[Athen] RE: RoboBraille

[Brian Richwine]

Thanks Sean, very thoughtful and complete reply as always.

I see that the robobraille website has resources that might answer most of
my questions. When you talk about using the system as a reliable conversion
tool, what is the source document format that you recommend content authors
start with?

Thanks,
  Brian


On Wed, Nov 20, 2013 at 12:54 PM, Sean Keegan <skeegan at stanford.edu> wrote:


> In terms of quality of the output, I have found that it is based on

> several parameters, including:

>

> 1. quality of the original scan

> 2. complexity of the material being submitted

> 3. final document format

> 4. if the document included accessibility information during authoring

> process

>

> There may be others, but these seem to be the major parts that affect the

> output quality.

>

> For instance, if the intent is to go from an image PDF to a text-based

> (tagged) PDF and the image scan is poor, then the resulting PDF document

> will not be that usable for a student. On the other hand, if the original

> image scan is high-quality and a simple 1-2 column layout, then I have

> found the output for both the tagged PDF and MS Word format to be very

> usable for a student.

>

> If the document author includes accessibility information into the

> original document format, then the output can be quite accessible. For

> example, if the author includes text descriptions for images, heading

> markup, etc., then the output can be very usable if going to either a DAISY

> file or a tagged PDF version.

>

> Now, where the system breaks down is when you are dealing with formats

> that do not function well in an automated process, such as math and/or

> science materials. We point this out to students and let them decide how

> they may choose to proceed.

>

> I want to point out that a system such as this is not intended to replace

> alternate format production at an institution; rather, the idea is that it

> can augment such a production facility. There are times a student receives

> academic materials that are not in the format that works best for their

> disability needs and the student needs the materials quickly. An automated

> conversion process that allows a student to submit a file and receive a

> result in less than an hour offers more independence and faster results

> than going through a full production process. Further, a student may not be

> sure they even want to read the full document. We have had students use the

> automated process to perform a quick document conversion so they may decide

> if they want to make a request through our production facility.

>

> I  would also suggest that there is the potentiality for such a system to

> act as an "accessibility converter" and support the creation of accessible

> instructional materials from an institutional perspective. Rather than

> focusing on how to use all the different tools and editors to make an

> accessible document, I can focus on the authoring techniques that result in

> accessible versions. I can talk about the benefits of headings, image

> descriptions, data tables, math content, etc. and then avoid the painful

> conversation of "here's how you use the Touch-Up Reading Order Tool" in

> Adobe Acrobat. Instead, I can point the document creators to a reliable

> conversion tool that, provided they have followed the correct authoring

> techniques, can result in an accessible versions.

>

> There will always be the need for an alt format production facility to

> provide conversion services for students with print-disabilities. While not

> perfect, an automated system can provide an alternative to the traditional

> conversion services and offer students a choice in how they may wish to

> have their materials processed.

>

> So, to provide a long answer to your short question regarding the quality

> of materials - it depends. <grin>

>

> Take care,

> Sean

>

>

> On Nov 20, 2013, at 7:46 AM, Brian Richwine <blrichwine at gmail.com> wrote:

>

> Can anyone comment on the quality of the materials produced by the

> Robobraille/SensusAccess portal? Our students are used to having our

> editing staff produce fully edited accessible word documents (with

> appropriate heading structures, page number matching, described images,

> linearized/described tables, etc.).

>

> What kind of student satisfaction feedback are you getting?

>

> Sincerely,

>   Brian

>

>

> On Tue, Nov 19, 2013 at 7:35 PM, Sean Keegan <skeegan at stanford.edu> wrote:

>

>> To follow-up on Joshua's comment - the SCRIBE system uses Abbyy

>> FineReader Corporate Edition and the reason for this is because we manage

>> the system on-campus vs. the hosted solution other institutions are

>> implementing. Overall, it works okay, but Recognition Server is far more

>> robust in terms of speed and recovery from corrupted PDF documents.

>>

>> When we built SCRIBE, FineReader Corporate Edition was the only viable

>> option for us as Recognition Server was too expensive. Every so often I do

>> have to go and restart the Abbyy Corporate Edition application on our

>> system as a corrupted PDF file will stop the automatic processing and, from

>> what I can tell, this does not happen with Recognition Server (on the other

>> hand, you don't suffer quite as many heart attacks from the RS

>> pricing). Getting access to Recognition Server is far better for overall

>> reliability.

>>

>> I am not too concerned about the future of the Robobraille/SensusAccess

>> portal as there is too much investment by other entities around the world

>> to let it expire. As Joshua mentioned, I also suggest the system for users

>> once they leave the institutional environment and believe this resource

>> will continue until something better comes along.

>>

>> Take care,

>> sean

>>

>>

>> On Nov 19, 2013, at 4:06 PM, Joshua Hori <jhori at ucdavis.edu> wrote:

>>

>> Hello everyone!

>>

>>

>> Along with UC Irvine, UC Davis has also purchased a hosted license of

>> SensusAccess to be used by faculty, students, or staff using a UCDavis

>> email addresses. This is to ensure that our students can provide their own

>> conversions if needed and to support development. Currently, SensusAccess

>> is paid by European governments for European students. When the

>> SensusAccess group came to the US Government for funding, they were told to

>> request funds from institutions instead.

>>

>>

>> *So, what do I get with my license?*

>>

>> I get an iframe to add to one of my web pages which allows students to

>> submit a file from their computer (default on SA’s main page), or I have

>> the ability to post a file via URL (non-authentication), or I can copy and

>> paste text into a textbox for conversion.

