixon
signed into law and in Section 504 of that, extended civil rights to people with disabilities
and provided a lot of opportunities in education and anywhere federal funds are involved.
Section 508 required that electronic and information technology developed by the federal government
be accessible to people with disabilities, in 1973.
So, they were fairly prescient, and we're still trying to figure out how to do that.
As scholarly content made its way into digital devices, electronic versions and online, colleges
and universities have been adoption these and finding that accessibility issues are
a major stumbling block, leading to lawsuits and typically settlement agreements.
These agreements are a trove of information and have served as a startling kick in the
pants for some institutions.
About two weeks ago, many of you have already seen, the 2017 NMC Horizon report on libraries
was released.
It helpfully identifies trends for the industry as well as challenges.
Accessibility of library resources and services is called out this year as one of those challenges.
T he good news is that NMC has labeled this challenge as "solvable."
It's a challenge we understand and know how to solve and that's, I think, where
we come in today.
There are many projects.
These will be just three that are trying to work on how we do that solving…how we chip
away at that challenge.
I will discuss accessible instructional materials or AIM.
Jody will talk about video captioning and I'll share a few words on behalf of Beth
regarding Daisy Files and the HATHI Trust digital library.
I do want to spend a few minutes in the beginning talking about the scope and nature of this
issue just in case some of you are also looking for some entry into the issues of accessibility.
I hope you can read this in the back.
According to the US government accountability office, 10.8% of students enrolled in post-secondary
institutions in 2008 had a disability.
While dated, these seem to be holding and are the best comprehensive statistics I could
find that array this way.
This represents more than 2 million students in post-secondary institutions across the
United States.
Students with disabilities share demographic distribution with the general population of
post-secondary students in terms of race, age, and schools attended.
So, this table shows you the categories of disabilities that have been tracked and how
that 10.8% population is distributed across those categories.
It's important to note here that learning disabilities is by far the largest population
group and that's where most of the growth has come from.
Statistics on disabilities in higher education are also based on self-identification by students
and so are widely believed to be underestimates, which is only logical.
We just doing know how dramatically their underestimated.
88% of institutions report enrolled students with disabilities and 99% of public institutions,
and 100% of medium and large institutions report students enrolled.
So, accessibility is an extensive and relatively urgent issue in higher education.
It also helps to talk about what do we mean when we say 'accessible'?
The federal government has used this language in many of the settlement agreements.
The repetition of it over and over again serves as a signal that this is not an evolving concept
but a clear expectation of compliance.
Accessible means a person with a disability is afforded the opportunity to acquire the
same information, engage in the same interactions, and enjoy the same services as a person without
disability in an equally effective and equally integrated manner with substantially equivalent
ease of use.
The person with a disability must be able to obtain the information as fully, equally,
and independently as a person without a disability.
Although this might not result in identical ease of use compared to that of persons without
disabilities, it still much ensure equal opportunity to the educational benefits and opportunities
afforded by the technology and equal treatment in the use of such technology.
Now, a basic element of trying to meet accessibility is the availability of required readings for
a class.
So, I'm going to start to talk about my project update, which deals with that material.
We've been working on an IMLS title grant, which was titled, Repository Services for
Accessible Course Content.
The core members of my steering committee on this project were Jamie Axelrod, Stephen
Downie, Mike Furlough and John Unsworth.
Major help came from David Wedaman, who is also a co-author on the White Paper, and from
Katrina Fenlon, who really did yeoman's work on the research we did, the focus groups we
conducted, and then the data analysis.
She is the first author on the paper that was published.
So, what is accessible instruction materials?
In order to provide accessible learning materials, institutions navigate a variety of sources
to see if digital formats already exist.
They also request electronic copies directly from publishers or through a mediated service.
Some publishers are relatively responsive to requests but others aren't.
Response rates range from one day to two weeks or more and there is no obligation for publishers
to be timely with responses.
As a last resort, a school may scan works from print, some of them do that a lot because
they don't want to wait.
Regardless of the original source, a digital file will then need to undergo significant
reformatting before delivery to the student.
Numerous institutions may be seeking the same text at any given time but they have no mechanism
for sharing.
