E-mail List Archives
Thread: accessibility web proxy
Number of posts in this thread: 4 (In chronological order)
From: julien
Date: Wed, Feb 12 2014 10:00AM
Subject: accessibility web proxy
No previous message | Next message →
Hello,
This is probably a silly question, but I was curious why there was no such thing as an accessibility web proxy service on the market (I'm thinking about a web proxy server sitting between the user's PC and the server, that fixes and enhances html code on the fly, before it is interpreted by the screen reader on the client machine).
With a large enough community of subscribers, wouldn't such a service be able to address common pain points for blind web users like optimizing bad html, solving captcha, making popular hard-to-access websites accessible without installing scripts? Or even augmenting web pages (adding summaries or image descriptions when needed)?
Anyway, I'm sure others have explored this idea before (I have found a few academic papers mentioning this concept); what do you think are the main reasons it doesn't exist?
Thanks!
Julien
From: Stanzel, Susan - FSA, Kansas City, MO
Date: Wed, Feb 12 2014 10:43AM
Subject: Re: accessibility web proxy
← Previous message | Next message →
I think Deque sells such a service.
Susie Stanzel
From: Greg Kraus
Date: Wed, Feb 12 2014 12:12PM
Subject: Re: accessibility web proxy
← Previous message | Next message →
I developed a system just like this a few years ago. You could
crowd-source the "rewrites" to a page that needed to happen to make it
more accessible, and users could automatically have those changes
applied to their page if they wanted to. There are some really cool
aspects about a system like this, but I think there are some
significant downsides too.
1. The "crowd" must respond to unannounced changes in the source document.
2. It takes 100% of the responsibility off of the developer so there
is no incentive to code things right in the first place.
3. While this model works adequately well for static content, once you
get into dynamic content it gets much more difficult to manage.
4. Duct tape and bandages can be good short-term, temporary patches,
but they are no way to build a quality system.
The best feature of a system like this, and one of the main reasons I
developed my system, was so that I could give feedback to developers.
I could manually evaluate their page for accessibility and record my
notes in a database. Then the developer could execute a bookmarklet
while viewing their page, and the notes and changes I made in the
database would be applied. They could see the code changes plus
contextual notes I would leave them.
Greg
--
Greg Kraus
University IT Accessibility Coordinator
NC State University
919.513.4087
= EMAIL ADDRESS REMOVED =
http://go.ncsu.edu/itaccess
On Wed, Feb 12, 2014 at 12:43 PM, Stanzel, Susan - FSA, Kansas City,
MO < = EMAIL ADDRESS REMOVED = > wrote:
> I think Deque sells such a service.
>
> Susie Stanzel
>
>
From: julien
Date: Wed, Feb 12 2014 3:25PM
Subject: Re: accessibility web proxy
← Previous message | No next message
@Susan: Interesting, thank you! I had a look and it looks a lot like what I had in mind, although their service is purely B2B. It seems to indicate that technically this is quite feasible, interesting… Really wondering how it could be transposed in a B2C environment.
@Greg: Thanks for sharing!
1&4. Great points. Also, not all users are able to create or maintain scripts. I guess you could have a three pronged system: 1/ automated accessibility optimization (like the Deque service pointed out by Susan) that is applied to all sites accessed through the service 2/ crowd-sourced optimizations developped by tech savvy users (although one could create a simplified scripting engine?) 3/ optimizations developed and maintened by professional staff for a small number of highly-visited websites that accound for a large part of the user-community's web traffic.
2. True. And as you say, you can also use the accessibility bugs you collect to do some "advocacy work" with website owners ;) Having a community of users behind you also gives more weight to your recommandations I would guess.
3. I have no technical background whatsoever, so I can only begin to imagine how complex this can be. I'm wondering what a data scientist could do with the browsing data generated by the user community to generate generic solutions to the most frequent problems?
Also I guess there is some interesting possibilities to connect the system to external APIs (images auto-tagging, captcha solvers, text summarization etc) to improve the users' experience.