I recall in a forum once, someone thought we should not allow recorded lectures to be available as podcasts because this would be unfair to deaf students.
So the spoken lecture is fine, but the podcast is not….
I think part of the problem is that people think in black and white terms, either/or and forget that we can have both or grey areas.
I was showing some staff an UMPC once, the Q1 Ultra, which I am thinking of using in our library, and first comment was that the 7″ screen would be too small for some students.
This is a fair comment, but I am not going to replace all the computers in the library with UMPCs, there would still be big computers with big screens for those that wanted them. The UMPCs would be in addition not a replacement. Some users will be fine with the UMPC, others will want what they see as a “normal” computer.
I would say it is similar with web services, just because a service is not accessible to everyone, doesn’t mean that it shouldn’t be used, but consideration needs to be given how you would support the users for whom it *may* be inaccessible.
In my lecture/podcast example, I would say that if a signer was provided for the lecture, then a signer could be provided for the podcast.
If “services must be accessible to all or they shouldn’t be deployed” then non-web services should be subject to the same constraints, in which case nothing would happen in an educational institution!
It’s not black and white, it’s grey.