Here you might think about best-selling textbooks and how often they would be used across the
country and if one of those is made accessible where it might also be needed in an accessible
format.
The work that goes into a single file may represent many hours of labor and the quality
of the results vary depending on the school's resources.
As those electronic files are created and provided to students, the disabilities resources
and services department must wrestle with a significant file management problem.
The files need to be secured for the sake of copyright.
But, most courses are taught multiple times with some re-use of common texts.
The ability to securely store, describe, and reuse these reformatted materials is necessary
on every college campus.
In our planning project, a major component was sitting down with disability services
staff to better understand how they create and manage accessible course content.
We held focus groups at the AHED Conference in 2015.
Ahead is the Association of Higher Education and Disabilities.
The paper that was published through the proceedings of the ASIST annual conference and the White
Paper we very recently released, both detail more of the findings from that research.
It is not the number of students or the volume of requests as much as the nature of the request
that determines the amount of effort and funding required to make materials accessible.
Focus group participants were very clear that there were certain disciplines and content
types that are exceedingly difficult and resource intensive.
STEM disciplines are by far most difficult and highest in demand.
The common use of equations, images, charts, and graphs make accessibility requests very
challenging.
Images need text descriptions far beyond an image caption that might already be provided.
Charts and graphs may need to be depicted tactilely for some students.
Staff may need to solicit help from the teaching faculty or graduate assistants to create those
descriptions and interpretations of the material for the student.
The second most often mentioned area of difficulty was foreign language materials, especially
when involving additional alphabets.
Furthermore, images, tables, multimedia may need special attention for conversion and
adaptation.
Music, theology, and professional school disciplines also present extra challenges.
Non-text formats are resource intensive, such as video and audio.
Focus group participants conveyed the need for flexibility and creativity in order to
provide disability accommodations.
They are a very flexible and very creative bunch.
Our White Paper describes in some detail additional factors such as the existing landscape of
resources and the behavior and practices of publishers.
Frankly, I decided at this hour of the day its too depressing to start that way.
But, they are important factors.
I hope you will peruse the White Paper for more information.
It's available through the links on the project-briefing page for this presentation.
Overall, it's very hard to quantify the cost associated with current ways schools
create and distribute accessible instructional materials.
But, I would like to offer and example for context.
One state institution that we know of, it has a population of roughly 28 thousand students,
the institution uses a mixture of full time staff, student workers, and third-party vendors.
They remediate, on average, 1,200 print items, and caption 1,500 multi-media items every
year.
In total, the annual budge for those processes is around 100 thousand dollars.
I believe that doesn't include some of the staffing costs for that single university
for that single year.
And that excludes any braille production, which varies significantly from year to year.
When those braille costs come in, they can have real spikes in terms of costs for people.
For a single student requesting an accessible version of an assigned text for class, accessibility
typically refers to a version of the text that can be or has been manipulated to meet
the student's need.
Whether the student will use the copy electronically, such as with screen reader software or other
assistive technologies, or needs a fixed version, like large-text or braille, an accessible
file will be a digital copy of a text with markup of headers and other formatting features,
alternative text to describe images and graphs, notifications of page-breaks and marginalia
and so on.
In short, the accessible file replicates the structure of the document in addition to the
content of the text.
I can't help but talk a little bit about the legal landscape here.
Our schools are prudent in attempts to mitigate and reduce risk.
There are two fundament legal areas that drive college and university behavior with accessible
course content, copyright infringement on the one hand, and civil rights violations
on the other.
Many schools remain concerned that the act of simply creating accessible version of course
materials is a violation of copyright and a legal liability.
The legal basis for providing reformatted and accessible course content and services
in libraries and universities is recently, strongly affirmed by the US Court of Appeals,
second circuit.
Ruled in Authors' Guild v. HathiTrust that, "the doctrine of fair use allows the Libraries,
to provide full digital access to copyrighted works to their print disabled patrons."
Through this ruling, the court removed considerable ambiguity regarding the rights of schools
and, for the first time, made clear that authorized entities can make copies for print disabled
users.
It is important to note, however, that the court was specifically addressing the provision
of works digitized, on mass, from library collections, not newly produced textbooks.