>>

>>

>> I also get the ability to request features to be implemented. One feature

>> that was implemented at our request was the ability to have files available

>> as download links instead of being attached to emails. This was due to some

>> documents being too large (over 30M) for some email systems, to which they

>> added code to recognize file size limitations and send users a link instead

>> of an attachment. Technical issues are usually resolved via email within an

>> hour as well.

>>

>>

>> A benefit is that I don’t have to pay for that outrageous pricetag that

>> ABBYY recognition server was requesting, which increased depending on how

>> many cores you plan on using within your processor (look underneath my

>> signature for the quotes I was given…). I also don’t have to pay, or

>> maintain, a server, or have to worry about the setup, licensing, or

>> maintenance of all the programs needed for the conversion process. (MS

>> Office, *OpenOffice*, *Calibre*, *DAISY Pipeline*, NeoSpeech voices,

>> Mailserver, FTP, RoboBraille, and all the plug-ins and specialized

>> software). There is a slight difference between ABBYY Pro, which I believe

>> SCRIBE is using, and ABBYY recognition server, which is what SA is using.

>>

>>

>> I instruct our students on how to use the UC Davis hosted service for

>> their needs, and that they can continue to use the open SensusAccess portal

>> for conversions after higher ed. If this service dies after a few years,

>> then all my plans to support my students after higher ed vanishes as well.

>> NOT GOOD.

>>

>>

>> Sorry, I went full nerd on everyone. I know…I’m not supposed to go full

>> nerd, but…I did.

>>

>>

>> *TL;DR – Yeah, we pay for the service…and it has its’ ups and downs, but

>> yet…it’s beautiful.*

>>

>> Joshua Hori

>> Accessible Technology Analyst

>> University of California, Davis

>> Student Disability Center

>>

>>

>> ABBYY Recognition Server (OCR support)

>> ·         $3,366 for 1 million page conversions, $1,346 annual

>> maintenance (locked to a single core)

>> ·         $2,993 for 300k page conversions, $1,197 annual maintenance

>> (locked to a single core)

>> ·         $1,850 for 100k page conversions, $740 annual maintenance

>> (locked to a single core)

>> ·         Dual core license: $12k, $4,800 annual maintenance

>> ·         Quad core license:  $18k

>>

>> o   Has the ability to convert 100k pages per night

>>

>> Remember…this is just for the OCR software.

>>

>> *From:* athen-list-bounces at mailman1.u.washington.edu [mailto:athen-

>> list-bounces at mailman1.u.washington.edu] *On Behalf Of *James Bailey

>> *Sent:* Tuesday, November 19, 2013 9:56 AM

>> *To:* athen-list-bounces at mailman1.u.washington.edu;

>> athen-list at u.washington.edu

>> *Subject:* [Athen] RoboBraille

>>

>>

>>

>>

>>

>>

>>

>> Hello all:

>>

>> I think the RoboBraille service is great and I tell our Alt-format users

>> about it. I particularly recommend it for those study sessions when a

>> picture-only PDF comes off of BlackBoard etc. and a the student needs it

>> converted quickly to keep working.

>>

>> I have received an e-mail or two from them suggesting we (U Oregon) need

>> an agreement with them. We do not use it at all in our production process.

>> I simply make students aware of it and the students then use it or not as

>> individuals.

>>

>>

>>

>> This is from their web site (it’s cut and paste so any typos are theirs):

>>

>> “RoboBraille is available 24/7 as a self-service solution, it is free of

>> charge to all individual, non-commercial users and users need not register

>> in order to use the service. The objective is to support and promote

>> self-sufficiency of people with special needs socially, throughout the

>> educational system and on the labour market. As an additional benefit,

>> RoboBraille helps to protect the privacy of of those who need material in

>> alternate formats.”

>>

>> This seems pretty cut and dried to me.

>>

>> If you have a take on this, please share it.

>>

>>

>>

>> Oh it gets better!

>>

>>

>>

>> From: Tanja Stevns <tanja at sensus.dk>

>>

>> Sent: Tuesday, November 19, 2013 9:10 AM

>>

>> To: James Bailey

>>

>> Subject: SV: Document conversion

>>

>> James,

>>

>> The users go through our SensusAccess engine. The requests is coming from

>> University of Oregon. I am not talking about

>>

>> personal private use of individuals using RoboBraille as you are

>> referring to in this mail.

>>

>> Now I have tried to approach this in the best way possible but we will

>> take action on further ongoing use from

>>

>> universities who do not wish to make an agreement with us.

>>

>> Have a nice day,

>>

>> Tanja

>>

>>

>>

>> My take on this is that they are seeing our student e-mail accounts as

>> being the University of Oregon.

>>

>>

>>

>> As Tanja says, “Have a nice day”

>>

>>

>>

>> James

>>

>>

>>

>> _______________________________________________

>>

>> athen-list mailing list

>> athen-list at mailman1.u.washington.edu

>> http://mailman1.u.washington.edu/mailman/listinfo/athen-list

>>

>>

>>

>> _______________________________________________

>> athen-list mailing list

>> athen-list at mailman1.u.washington.edu

>> http://mailman1.u.washington.edu/mailman/listinfo/athen-list

>>

>>

> _______________________________________________

> athen-list mailing list

> athen-list at mailman1.u.washington.edu

> http://mailman1.u.washington.edu/mailman/listinfo/athen-list

>

>

>

> _______________________________________________

> athen-list mailing list

> athen-list at mailman1.u.washington.edu

> http://mailman1.u.washington.edu/mailman/listinfo/athen-list

>

>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman12.u.washington.edu/pipermail/athen-list/attachments/20131120/7e78c566/attachment.html>