Further, although the logic of the ruling could be extended to other formats, such as
audio and video materials, the ruling doesn't specifically address their reproduction.
Nevertheless, the second circuit ruling should reassure libraries and educational institutions
they can lawfully make and provide specialized copies of educational materials for their
eligible students.
And it has encouraged many to expand their services, which is a good thing.
Libraries and their parent institutions must be attentive to numerous elements of potential
violation of copyrighted status with printed published scholarly works in media.
But in the context of accessible course materials, protection of copyright owners might take
a backseat to a greater area of legal risk.
The potential of a lawsuit or investigation for failure to provide adequate accommodation
for students with disabilities by the office of Civil Rights, by the Department of Justice,
or by individuals and advocacy agencies.
The office of Civil Rights is focused on the availability of alternate formats of materials
or other accommodations to provide access when a student need arises.
They have stipulated that the timeliness and ease of use of those materials is a critical
factor in ensuring equity.
Most importantly, institutions should be considering accessibility before an individual need may
arise.
This includes the holdings of the institution's library and all of the services it offers.
Libraries need to be able to provide equal access to the educational opportunities and
benefits of the library, all the collections, and all the services.
Despite government expectations, there is little evidence of widespread proactive efforts
by our schools to ensure readiness for requests.
While there's increased attention in the literature of libraries around accessibility, relatively
little is found around course materials or general collections or other types of services.
Through the combined pressures of liability, to demonstrate ADA compliance, and reassurances
of safe harbor through the doctrine of fair use, really makes this an excellent time for
further action to address this growing need.
Like it or not, there's at least one digital library on your campus that could use some
help.
Disability resources and services staff needs help with storage, metadata, multiple formats
and discovery.
Armed with the research of our project, we have a really, fairly clear, understanding
of the needs of the community so we're trying to move ahead with a project to create repository
series.
This will be a specification driven process.
We're now evaluating platform technology options in light of the awareness of needs to see
what is our best option going forward.
These are the current collaborators for the next part of this project as we seek our next
round of support.
Each one of these schools has identified participants from both the disability resource services
area and the library.
So, the project intends to build collaboration with the institutions while it also builds
collaboration across institutions.
We're hardly the first to pose file sharing.
Various groups and organizations have attempted this approach previously in areas around accessibility.
However, we're finding that as the population grows, the technology's improving and the
legal environment shifts.
It now is a better time and a good time for us to focus on sharing.
This is my favorite quote from our focus groups.
It gives me motivation when I feel like this is an impossible task to keep it going.
So, before I turn it over to Jody, one more point.
There's much more work to do.
The Horizon report says it very well…
"Libraries can pave the way for their campuses by working with other institutional stakeholders
to implement policies that ensure equality of opportunity for disabled students, faculty
and scholars."
In addition to our analysis of the environment, our White Paper includes as set of recommendations
directed towards libraries.
One, directed towards university and college administration, and one directed towards publishers.
We can always hope.
Any development of repository services is simply one tool in a very complex environment.
We're calling on libraries to take seriously the information needs of students with disabilities.
For those willing to connect with their disability services staff, we provide an appendix with
suggested topics for discussion and questions for working together to develop new service
workflow's.
For libraries interested in a more internal review of how the library handles accessibility,
there's an appendix of reflection question and topics to support making improvements
in your own environment.
By analyzing the local environment at a single institution, the library can create a set
of priorities for action, which will best assist their community and enable librarians
to be a bigger part and a better, stronger part of the community puzzle.
Our work has to be proactive.
It has to be collaborative and we really need to be persistent.
With that, Jody's turn.
Waypoints for a Roadmap
Okay, so I've been working with the accessibility side of things for ARL on the captioning project
starting in January and came into this conversation almost immediately as a pre-proposal was being
developed for a larger grant.
As part of that conversation, along with a lot of conversations with many of you.
I think I've probably talked to half you on the phone at some point or another, or
in a conference call.
I've been trying to put together, you know, pieces and parts of a puzzle that might lead
to a roadmap for project Plan.
Also, trying to spin up as fast as possible as far as where the state of the art is for
captioning, as it sits in this larger context of accessible materials.
So, I've started categorizing certain aspects of the conversations that I've had with
you, and put them together into the next slide, which is sort of way points for the roadmap.
Its not quite a roadmap.
I'm sorry its so text dense, but I didn't want to go through multiple, multiple slides.
These are areas where we've had some really fruitful, I think, conversations that have
helped me understand what might work, what's feasible and what's not.
The first part had to do with the discoverability issue.
One of the things that I think frustrates all of us is building a tool or service that
nobody uses because they can't find it.
That led me to questions around Laura's project about what kinds of metadata standards
exist for accessibility.
There have, in fact, been a fairly number of projects aimed at this, starting around
2011, really taking traction around 2013.
There is a W3C standard that includes accessible tagging and there are going to be slides here
with links to this material but I didn't want to waste our time together today going
over each piece and part.
We can do that in the Q and A, if you like.
Zeroing in on captioning, version one of that standard, which was developed by schema.org
and then adopted, has one tag for captioning and it is captioned—that's it.
Not quality of the caption, how the caption was generated, not whether it has error correction
or anything like that done to it.
So, one could hope for more articulate metadata but it's a start.
That means we wouldn't have to start from zero.
There is much more tagging associated with other types of accessible formats that I think
merits some revisiting.
A lot of this work came out of funding by the Gates Foundation through, I don't know
how to pronounce this, but it's A-11-Y metadata, which was a workgroup that you'll have a
link to.
So, the discoverability issue, I think, is one that is feasible, fungible, manageable,
needing a little additional work.
The next think I wanted to look at were the tools and technologies for captioning and
so have spent a fair amount of time asking and begging people to test things for me or
inform me about what they are currently using, what's working, what's clunky, what doesn't
work, and what gaps are there where we could maybe ask for better tools.
And, have played a little bit…have talked with YouTube about their automated captioning
process, which, by the way, has gotten better, which is not to say…the bar was pretty low
to begin with.
So, it's a comparative thing.
But, I would encourage you, if you're interested, to have a look again.
I've had some of our folks test it out on short videos and the accuracy rate seems to
be around 90-95%.
And, you can go back and do error correction or enhancement.
The challenge for automated captioning, of course, has to do with descriptions of sound
that are not words.
So, if there's an explosion in the background, someone has to type in and machines aren't
terribly good at figuring that out yet, though maybe eventually they will be.
The availability of desktop software has also increased as has commercial providers of,
of providing caption that many of you use.
I've played around myself with a couple of desktop things and I'm one, impressed by
how friendly they generally are but two, how labor-intensive of a process it is.
This is something that cries out for crowd sourcing, I think.
I think there are folks that would spend a couple of hours on a weekend doing error correction
if the tools were available for it.
The intellectual property considerations around video, I have some familiarity with having
an association with the Vanderbilt Television News Archive and lots of very friendly conversations
with legal council for national network news organizations.
Every time they get a new vice-president, I usually get a call to say, "What are you
doing with our stuff?" and "How dare you."
And "Oh, there is actually a section of the copyright law that allows us to do this."
So, dealing with the intellectual property issues, I think, a dicy thing, particularly,
as we first imagined we'd be storing video with embedded captioning in it.
But, in conversations since I've come on board, we thought about not storing video
with embedded captioning but just in the caption files in multiple formats.
Players, video players, including web-based HTML5 players, are able to merge the caption
file with the video at the point that it's being streamed.
And this would do two things for us, we think.
One, reduce the liability issue.
Obviously, I don't think we would be storing captioned files that were already on sale
by a publisher somewhere.
And two, would greatly limit the storage problem that video tends to present.
Caption files are small.
Text files, storing them would be a trivial matter.
Working with partners for Laura's project might be a way to deal with the issue of storage
and making them discoverable and marking them up and that sort of thing.
So, the intellectual property considerations, I think, are, we think, are also something
that we can work with.
Interesting side here…when I talked with YouTube about…I mean, clearly, they have
this problem, right?
They have developed a service called 'Content ID' where, if I upload your video and you
discover it, you can claim it as yours.
And effectively, any ad revenue that's generated by that item goes to you, not me.
In other words, I'm doing free labor for you if you're a publisher, for example.
And, I'm not sure that it wouldn't be a bad idea to tell publishers, "If you haven't,
if you don't have a captioned version of this, we're going to place it up into a shared
repository, lets have a conversation about whether we just give it to you or somehow
or another work out an arrangement where you benefit and therefore its not a question of
liability but a question of a business arrangement."
I know that's unheard of, but, you know, give it a try.
I've also spent a fair amount of time looking at the landscape of commercial providers.
As I say, I met and talked with YouTube.
Many of you use 3playmedia, I'm sure.
If you have this service on your campus, there fairly economical ways to have captions generated
if you want to.
At least one university has a site-wide subscription to deal with this issue of accommodation with
replay.
So, effectively its on-demand.
I may not be the fastest way, the most prompt way to get things to students but at least
its there.
And no, it is not cheap.
So this might be something that many of us going together might be able to look at.
Lots of you are interested, obviously, or you wouldn't be here, but there's a very
large list, growing list, of universities that are interested in dealing with this.
Partly, I think, because of compliance issues but also partly because it's the right thing
to do.
Many times, these issues are being dealt with as compliance problems but they're also part
and parcel of the diversity and inclusion movements that are sweeping our campuses.
This is a diverse…it's a diversity question.
This is a group of people with a different perspective that bring different abilities
and different views of the conversation to the table.
And as institutions of higher education, that is, after all, part of our mission.
Maybe a central part.
I have looked at architectural considerations and talked with many of you about whether
a centralized repository is the preferred route, ala Hathi or one of the others.
Or, whether we would want to have a linked set of repositories, possibly piggybacking
on the institutional repository structure.
I don't think there's any clear direction on that other than the answer seems to be
both.
It would be good to have a centralized, easily findable repository, but the ability to link
out to repositories around the country.
Types of materials of interest is also another question…how am I doing on time?
Okay.
As we think of the issues of what kinds of things go into a repository, at least initially,
some kind of prioritization might be in order.
Whether those are items that are, for example, a part of our collection that need to be captioned
or their instructional videos that are of general interest to many of us or they are
associational webinars.
One of the things I asked about the PowerPoint today was, "Will it be captioned?" and
the answer is, "yes.
It will be captioned."
So, you know, those kinds of things that are widespread in general interest, I think, are
natural targets for this kind of activity.
The business model and sustainability question, I really haven't made a lot of progress
on but it will have to be asked and answered at some point.
Grant funding will go someway and I think this is an eminently fundable area.
In the long run, we have to think through the issue of how do you keep it going into
the future.
And the same, I haven't made a lot of progress on identifying possible funding for pilot
projects.
I will, though, say there are a number of foundations that have expressed interest in
the past on this kind of work—Gates Foundation is one that did the metadata schema.org thing.
And, so I'm kind of confident that those things can be turned up.
And, I think I'll stop there and we'll wait.
I hope we'll have some good question and we can have a conversation.
The next two slides, or maybe there are more, are items that will be part of this deck that
you'll have access too.
These are just things that I've researched that you might want to visit if you haven't
already.
And, is there one more.
Yeah.
Here's the standards—that the A-11-Y metadata.
I was very pleased to find this.
I mean, this is a route by which tagged material makes its way into the large search engines.
Laura: Jody and I flipped a coin and decided that
I do a better Beth imitation, so I'll go it my best shot.
So, the University of Illinois has been working with HathiTrust Digital Library and their
ultimate goal is to develop a scalable approach to scanning, generating metadata and DAISY
files and making digital accessible text accessible to print disabled reader through the HathiTrust
Digital Library.
The pilot partners are the library and the digital resources in educational services—they
refer to them as DRES—unit of the University of Illinois at Urbana-Champaign, and the HathiTrust
Digital Library.
The initial goal is to develop and test and evaluate a workflow and access method for
the University of Illinois to produce digitized texts, metadata, and accessible digital text
(DAISY files), ingest them into the HathiTrust Digital Library, make them accessible to proxies
for download on behalf of print-disabled users.
Our initial goal was to have a pilot workflow outlined by December 1, 2016, so we could
put a test volume through the steps.
This would be utilized to have a more significant pilot of about 50 volumes ready to launch
in 2017.
The pilot would be used to refine the service, develop a more accurate cost model, and launch
a full program for late 2017/early 2018.
With responsibilities, the University of Illinois Library agreed to fund the acquisitions needed,
the associated Internet archive scanning, and the staff resources to work on the HathiTrust
ingest process.
DRES has assumed responsibility for generating the DAISY files, working with the JPEG 2000
images and OCR generated from the Internet archive scans.
So, to date, with scanning and OCR production, using Internet archives, to date the library
chose to scan the contents so it would mesh well with the Internet archives existing operating
and invoicing procedure.
In order to do so, the content had to be placed online.
If we want the images to be cropped and OCR processed, which we do, they need to be uploaded
and derived.
After discussing with staff from Internet archive, the recommendation was to load this
content into Internet Archives print-disabled collection, which is not findable, as it consists
of in-copyright content.
This interferes with Illinois' internet archive workflow the least and subsequently
would allow the internet archive to host and share the JPEG 2000 files for those items
so that with the login, DRES can go in a download the JPEG 2 files and manipulate them as needed
to create the proper OCR DAISY content.
This removes the added need for setting up shared server workspace and keeping it updated.
For availability in restricted Internet archive collections, from there, Internet archive
and I, Beth, decided that designating this content in the future as DRES would entail
the scanners uploading it to an access-restricted collection, which content appears only to
users who have the URL and proper permissions through an Internet archive log-in, which
they've made for preservation services to be shared with the cataloging and metadata
unit and eventually with DRES.
Following this, conversations with Hathi about how Hathi could inject this content resulted
in the solution of having Internet archive link the UIUC and HathiTrust accounts and
grant Hathi Internet archive with equal permissions to access and manipulate the disability resources
educational services collection.
The HathiTrust metadata access, Hathi was subsequently able to get this working in their
end and found they were able to get content working with their ingest tools in their development
environment, which means that once there's metadata for it in the HathiTrust Zephyr metadata
system, it can then be fully ingested.
On Illinois's part, we have been able to download and modify the metadata for a test
item.
In theory, we are on the cusp of being able to fully ingest the content.
T he final step for the pilot is to ensure that this content is fully accessible to a
proxy user.
For the future, the HathiTrust intends to start developing a workflow for ingesting
the metadata and making the DAISY files accessible to proxy users sometime in 2017.
From the usability end, our applied health sciences librarian is to focus on how these
items are used from a patron perception.
HathiTrust proxy system, in essence, allows her, as a proxy for the print-disabled patron,
to download the full PDF text from Hathi.
In future development, Hathi may also place a DAISY download under it, similar to what
Internet archive offers.
Their DAISY files are automatically generated but future unacceptable OCR quality for the
Illinois DRES level needs.
Where to store DAISY files and how to make the available to proxy's who need to download
them for print-disabled users is on the agenda for HathiTrust to consider in the future.
There is strong interests among the big ten academic alliance in seeing Hathi develop
the capability to ingest daisy files, OCR and metadata associated with the accessible
text; and to make this information available within the HathiTrust proxy service.
HathiTrust is clearly interested in working towards this, resources permitting.
Access to this content has the potential to be one of several federated components that
support more streamlined access to text and media for users with disabilities.
She lists here members of the project team with some of their contact information and
you can see, as she has mentioned, this is bringing people in from across the university
to work together.
So, now, we have time for questions.
For more infomation >> [ Animal Coloring Pages for Kids ] | Drawing and Coloring a Pearl - Duration: 2:34. 
For more infomation >> Uncut interview_lessourciersPart2 2016 (test for eng subs) - Duration: 20:48. 
For more infomation >> Angelica om Senter for Voksenopplæring, Stord - Duration: 1:24. 
For more infomation >> Microsoft SQL Server 2012 Tutorial for Beginners # 2 | Insert, sort,delete data in SQL Table - Duration: 43:08. 
For more infomation >> MH series built for speed.mp4 - Duration: 2:09. 
For more infomation >> Alex Ahrendtsen - Hvilket rum for musik drømte I om med Folkeskolereformen? - Duration: 0:42. 
For more infomation >> I've found my dream way of working... what's yours? - Duration: 3:42. 
Không có nhận xét nào:
Đăng nhận xét